diff --git a/swh/storage/buffer.py b/swh/storage/buffer.py index 8f86455b..7c73d3a9 100644 --- a/swh/storage/buffer.py +++ b/swh/storage/buffer.py @@ -1,109 +1,112 @@ # Copyright (C) 2019 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information from collections import deque from functools import partial from typing import Optional, Iterable, Dict from swh.core.utils import grouper from swh.storage import get_storage class BufferingProxyStorage: """Storage implementation in charge of accumulating objects prior to discussing with the "main" storage. Sample configuration use case for buffering storage: .. code-block:: yaml storage: cls: buffer args: storage: cls: remote args: http://storage.internal.staging.swh.network:5002/ min_batch_size: content: 10000 content_bytes: 100000000 + skipped_content: 10000 directory: 5000 revision: 1000 release: 10000 """ def __init__(self, storage, min_batch_size=None): self.storage = get_storage(**storage) if min_batch_size is None: min_batch_size = {} self.min_batch_size = { 'content': min_batch_size.get('content', 10000), 'content_bytes': min_batch_size.get('content_bytes', 100*1024*1024), + 'skipped_content': min_batch_size.get('skipped_content', 10000), 'directory': min_batch_size.get('directory', 25000), 'revision': min_batch_size.get('revision', 100000), 'release': min_batch_size.get('release', 100000), } - self.object_types = ['content', 'directory', 'revision', 'release'] + self.object_types = [ + 'content', 'skipped_content', 'directory', 'revision', 'release'] self._objects = {k: deque() for k in self.object_types} def __getattr__(self, key): if key.endswith('_add'): - object_type = key.split('_')[0] + object_type = key.rsplit('_', 1)[0] if object_type in self.object_types: return partial( self.object_add, object_type=object_type ) return getattr(self.storage, key) def content_add(self, content: Iterable[Dict]) -> Dict: """Enqueue contents to write to the storage. Following policies apply: - First, check if the queue's threshold is hit. If it is flush content to the storage. - If not, check if the total size of enqueued contents's threshold is hit. If it is flush content to the storage. """ s = self.object_add(content, object_type='content') if not s: q = self._objects['content'] total_size = sum(c['length'] for c in q) if total_size >= self.min_batch_size['content_bytes']: return self.flush(['content']) return s def flush(self, object_types: Optional[Iterable[str]] = None) -> Dict: if object_types is None: object_types = self.object_types summary = {} # type: Dict[str, Dict] for object_type in object_types: q = self._objects[object_type] for objs in grouper(q, n=self.min_batch_size[object_type]): add_fn = getattr(self.storage, '%s_add' % object_type) s = add_fn(objs) summary = {k: v + summary.get(k, 0) for k, v in s.items()} q.clear() return summary def object_add(self, objects: Iterable[Dict], *, object_type: str) -> Dict: """Enqueue objects to write to the storage. This checks if the queue's threshold is hit. If it is actually write those to the storage. """ q = self._objects[object_type] threshold = self.min_batch_size[object_type] q.extend(objects) if len(q) >= threshold: return self.flush() return {} diff --git a/swh/storage/cassandra/cql.py b/swh/storage/cassandra/cql.py index a42abf42..f521f639 100644 --- a/swh/storage/cassandra/cql.py +++ b/swh/storage/cassandra/cql.py @@ -1,623 +1,678 @@ # Copyright (C) 2019-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import functools import json import logging import random from typing import ( Any, Callable, Dict, Generator, Iterable, List, Optional, TypeVar ) from cassandra.cluster import ( Cluster, EXEC_PROFILE_DEFAULT, ExecutionProfile, ResultSet) from cassandra.policies import DCAwareRoundRobinPolicy, TokenAwarePolicy from cassandra.query import PreparedStatement from tenacity import retry, stop_after_attempt, wait_random_exponential from swh.model.model import ( Sha1Git, TimestampWithTimezone, Timestamp, Person, Content, + SkippedContent, ) from .common import Row, TOKEN_BEGIN, TOKEN_END, hash_url from .schema import CREATE_TABLES_QUERIES, HASH_ALGORITHMS logger = logging.getLogger(__name__) _execution_profiles = { EXEC_PROFILE_DEFAULT: ExecutionProfile( load_balancing_policy=TokenAwarePolicy(DCAwareRoundRobinPolicy())), } # Configuration for cassandra-driver's access to servers: # * hit the right server directly when sending a query (TokenAwarePolicy), # * if there's more than one, then pick one at random that's in the same # datacenter as the client (DCAwareRoundRobinPolicy) def create_keyspace(hosts: List[str], keyspace: str, port: int = 9042, *, durable_writes=True): cluster = Cluster( hosts, port=port, execution_profiles=_execution_profiles) session = cluster.connect() extra_params = '' if not durable_writes: extra_params = 'AND durable_writes = false' session.execute('''CREATE KEYSPACE IF NOT EXISTS "%s" WITH REPLICATION = { 'class' : 'SimpleStrategy', 'replication_factor' : 1 } %s; ''' % (keyspace, extra_params)) session.execute('USE "%s"' % keyspace) for query in CREATE_TABLES_QUERIES: session.execute(query) T = TypeVar('T') def _prepared_statement( query: str) -> Callable[[Callable[..., T]], Callable[..., T]]: """Returns a decorator usable on methods of CqlRunner, to inject them with a 'statement' argument, that is a prepared statement corresponding to the query. This only works on methods of CqlRunner, as preparing a statement requires a connection to a Cassandra server.""" def decorator(f): @functools.wraps(f) def newf(self, *args, **kwargs) -> T: if f.__name__ not in self._prepared_statements: statement: PreparedStatement = self._session.prepare(query) self._prepared_statements[f.__name__] = statement return f(self, *args, **kwargs, statement=self._prepared_statements[f.__name__]) return newf return decorator def _prepared_insert_statement(table_name: str, columns: List[str]): """Shorthand for using `_prepared_statement` for `INSERT INTO` statements.""" return _prepared_statement( 'INSERT INTO %s (%s) VALUES (%s)' % ( table_name, ', '.join(columns), ', '.join('?' for _ in columns), ) ) def _prepared_exists_statement(table_name: str): """Shorthand for using `_prepared_statement` for queries that only check which ids in a list exist in the table.""" return _prepared_statement(f'SELECT id FROM {table_name} WHERE id IN ?') class CqlRunner: """Class managing prepared statements and building queries to be sent to Cassandra.""" def __init__(self, hosts: List[str], keyspace: str, port: int): self._cluster = Cluster( hosts, port=port, execution_profiles=_execution_profiles) self._session = self._cluster.connect(keyspace) self._cluster.register_user_type( keyspace, 'microtimestamp_with_timezone', TimestampWithTimezone) self._cluster.register_user_type( keyspace, 'microtimestamp', Timestamp) self._cluster.register_user_type( keyspace, 'person', Person) self._prepared_statements: Dict[str, PreparedStatement] = {} ########################## # Common utility functions ########################## MAX_RETRIES = 3 @retry(wait=wait_random_exponential(multiplier=1, max=10), stop=stop_after_attempt(MAX_RETRIES)) def _execute_with_retries(self, statement, args) -> ResultSet: return self._session.execute(statement, args, timeout=1000.) @_prepared_statement('UPDATE object_count SET count = count + ? ' 'WHERE partition_key = 0 AND object_type = ?') def _increment_counter( self, object_type: str, nb: int, *, statement: PreparedStatement ) -> None: self._execute_with_retries(statement, [nb, object_type]) def _add_one( self, statement, object_type: str, obj, keys: List[str] ) -> None: self._increment_counter(object_type, 1) self._execute_with_retries( statement, [getattr(obj, key) for key in keys]) def _get_random_row(self, statement) -> Optional[Row]: """Takes a prepared statement of the form "SELECT * FROM WHERE token() > ? LIMIT 1" and uses it to return a random row""" token = random.randint(TOKEN_BEGIN, TOKEN_END) rows = self._execute_with_retries(statement, [token]) if not rows: # There are no row with a greater token; wrap around to get # the row with the smallest token rows = self._execute_with_retries(statement, [TOKEN_BEGIN]) if rows: return rows.one() else: return None def _missing(self, statement, ids): res = self._execute_with_retries(statement, [ids]) found_ids = {id_ for (id_,) in res} return [id_ for id_ in ids if id_ not in found_ids] ########################## # 'content' table ########################## _content_pk = ['sha1', 'sha1_git', 'sha256', 'blake2s256'] _content_keys = [ 'sha1', 'sha1_git', 'sha256', 'blake2s256', 'length', 'ctime', 'status'] + @_prepared_insert_statement('content', _content_keys) + def content_add_one(self, content, *, statement) -> None: + self._add_one(statement, 'content', content, self._content_keys) + @_prepared_statement('SELECT * FROM content WHERE ' + ' AND '.join(map('%s = ?'.__mod__, HASH_ALGORITHMS))) def content_get_from_pk( self, content_hashes: Dict[str, bytes], *, statement ) -> Optional[Row]: rows = list(self._execute_with_retries( statement, [content_hashes[algo] for algo in HASH_ALGORITHMS])) assert len(rows) <= 1 if rows: return rows[0] else: return None @_prepared_statement('SELECT * FROM content WHERE token(%s) > ? LIMIT 1' % ', '.join(_content_pk)) def content_get_random(self, *, statement) -> Optional[Row]: return self._get_random_row(statement) @_prepared_statement(('SELECT token({0}) AS tok, {1} FROM content ' 'WHERE token({0}) >= ? AND token({0}) <= ? LIMIT ?') .format(', '.join(_content_pk), ', '.join(_content_keys))) def content_get_token_range( self, start: int, end: int, limit: int, *, statement) -> Row: return self._execute_with_retries(statement, [start, end, limit]) ########################## # 'content_by_*' tables ########################## @_prepared_statement('SELECT sha1_git FROM content_by_sha1_git ' 'WHERE sha1_git IN ?') def content_missing_by_sha1_git( self, ids: List[bytes], *, statement) -> List[bytes]: return self._missing(statement, ids) - @_prepared_insert_statement('content', _content_keys) - def content_add_one(self, content, *, statement) -> None: - self._add_one(statement, 'content', content, self._content_keys) - def content_index_add_one(self, main_algo: str, content: Content) -> None: query = 'INSERT INTO content_by_{algo} ({cols}) VALUES ({values})' \ .format(algo=main_algo, cols=', '.join(self._content_pk), values=', '.join('%s' for _ in self._content_pk)) self._execute_with_retries( query, [content.get_hash(algo) for algo in self._content_pk]) def content_get_pks_from_single_hash( self, algo: str, hash_: bytes) -> List[Row]: assert algo in HASH_ALGORITHMS query = 'SELECT * FROM content_by_{algo} WHERE {algo} = %s'.format( algo=algo) return list(self._execute_with_retries(query, [hash_])) + ########################## + # 'skipped_content' table + ########################## + + _skipped_content_pk = ['sha1', 'sha1_git', 'sha256', 'blake2s256'] + _skipped_content_keys = [ + 'sha1', 'sha1_git', 'sha256', 'blake2s256', 'length', + 'ctime', 'status', 'reason', 'origin'] + _magic_null_pk = b'' + """ + NULLs are not allowed in primary keys; instead use an empty + value + """ + + @_prepared_insert_statement('skipped_content', _skipped_content_keys) + def skipped_content_add_one(self, content, *, statement) -> None: + content = content.to_dict() + for key in self._skipped_content_pk: + if content[key] is None: + content[key] = self._magic_null_pk + content = SkippedContent.from_dict(content) + self._add_one(statement, 'skipped_content', content, + self._skipped_content_keys) + + @_prepared_statement('SELECT * FROM skipped_content WHERE ' + + ' AND '.join(map('%s = ?'.__mod__, HASH_ALGORITHMS))) + def skipped_content_get_from_pk( + self, content_hashes: Dict[str, bytes], *, statement + ) -> Optional[Row]: + rows = list(self._execute_with_retries( + statement, [content_hashes[algo] or self._magic_null_pk + for algo in HASH_ALGORITHMS])) + assert len(rows) <= 1 + if rows: + # TODO: convert _magic_null_pk back to None? + return rows[0] + else: + return None + + ########################## + # 'skipped_content_by_*' tables + ########################## + + def skipped_content_index_add_one( + self, main_algo: str, content: Content) -> None: + assert content.get_hash(main_algo) is not None + query = ('INSERT INTO skipped_content_by_{algo} ({cols}) ' + 'VALUES ({values})').format( + algo=main_algo, cols=', '.join(self._content_pk), + values=', '.join('%s' for _ in self._content_pk)) + self._execute_with_retries( + query, [content.get_hash(algo) or self._magic_null_pk + for algo in self._content_pk]) + ########################## # 'revision' table ########################## _revision_keys = [ 'id', 'date', 'committer_date', 'type', 'directory', 'message', 'author', 'committer', 'synthetic', 'metadata'] @_prepared_exists_statement('revision') def revision_missing(self, ids: List[bytes], *, statement) -> List[bytes]: return self._missing(statement, ids) @_prepared_insert_statement('revision', _revision_keys) def revision_add_one(self, revision: Dict[str, Any], *, statement) -> None: self._add_one(statement, 'revision', revision, self._revision_keys) @_prepared_statement('SELECT id FROM revision WHERE id IN ?') def revision_get_ids(self, revision_ids, *, statement) -> ResultSet: return self._execute_with_retries( statement, [revision_ids]) @_prepared_statement('SELECT * FROM revision WHERE id IN ?') def revision_get(self, revision_ids, *, statement) -> ResultSet: return self._execute_with_retries( statement, [revision_ids]) @_prepared_statement('SELECT * FROM revision WHERE token(id) > ? LIMIT 1') def revision_get_random(self, *, statement) -> Optional[Row]: return self._get_random_row(statement) ########################## # 'revision_parent' table ########################## _revision_parent_keys = ['id', 'parent_rank', 'parent_id'] @_prepared_insert_statement('revision_parent', _revision_parent_keys) def revision_parent_add_one( self, id_: Sha1Git, parent_rank: int, parent_id: Sha1Git, *, statement) -> None: self._execute_with_retries( statement, [id_, parent_rank, parent_id]) @_prepared_statement('SELECT parent_id FROM revision_parent WHERE id = ?') def revision_parent_get( self, revision_id: Sha1Git, *, statement) -> ResultSet: return self._execute_with_retries( statement, [revision_id]) ########################## # 'release' table ########################## _release_keys = [ 'id', 'target', 'target_type', 'date', 'name', 'message', 'author', 'synthetic'] @_prepared_exists_statement('release') def release_missing(self, ids: List[bytes], *, statement) -> List[bytes]: return self._missing(statement, ids) @_prepared_insert_statement('release', _release_keys) def release_add_one(self, release: Dict[str, Any], *, statement) -> None: self._add_one(statement, 'release', release, self._release_keys) @_prepared_statement('SELECT * FROM release WHERE id in ?') def release_get(self, release_ids: List[str], *, statement) -> None: return self._execute_with_retries(statement, [release_ids]) @_prepared_statement('SELECT * FROM release WHERE token(id) > ? LIMIT 1') def release_get_random(self, *, statement) -> Optional[Row]: return self._get_random_row(statement) ########################## # 'directory' table ########################## _directory_keys = ['id'] @_prepared_exists_statement('directory') def directory_missing(self, ids: List[bytes], *, statement) -> List[bytes]: return self._missing(statement, ids) @_prepared_insert_statement('directory', _directory_keys) def directory_add_one(self, directory_id: Sha1Git, *, statement) -> None: """Called after all calls to directory_entry_add_one, to commit/finalize the directory.""" self._execute_with_retries(statement, [directory_id]) self._increment_counter('directory', 1) @_prepared_statement('SELECT * FROM directory WHERE token(id) > ? LIMIT 1') def directory_get_random(self, *, statement) -> Optional[Row]: return self._get_random_row(statement) ########################## # 'directory_entry' table ########################## _directory_entry_keys = ['directory_id', 'name', 'type', 'target', 'perms'] @_prepared_insert_statement('directory_entry', _directory_entry_keys) def directory_entry_add_one( self, entry: Dict[str, Any], *, statement) -> None: self._execute_with_retries( statement, [entry[key] for key in self._directory_entry_keys]) @_prepared_statement('SELECT * FROM directory_entry ' 'WHERE directory_id IN ?') def directory_entry_get(self, directory_ids, *, statement) -> ResultSet: return self._execute_with_retries( statement, [directory_ids]) ########################## # 'snapshot' table ########################## _snapshot_keys = ['id'] @_prepared_exists_statement('snapshot') def snapshot_missing(self, ids: List[bytes], *, statement) -> List[bytes]: return self._missing(statement, ids) @_prepared_insert_statement('snapshot', _snapshot_keys) def snapshot_add_one(self, snapshot_id: Sha1Git, *, statement) -> None: self._execute_with_retries(statement, [snapshot_id]) self._increment_counter('snapshot', 1) @_prepared_statement('SELECT * FROM snapshot ' 'WHERE id = ?') def snapshot_get(self, snapshot_id: Sha1Git, *, statement) -> ResultSet: return self._execute_with_retries(statement, [snapshot_id]) @_prepared_statement('SELECT * FROM snapshot WHERE token(id) > ? LIMIT 1') def snapshot_get_random(self, *, statement) -> Optional[Row]: return self._get_random_row(statement) ########################## # 'snapshot_branch' table ########################## _snapshot_branch_keys = ['snapshot_id', 'name', 'target_type', 'target'] @_prepared_insert_statement('snapshot_branch', _snapshot_branch_keys) def snapshot_branch_add_one( self, branch: Dict[str, Any], *, statement) -> None: self._execute_with_retries( statement, [branch[key] for key in self._snapshot_branch_keys]) @_prepared_statement('SELECT ascii_bins_count(target_type) AS counts ' 'FROM snapshot_branch ' 'WHERE snapshot_id = ? ') def snapshot_count_branches( self, snapshot_id: Sha1Git, *, statement) -> ResultSet: return self._execute_with_retries(statement, [snapshot_id]) @_prepared_statement('SELECT * FROM snapshot_branch ' 'WHERE snapshot_id = ? AND name >= ?' 'LIMIT ?') def snapshot_branch_get( self, snapshot_id: Sha1Git, from_: bytes, limit: int, *, statement) -> None: return self._execute_with_retries( statement, [snapshot_id, from_, limit]) ########################## # 'origin' table ########################## origin_keys = ['sha1', 'url', 'type', 'next_visit_id'] @_prepared_statement('INSERT INTO origin (sha1, url, next_visit_id) ' 'VALUES (?, ?, 1) IF NOT EXISTS') def origin_add_one(self, origin: Dict[str, Any], *, statement) -> None: self._execute_with_retries( statement, [hash_url(origin['url']), origin['url']]) self._increment_counter('origin', 1) @_prepared_statement('SELECT * FROM origin WHERE sha1 = ?') def origin_get_by_sha1(self, sha1: bytes, *, statement) -> ResultSet: return self._execute_with_retries(statement, [sha1]) def origin_get_by_url(self, url: str) -> ResultSet: return self.origin_get_by_sha1(hash_url(url)) @_prepared_statement( f'SELECT token(sha1) AS tok, {", ".join(origin_keys)} ' f'FROM origin WHERE token(sha1) >= ? LIMIT ?') def origin_list( self, start_token: int, limit: int, *, statement) -> ResultSet: return self._execute_with_retries( statement, [start_token, limit]) @_prepared_statement('SELECT * FROM origin') def origin_iter_all(self, *, statement) -> ResultSet: return self._execute_with_retries(statement, []) @_prepared_statement('SELECT next_visit_id FROM origin WHERE sha1 = ?') def _origin_get_next_visit_id( self, origin_sha1: bytes, *, statement) -> int: rows = list(self._execute_with_retries(statement, [origin_sha1])) assert len(rows) == 1 # TODO: error handling return rows[0].next_visit_id @_prepared_statement('UPDATE origin SET next_visit_id=? ' 'WHERE sha1 = ? IF next_visit_id=?') def origin_generate_unique_visit_id( self, origin_url: str, *, statement) -> int: origin_sha1 = hash_url(origin_url) next_id = self._origin_get_next_visit_id(origin_sha1) while True: res = list(self._execute_with_retries( statement, [next_id+1, origin_sha1, next_id])) assert len(res) == 1 if res[0].applied: # No data race return next_id else: # Someone else updated it before we did, let's try again next_id = res[0].next_visit_id # TODO: abort after too many attempts return next_id ########################## # 'origin_visit' table ########################## _origin_visit_keys = [ 'origin', 'visit', 'type', 'date', 'status', 'metadata', 'snapshot'] _origin_visit_update_keys = [ 'type', 'date', 'status', 'metadata', 'snapshot'] @_prepared_statement('SELECT * FROM origin_visit ' 'WHERE origin = ? AND visit > ?') def _origin_visit_get_no_limit( self, origin_url: str, last_visit: int, *, statement) -> ResultSet: return self._execute_with_retries(statement, [origin_url, last_visit]) @_prepared_statement('SELECT * FROM origin_visit ' 'WHERE origin = ? AND visit > ? LIMIT ?') def _origin_visit_get_limit( self, origin_url: str, last_visit: int, limit: int, *, statement ) -> ResultSet: return self._execute_with_retries( statement, [origin_url, last_visit, limit]) def origin_visit_get( self, origin_url: str, last_visit: Optional[int], limit: Optional[int]) -> ResultSet: if last_visit is None: last_visit = -1 if limit is None: return self._origin_visit_get_no_limit(origin_url, last_visit) else: return self._origin_visit_get_limit(origin_url, last_visit, limit) def origin_visit_update( self, origin_url: str, visit_id: int, updates: Dict[str, Any] ) -> None: set_parts = [] args: List[Any] = [] for (column, value) in updates.items(): set_parts.append(f'{column} = %s') if column == 'metadata': args.append(json.dumps(value)) else: args.append(value) if not set_parts: return query = ('UPDATE origin_visit SET ' + ', '.join(set_parts) + ' WHERE origin = %s AND visit = %s') self._execute_with_retries( query, args + [origin_url, visit_id]) @_prepared_insert_statement('origin_visit', _origin_visit_keys) def origin_visit_add_one( self, visit: Dict[str, Any], *, statement) -> None: self._execute_with_retries( statement, [visit[key] for key in self._origin_visit_keys]) self._increment_counter('origin_visit', 1) @_prepared_statement( 'UPDATE origin_visit SET ' + ', '.join('%s = ?' % key for key in _origin_visit_update_keys) + ' WHERE origin = ? AND visit = ?') def origin_visit_upsert( self, visit: Dict[str, Any], *, statement) -> None: self._execute_with_retries( statement, [visit.get(key) for key in self._origin_visit_update_keys] + [visit['origin'], visit['visit']]) # TODO: check if there is already one self._increment_counter('origin_visit', 1) @_prepared_statement('SELECT * FROM origin_visit ' 'WHERE origin = ? AND visit = ?') def origin_visit_get_one( self, origin_url: str, visit_id: int, *, statement) -> Optional[Row]: # TODO: error handling rows = list(self._execute_with_retries( statement, [origin_url, visit_id])) if rows: return rows[0] else: return None @_prepared_statement('SELECT * FROM origin_visit ' 'WHERE origin = ?') def origin_visit_get_all(self, origin_url: str, *, statement) -> ResultSet: return self._execute_with_retries( statement, [origin_url]) @_prepared_statement('SELECT * FROM origin_visit WHERE origin = ?') def origin_visit_get_latest( self, origin: str, allowed_statuses: Optional[Iterable[str]], require_snapshot: bool, *, statement) -> Optional[Row]: # TODO: do the ordering and filtering in Cassandra rows = list(self._execute_with_retries(statement, [origin])) rows.sort(key=lambda row: (row.date, row.visit), reverse=True) for row in rows: if require_snapshot and row.snapshot is None: continue if allowed_statuses is not None \ and row.status not in allowed_statuses: continue if row.snapshot is not None and \ self.snapshot_missing([row.snapshot]): raise ValueError('visit references unknown snapshot') return row else: return None @_prepared_statement('SELECT * FROM origin_visit WHERE token(origin) >= ?') def _origin_visit_iter_from( self, min_token: int, *, statement) -> Generator[Row, None, None]: yield from self._execute_with_retries(statement, [min_token]) @_prepared_statement('SELECT * FROM origin_visit WHERE token(origin) < ?') def _origin_visit_iter_to( self, max_token: int, *, statement) -> Generator[Row, None, None]: yield from self._execute_with_retries(statement, [max_token]) def origin_visit_iter( self, start_token: int) -> Generator[Row, None, None]: """Returns all origin visits in order from this token, and wraps around the token space.""" yield from self._origin_visit_iter_from(start_token) yield from self._origin_visit_iter_to(start_token) ########################## # 'tool' table ########################## _tool_keys = ['id', 'name', 'version', 'configuration'] @_prepared_insert_statement('tool_by_uuid', _tool_keys) def tool_by_uuid_add_one(self, tool: Dict[str, Any], *, statement) -> None: self._execute_with_retries( statement, [tool[key] for key in self._tool_keys]) @_prepared_insert_statement('tool', _tool_keys) def tool_add_one(self, tool: Dict[str, Any], *, statement) -> None: self._execute_with_retries( statement, [tool[key] for key in self._tool_keys]) self._increment_counter('tool', 1) @_prepared_statement('SELECT id FROM tool ' 'WHERE name = ? AND version = ? ' 'AND configuration = ?') def tool_get_one_uuid( self, name: str, version: str, configuration: Dict[str, Any], *, statement) -> Optional[str]: rows = list(self._execute_with_retries( statement, [name, version, configuration])) if rows: assert len(rows) == 1 return rows[0].id else: return None ########################## # Miscellaneous ########################## @_prepared_statement('SELECT uuid() FROM revision LIMIT 1;') def check_read(self, *, statement): self._execute_with_retries(statement, []) @_prepared_statement('SELECT object_type, count FROM object_count ' 'WHERE partition_key=0') def stat_counters(self, *, statement) -> ResultSet: return self._execute_with_retries( statement, []) diff --git a/swh/storage/cassandra/schema.py b/swh/storage/cassandra/schema.py index 34e683ae..cb4cf182 100644 --- a/swh/storage/cassandra/schema.py +++ b/swh/storage/cassandra/schema.py @@ -1,202 +1,226 @@ # Copyright (C) 2019-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information CREATE_TABLES_QUERIES = ''' CREATE OR REPLACE FUNCTION ascii_bins_count_sfunc ( state tuple>, -- (nb_none, map) bin_name ascii ) CALLED ON NULL INPUT RETURNS tuple> LANGUAGE java AS $$ if (bin_name == null) { state.setInt(0, state.getInt(0) + 1); } else { Map counters = state.getMap( 1, String.class, Integer.class); Integer nb = counters.get(bin_name); if (nb == null) { nb = 0; } counters.put(bin_name, nb + 1); state.setMap(1, counters, String.class, Integer.class); } return state; $$ ; CREATE OR REPLACE AGGREGATE ascii_bins_count ( ascii ) SFUNC ascii_bins_count_sfunc STYPE tuple> INITCOND (0, {}) ; CREATE TYPE IF NOT EXISTS microtimestamp ( seconds bigint, microseconds int ); CREATE TYPE IF NOT EXISTS microtimestamp_with_timezone ( timestamp frozen, offset smallint, negative_utc boolean ); CREATE TYPE IF NOT EXISTS person ( fullname blob, name blob, email blob ); CREATE TABLE IF NOT EXISTS content ( sha1 blob, sha1_git blob, sha256 blob, blake2s256 blob, length bigint, ctime timestamp, -- creation time, i.e. time of (first) injection into the storage status ascii, PRIMARY KEY ((sha1, sha1_git, sha256, blake2s256)) ); +CREATE TABLE IF NOT EXISTS skipped_content ( + sha1 blob, + sha1_git blob, + sha256 blob, + blake2s256 blob, + length bigint, + ctime timestamp, + -- creation time, i.e. time of (first) injection into the storage + status ascii, + reason text, + origin text, + PRIMARY KEY ((sha1, sha1_git, sha256, blake2s256)) +); + CREATE TABLE IF NOT EXISTS revision ( id blob PRIMARY KEY, date microtimestamp_with_timezone, committer_date microtimestamp_with_timezone, type ascii, directory blob, -- source code "root" directory message blob, author person, committer person, synthetic boolean, -- true iff revision has been created by Software Heritage metadata text -- extra metadata as JSON(tarball checksums, -- extra commit information, etc...) ); CREATE TABLE IF NOT EXISTS revision_parent ( id blob, parent_rank int, -- parent position in merge commits, 0-based parent_id blob, PRIMARY KEY ((id), parent_rank) ); CREATE TABLE IF NOT EXISTS release ( id blob PRIMARY KEY, target_type ascii, target blob, date microtimestamp_with_timezone, name blob, message blob, author person, synthetic boolean, -- true iff release has been created by Software Heritage ); CREATE TABLE IF NOT EXISTS directory ( id blob PRIMARY KEY, ); CREATE TABLE IF NOT EXISTS directory_entry ( directory_id blob, name blob, -- path name, relative to containing dir target blob, perms int, -- unix-like permissions type ascii, -- target type PRIMARY KEY ((directory_id), name) ); CREATE TABLE IF NOT EXISTS snapshot ( id blob PRIMARY KEY, ); -- For a given snapshot_id, branches are sorted by their name, -- allowing easy pagination. CREATE TABLE IF NOT EXISTS snapshot_branch ( snapshot_id blob, name blob, target_type ascii, target blob, PRIMARY KEY ((snapshot_id), name) ); CREATE TABLE IF NOT EXISTS origin_visit ( origin text, visit bigint, date timestamp, type text, status ascii, metadata text, snapshot blob, PRIMARY KEY ((origin), visit) ); CREATE TABLE IF NOT EXISTS origin ( sha1 blob PRIMARY KEY, url text, type text, next_visit_id int, -- We need integer visit ids for compatibility with the pgsql -- storage, so we're using lightweight transactions with this trick: -- https://stackoverflow.com/a/29391877/539465 ); CREATE TABLE IF NOT EXISTS tool_by_uuid ( id timeuuid PRIMARY KEY, name ascii, version ascii, configuration blob, ); CREATE TABLE IF NOT EXISTS tool ( id timeuuid, name ascii, version ascii, configuration blob, PRIMARY KEY ((name, version, configuration)) ) CREATE TABLE IF NOT EXISTS object_count ( partition_key smallint, -- Constant, must always be 0 object_type ascii, count counter, PRIMARY KEY ((partition_key), object_type) ); '''.split('\n\n') CONTENT_INDEX_TEMPLATE = ''' CREATE TABLE IF NOT EXISTS content_by_{main_algo} ( sha1 blob, sha1_git blob, sha256 blob, blake2s256 blob, PRIMARY KEY (({main_algo}), {other_algos}) -);''' +); + +CREATE TABLE IF NOT EXISTS skipped_content_by_{main_algo} ( + sha1 blob, + sha1_git blob, + sha256 blob, + blake2s256 blob, + PRIMARY KEY (({main_algo}), {other_algos}) +); +''' -TABLES = ('content revision revision_parent release directory ' - 'directory_entry snapshot snapshot_branch origin_visit ' - 'origin tool_by_uuid tool object_count').split() +TABLES = ('skipped_content content revision revision_parent release ' + 'directory directory_entry snapshot snapshot_branch ' + 'origin_visit origin tool_by_uuid tool object_count').split() HASH_ALGORITHMS = ['sha1', 'sha1_git', 'sha256', 'blake2s256'] for main_algo in HASH_ALGORITHMS: - CREATE_TABLES_QUERIES.append(CONTENT_INDEX_TEMPLATE.format( + CREATE_TABLES_QUERIES.extend(CONTENT_INDEX_TEMPLATE.format( main_algo=main_algo, other_algos=', '.join( [algo for algo in HASH_ALGORITHMS if algo != main_algo]) - )) + ).split('\n\n')) TABLES.append('content_by_%s' % main_algo) + TABLES.append('skipped_content_by_%s' % main_algo) diff --git a/swh/storage/cassandra/storage.py b/swh/storage/cassandra/storage.py index 92449780..454c2954 100644 --- a/swh/storage/cassandra/storage.py +++ b/swh/storage/cassandra/storage.py @@ -1,952 +1,988 @@ # Copyright (C) 2019-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import datetime import json import random import re from typing import Any, Dict, List, Optional import uuid import attr import dateutil from swh.model.model import ( - Revision, Release, Directory, DirectoryEntry, Content, OriginVisit, + Revision, Release, Directory, DirectoryEntry, Content, SkippedContent, + OriginVisit, ) from swh.objstorage import get_objstorage from swh.objstorage.exc import ObjNotFoundError try: from swh.journal.writer import get_journal_writer except ImportError: get_journal_writer = None # type: ignore # mypy limitation, see https://github.com/python/mypy/issues/1153 from .common import TOKEN_BEGIN, TOKEN_END from .converters import ( revision_to_db, revision_from_db, release_to_db, release_from_db, ) from .cql import CqlRunner from .schema import HASH_ALGORITHMS # Max block size of contents to return BULK_BLOCK_CONTENT_LEN_MAX = 10000 def now(): return datetime.datetime.now(tz=datetime.timezone.utc) class CassandraStorage: def __init__(self, hosts, keyspace, objstorage, port=9042, journal_writer=None): self._cql_runner = CqlRunner(hosts, keyspace, port) self.objstorage = get_objstorage(**objstorage) if journal_writer: self.journal_writer = get_journal_writer(**journal_writer) else: self.journal_writer = None def check_config(self, *, check_write): self._cql_runner.check_read() return True def _content_add(self, contents, with_data): contents = [Content.from_dict(c) for c in contents] # Filter-out content already in the database. contents = [c for c in contents if not self._cql_runner.content_get_from_pk(c.to_dict())] if self.journal_writer: for content in contents: content = content.to_dict() if 'data' in content: del content['data'] self.journal_writer.write_addition('content', content) count_contents = 0 count_content_added = 0 count_content_bytes_added = 0 for content in contents: # First insert to the objstorage, if the endpoint is # `content_add` (as opposed to `content_add_metadata`). # TODO: this should probably be done in concurrently to inserting # in index tables (but still before the main table; so an entry is # only added to the main table after everything else was # successfully inserted. count_contents += 1 if content.status != 'absent': count_content_added += 1 if with_data: content_data = content.data count_content_bytes_added += len(content_data) self.objstorage.add(content_data, content.sha1) # Then add to index tables for algo in HASH_ALGORITHMS: self._cql_runner.content_index_add_one(algo, content) # Then to the main table self._cql_runner.content_add_one(content) # Note that we check for collisions *after* inserting. This # differs significantly from the pgsql storage, but checking # before insertion does not provide any guarantee in case # another thread inserts the colliding hash at the same time. # # The proper way to do it would probably be a BATCH, but this # would be inefficient because of the number of partitions we # need to affect (len(HASH_ALGORITHMS)+1, which is currently 5) for algo in {'sha1', 'sha1_git'}: pks = self._cql_runner.content_get_pks_from_single_hash( algo, content.get_hash(algo)) if len(pks) > 1: # There are more than the one we just inserted. from .. import HashCollision raise HashCollision(algo, content.get_hash(algo), pks) summary = { 'content:add': count_content_added, - 'skipped_content:add': count_contents - count_content_added, } if with_data: summary['content:add:bytes'] = count_content_bytes_added return summary def content_add(self, content): content = [c.copy() for c in content] # semi-shallow copy for item in content: item['ctime'] = now() return self._content_add(content, with_data=True) def content_update(self, content, keys=[]): raise NotImplementedError( 'content_update is not supported by the Cassandra backend') def content_add_metadata(self, content): return self._content_add(content, with_data=False) def content_get(self, content): if len(content) > BULK_BLOCK_CONTENT_LEN_MAX: raise ValueError( "Sending at most %s contents." % BULK_BLOCK_CONTENT_LEN_MAX) for obj_id in content: try: data = self.objstorage.get(obj_id) except ObjNotFoundError: yield None continue yield {'sha1': obj_id, 'data': data} def content_get_partition( self, partition_id: int, nb_partitions: int, limit: int = 1000, page_token: str = None): if limit is None: raise ValueError('Development error: limit should not be None') # Compute start and end of the range of tokens covered by the # requested partition partition_size = (TOKEN_END-TOKEN_BEGIN)//nb_partitions range_start = TOKEN_BEGIN + partition_id*partition_size range_end = TOKEN_BEGIN + (partition_id+1)*partition_size # offset the range start according to the `page_token`. if page_token is not None: if not (range_start <= int(page_token) <= range_end): raise ValueError('Invalid page_token.') range_start = int(page_token) # Get the first rows of the range rows = self._cql_runner.content_get_token_range( range_start, range_end, limit) rows = list(rows) if len(rows) == limit: next_page_token: Optional[str] = str(rows[-1].tok+1) else: next_page_token = None return { 'contents': [row._asdict() for row in rows if row.status != 'absent'], 'next_page_token': next_page_token, } def content_get_metadata( self, contents: List[bytes]) -> Dict[bytes, List[Dict]]: result: Dict[bytes, List[Dict]] = {sha1: [] for sha1 in contents} for sha1 in contents: # Get all (sha1, sha1_git, sha256, blake2s256) whose sha1 # matches the argument, from the index table ('content_by_sha1') pks = self._cql_runner.content_get_pks_from_single_hash( 'sha1', sha1) if pks: # TODO: what to do if there are more than one? pk = pks[0] # Query the main table ('content') res = self._cql_runner.content_get_from_pk(pk._asdict()) # Rows in 'content' are inserted after corresponding # rows in 'content_by_*', so we might be missing it if res: content_metadata = res._asdict() content_metadata.pop('ctime') result[content_metadata['sha1']].append(content_metadata) return result - def skipped_content_missing(self, contents): - # TODO - raise NotImplementedError('not yet supported for Cassandra') - def content_find(self, content): # Find an algorithm that is common to all the requested contents. # It will be used to do an initial filtering efficiently. filter_algos = list(set(content).intersection(HASH_ALGORITHMS)) if not filter_algos: raise ValueError('content keys must contain at least one of: ' '%s' % ', '.join(sorted(HASH_ALGORITHMS))) common_algo = filter_algos[0] # Find all contents whose common_algo matches at least one # of the requests. found_pks = self._cql_runner.content_get_pks_from_single_hash( common_algo, content[common_algo]) found_pks = [pk._asdict() for pk in found_pks] # Filter with the other hash algorithms. for algo in filter_algos[1:]: found_pks = [pk for pk in found_pks if pk[algo] == content[algo]] results = [] for pk in found_pks: # Query the main table ('content'). res = self._cql_runner.content_get_from_pk(pk) # Rows in 'content' are inserted after corresponding # rows in 'content_by_*', so we might be missing it if res: results.append({ **res._asdict(), 'ctime': res.ctime.replace(tzinfo=datetime.timezone.utc) }) return results def content_missing(self, content, key_hash='sha1'): for cont in content: res = self.content_find(cont) if not res: yield cont[key_hash] if any(c['status'] == 'missing' for c in res): yield cont[key_hash] def content_missing_per_sha1(self, contents): return self.content_missing([{'sha1': c for c in contents}]) def content_missing_per_sha1_git(self, contents): return self.content_missing([{'sha1_git': c for c in contents}], key_hash='sha1_git') def content_get_random(self): return self._cql_runner.content_get_random().sha1_git + def _skipped_content_add(self, contents): + contents = [SkippedContent.from_dict(c) for c in contents] + + # Filter-out content already in the database. + contents = [ + c for c in contents + if not self._cql_runner.skipped_content_get_from_pk(c.to_dict())] + + if self.journal_writer: + for content in contents: + content = content.to_dict() + if 'data' in content: + del content['data'] + self.journal_writer.write_addition('content', content) + + for content in contents: + # Add to index tables + for algo in HASH_ALGORITHMS: + if content.get_hash(algo) is not None: + self._cql_runner.skipped_content_index_add_one( + algo, content) + + # Then to the main table + self._cql_runner.skipped_content_add_one(content) + + return { + 'skipped_content:add': len(contents) + } + + def skipped_content_add(self, content): + content = [c.copy() for c in content] # semi-shallow copy + for item in content: + item['ctime'] = now() + return self._skipped_content_add(content) + + def skipped_content_missing(self, contents): + for content in contents: + if not self._cql_runner.skipped_content_get_from_pk(content): + yield content + def directory_add(self, directories): directories = list(directories) # Filter out directories that are already inserted. missing = self.directory_missing([dir_['id'] for dir_ in directories]) directories = [dir_ for dir_ in directories if dir_['id'] in missing] if self.journal_writer: self.journal_writer.write_additions('directory', directories) for directory in directories: directory = Directory.from_dict(directory) # Add directory entries to the 'directory_entry' table for entry in directory.entries: entry = entry.to_dict() entry['directory_id'] = directory.id self._cql_runner.directory_entry_add_one(entry) # Add the directory *after* adding all the entries, so someone # calling snapshot_get_branch in the meantime won't end up # with half the entries. self._cql_runner.directory_add_one(directory.id) return {'directory:add': len(missing)} def directory_missing(self, directories): return self._cql_runner.directory_missing(directories) def _join_dentry_to_content(self, dentry): keys = ( 'status', 'sha1', 'sha1_git', 'sha256', 'length', ) ret = dict.fromkeys(keys) ret.update(dentry.to_dict()) if ret['type'] == 'file': content = self.content_find({'sha1_git': ret['target']}) if content: content = content[0] for key in keys: ret[key] = content[key] return ret def _directory_ls(self, directory_id, recursive, prefix=b''): if self.directory_missing([directory_id]): return rows = list(self._cql_runner.directory_entry_get([directory_id])) for row in rows: # Build and yield the directory entry dict entry = row._asdict() del entry['directory_id'] entry = DirectoryEntry.from_dict(entry) ret = self._join_dentry_to_content(entry) ret['name'] = prefix + ret['name'] ret['dir_id'] = directory_id yield ret if recursive and ret['type'] == 'dir': yield from self._directory_ls( ret['target'], True, prefix + ret['name'] + b'/') def directory_entry_get_by_path(self, directory, paths): return self._directory_entry_get_by_path(directory, paths, b'') def _directory_entry_get_by_path(self, directory, paths, prefix): if not paths: return contents = list(self.directory_ls(directory)) if not contents: return def _get_entry(entries, name): """Finds the entry with the requested name, prepends the prefix (to get its full path), and returns it. If no entry has that name, returns None.""" for entry in entries: if entry['name'] == name: entry = entry.copy() entry['name'] = prefix + entry['name'] return entry first_item = _get_entry(contents, paths[0]) if len(paths) == 1: return first_item if not first_item or first_item['type'] != 'dir': return return self._directory_entry_get_by_path( first_item['target'], paths[1:], prefix + paths[0] + b'/') def directory_ls(self, directory, recursive=False): yield from self._directory_ls(directory, recursive) def directory_get_random(self): return self._cql_runner.directory_get_random().id def revision_add(self, revisions): revisions = list(revisions) # Filter-out revisions already in the database missing = self.revision_missing([rev['id'] for rev in revisions]) revisions = [rev for rev in revisions if rev['id'] in missing] if self.journal_writer: self.journal_writer.write_additions('revision', revisions) for revision in revisions: revision = revision_to_db(revision) if revision: # Add parents first for (rank, parent) in enumerate(revision.parents): self._cql_runner.revision_parent_add_one( revision.id, rank, parent) # Then write the main revision row. # Writing this after all parents were written ensures that # read endpoints don't return a partial view while writing # the parents self._cql_runner.revision_add_one(revision) return {'revision:add': len(revisions)} def revision_missing(self, revisions): return self._cql_runner.revision_missing(revisions) def revision_get(self, revisions): rows = self._cql_runner.revision_get(revisions) revs = {} for row in rows: # TODO: use a single query to get all parents? # (it might have lower latency, but requires more code and more # bandwidth, because revision id would be part of each returned # row) parent_rows = self._cql_runner.revision_parent_get(row.id) # parent_rank is the clustering key, so results are already # sorted by rank. parents = [row.parent_id for row in parent_rows] rev = Revision(**row._asdict(), parents=parents) rev = revision_from_db(rev) revs[rev.id] = rev.to_dict() for rev_id in revisions: yield revs.get(rev_id) def _get_parent_revs(self, rev_ids, seen, limit, short): if limit and len(seen) >= limit: return rev_ids = [id_ for id_ in rev_ids if id_ not in seen] if not rev_ids: return seen |= set(rev_ids) # We need this query, even if short=True, to return consistent # results (ie. not return only a subset of a revision's parents # if it is being written) if short: rows = self._cql_runner.revision_get_ids(rev_ids) else: rows = self._cql_runner.revision_get(rev_ids) for row in rows: # TODO: use a single query to get all parents? # (it might have less latency, but requires less code and more # bandwidth (because revision id would be part of each returned # row) parent_rows = self._cql_runner.revision_parent_get(row.id) # parent_rank is the clustering key, so results are already # sorted by rank. parents = [row.parent_id for row in parent_rows] if short: yield (row.id, parents) else: rev = revision_from_db(Revision( **row._asdict(), parents=parents)) yield rev.to_dict() yield from self._get_parent_revs(parents, seen, limit, short) def revision_log(self, revisions, limit=None): seen = set() yield from self._get_parent_revs(revisions, seen, limit, False) def revision_shortlog(self, revisions, limit=None): seen = set() yield from self._get_parent_revs(revisions, seen, limit, True) def revision_get_random(self): return self._cql_runner.revision_get_random().id def release_add(self, releases): releases = list(releases) missing = self.release_missing([rel['id'] for rel in releases]) releases = [rel for rel in releases if rel['id'] in missing] if self.journal_writer: self.journal_writer.write_additions('release', releases) for release in releases: release = release_to_db(release) if release: self._cql_runner.release_add_one(release) return {'release:add': len(missing)} def release_missing(self, releases): return self._cql_runner.release_missing(releases) def release_get(self, releases): rows = self._cql_runner.release_get(releases) rels = {} for row in rows: release = Release(**row._asdict()) release = release_from_db(release) rels[row.id] = release.to_dict() for rel_id in releases: yield rels.get(rel_id) def release_get_random(self): return self._cql_runner.release_get_random().id def snapshot_add(self, snapshots): snapshots = list(snapshots) missing = self._cql_runner.snapshot_missing( [snp['id'] for snp in snapshots]) snapshots = [snp for snp in snapshots if snp['id'] in missing] for snapshot in snapshots: if self.journal_writer: self.journal_writer.write_addition('snapshot', snapshot) # Add branches for (branch_name, branch) in snapshot['branches'].items(): if branch is None: branch = {'target_type': None, 'target': None} self._cql_runner.snapshot_branch_add_one({ 'snapshot_id': snapshot['id'], 'name': branch_name, 'target_type': branch['target_type'], 'target': branch['target'], }) # Add the snapshot *after* adding all the branches, so someone # calling snapshot_get_branch in the meantime won't end up # with half the branches. self._cql_runner.snapshot_add_one(snapshot['id']) return {'snapshot:add': len(snapshots)} def snapshot_missing(self, snapshots): return self._cql_runner.snapshot_missing(snapshots) def snapshot_get(self, snapshot_id): return self.snapshot_get_branches(snapshot_id) def snapshot_get_by_origin_visit(self, origin, visit): try: visit = self._cql_runner.origin_visit_get_one(origin, visit) except IndexError: return None return self.snapshot_get(visit.snapshot) def snapshot_get_latest(self, origin, allowed_statuses=None): visit = self.origin_visit_get_latest( origin, allowed_statuses=allowed_statuses, require_snapshot=True) if visit: assert visit['snapshot'] if self._cql_runner.snapshot_missing([visit['snapshot']]): raise ValueError('Visit references unknown snapshot') return self.snapshot_get_branches(visit['snapshot']) def snapshot_count_branches(self, snapshot_id): if self._cql_runner.snapshot_missing([snapshot_id]): # Makes sure we don't fetch branches for a snapshot that is # being added. return None rows = list(self._cql_runner.snapshot_count_branches(snapshot_id)) assert len(rows) == 1 (nb_none, counts) = rows[0].counts counts = dict(counts) if nb_none: counts[None] = nb_none return counts def snapshot_get_branches(self, snapshot_id, branches_from=b'', branches_count=1000, target_types=None): if self._cql_runner.snapshot_missing([snapshot_id]): # Makes sure we don't fetch branches for a snapshot that is # being added. return None branches = [] while len(branches) < branches_count+1: new_branches = list(self._cql_runner.snapshot_branch_get( snapshot_id, branches_from, branches_count+1)) if not new_branches: break branches_from = new_branches[-1].name new_branches_filtered = new_branches # Filter by target_type if target_types: new_branches_filtered = [ branch for branch in new_branches_filtered if branch.target is not None and branch.target_type in target_types] branches.extend(new_branches_filtered) if len(new_branches) < branches_count+1: break if len(branches) > branches_count: last_branch = branches.pop(-1).name else: last_branch = None branches = { branch.name: { 'target': branch.target, 'target_type': branch.target_type, } if branch.target else None for branch in branches } return { 'id': snapshot_id, 'branches': branches, 'next_branch': last_branch, } def snapshot_get_random(self): return self._cql_runner.snapshot_get_random().id def object_find_by_sha1_git(self, ids): results = {id_: [] for id_ in ids} missing_ids = set(ids) # Mind the order, revision is the most likely one for a given ID, # so we check revisions first. queries = [ ('revision', self._cql_runner.revision_missing), ('release', self._cql_runner.release_missing), ('content', self._cql_runner.content_missing_by_sha1_git), ('directory', self._cql_runner.directory_missing), ] for (object_type, query_fn) in queries: found_ids = missing_ids - set(query_fn(missing_ids)) for sha1_git in found_ids: results[sha1_git].append({ 'sha1_git': sha1_git, 'type': object_type, }) missing_ids.remove(sha1_git) if not missing_ids: # We found everything, skipping the next queries. break return results def origin_get(self, origins): if isinstance(origins, dict): # Old API return_single = True origins = [origins] else: return_single = False if any('id' in origin for origin in origins): raise ValueError('Origin ids are not supported.') results = [self.origin_get_one(origin) for origin in origins] if return_single: assert len(results) == 1 return results[0] else: return results def origin_get_one(self, origin): if 'id' in origin: raise ValueError('Origin ids are not supported.') rows = self._cql_runner.origin_get_by_url(origin['url']) rows = list(rows) if rows: assert len(rows) == 1 result = rows[0]._asdict() return { 'url': result['url'], } else: return None def origin_get_by_sha1(self, sha1s): results = [] for sha1 in sha1s: rows = self._cql_runner.origin_get_by_sha1(sha1) if rows: results.append({'url': rows.one().url}) else: results.append(None) return results def origin_list(self, page_token: Optional[str] = None, limit: int = 100 ) -> dict: # Compute what token to begin the listing from start_token = TOKEN_BEGIN if page_token: start_token = int(page_token) if not (TOKEN_BEGIN <= start_token <= TOKEN_END): raise ValueError('Invalid page_token.') rows = self._cql_runner.origin_list(start_token, limit) rows = list(rows) if len(rows) == limit: next_page_token: Optional[str] = str(rows[-1].tok+1) else: next_page_token = None return { 'origins': [{'url': row.url} for row in rows], 'next_page_token': next_page_token, } def origin_search(self, url_pattern, offset=0, limit=50, regexp=False, with_visit=False): # TODO: remove this endpoint, swh-search should be used instead. origins = self._cql_runner.origin_iter_all() if regexp: pat = re.compile(url_pattern) origins = [orig for orig in origins if pat.search(orig.url)] else: origins = [orig for orig in origins if url_pattern in orig.url] if with_visit: origins = [orig for orig in origins if orig.next_visit_id > 1] return [ { 'url': orig.url, } for orig in origins[offset:offset+limit]] def origin_add(self, origins): origins = list(origins) if any('id' in origin for origin in origins): raise ValueError('Origins must not already have an id.') results = [] for origin in origins: self.origin_add_one(origin) results.append(origin) return results def origin_add_one(self, origin): known_origin = self.origin_get_one(origin) if known_origin: origin_url = known_origin['url'] else: if self.journal_writer: self.journal_writer.write_addition('origin', origin) self._cql_runner.origin_add_one(origin) origin_url = origin['url'] return origin_url def origin_visit_add(self, origin, date, type): origin_url = origin # TODO: rename the argument if isinstance(date, str): date = dateutil.parser.parse(date) origin = self.origin_get_one({'url': origin_url}) if not origin: return None visit_id = self._cql_runner.origin_generate_unique_visit_id(origin_url) visit = { 'origin': origin_url, 'date': date, 'type': type, 'status': 'ongoing', 'snapshot': None, 'metadata': None, 'visit': visit_id } if self.journal_writer: self.journal_writer.write_addition('origin_visit', visit) self._cql_runner.origin_visit_add_one(visit) return { 'origin': origin_url, 'visit': visit_id, } def origin_visit_update(self, origin, visit_id, status=None, metadata=None, snapshot=None): origin_url = origin # TODO: rename the argument # Get the existing data of the visit row = self._cql_runner.origin_visit_get_one(origin_url, visit_id) if not row: raise ValueError('This origin visit does not exist.') visit = OriginVisit.from_dict(self._format_origin_visit_row(row)) updates = {} if status: updates['status'] = status if metadata: updates['metadata'] = metadata if snapshot: updates['snapshot'] = snapshot visit = attr.evolve(visit, **updates) if self.journal_writer: self.journal_writer.write_update('origin_visit', visit) self._cql_runner.origin_visit_update(origin_url, visit_id, updates) def origin_visit_upsert(self, visits): visits = [visit.copy() for visit in visits] for visit in visits: if isinstance(visit['date'], str): visit['date'] = dateutil.parser.parse(visit['date']) if self.journal_writer: for visit in visits: self.journal_writer.write_addition('origin_visit', visit) for visit in visits: visit = visit.copy() if visit.get('metadata'): visit['metadata'] = json.dumps(visit['metadata']) self._cql_runner.origin_visit_upsert(visit) @staticmethod def _format_origin_visit_row(visit): return { **visit._asdict(), 'origin': visit.origin, 'date': visit.date.replace(tzinfo=datetime.timezone.utc), 'metadata': (json.loads(visit.metadata) if visit.metadata else None), } def origin_visit_get(self, origin, last_visit=None, limit=None): rows = self._cql_runner.origin_visit_get(origin, last_visit, limit) yield from map(self._format_origin_visit_row, rows) def origin_visit_find_by_date(self, origin, visit_date): # Iterator over all the visits of the origin # This should be ok for now, as there aren't too many visits # per origin. visits = list(self._cql_runner.origin_visit_get_all(origin)) def key(visit): dt = visit.date.replace(tzinfo=datetime.timezone.utc) - visit_date return (abs(dt), -visit.visit) if visits: visit = min(visits, key=key) return visit._asdict() def origin_visit_get_by(self, origin, visit): visit = self._cql_runner.origin_visit_get_one(origin, visit) if visit: return self._format_origin_visit_row(visit) def origin_visit_get_latest( self, origin, allowed_statuses=None, require_snapshot=False): visit = self._cql_runner.origin_visit_get_latest( origin, allowed_statuses=allowed_statuses, require_snapshot=require_snapshot) if visit: return self._format_origin_visit_row(visit) def origin_visit_get_random(self, type: str) -> Optional[Dict[str, Any]]: back_in_the_day = now() - datetime.timedelta(weeks=12) # 3 months back # Random position to start iteration at start_token = random.randint(TOKEN_BEGIN, TOKEN_END) # Iterator over all visits, ordered by token(origins) then visit_id rows = self._cql_runner.origin_visit_iter(start_token) for row in rows: visit = self._format_origin_visit_row(row) if visit['date'] > back_in_the_day \ and visit['status'] == 'full': return visit else: return None def tool_add(self, tools): inserted = [] for tool in tools: tool = tool.copy() tool_json = tool.copy() tool_json['configuration'] = json.dumps( tool['configuration'], sort_keys=True).encode() id_ = self._cql_runner.tool_get_one_uuid(**tool_json) if not id_: id_ = uuid.uuid1() tool_json['id'] = id_ self._cql_runner.tool_by_uuid_add_one(tool_json) self._cql_runner.tool_add_one(tool_json) tool['id'] = id_ inserted.append(tool) return inserted def tool_get(self, tool): id_ = self._cql_runner.tool_get_one_uuid( tool['name'], tool['version'], json.dumps(tool['configuration'], sort_keys=True).encode()) if id_: tool = tool.copy() tool['id'] = id_ return tool else: return None def stat_counters(self): rows = self._cql_runner.stat_counters() keys = ( 'content', 'directory', 'origin', 'origin_visit', 'release', 'revision', 'skipped_content', 'snapshot') stats = {key: 0 for key in keys} stats.update({row.object_type: row.count for row in rows}) return stats def refresh_stat_counters(self): pass def origin_metadata_add(self, origin_url, ts, provider, tool, metadata): # TODO raise NotImplementedError('not yet supported for Cassandra') def origin_metadata_get_by(self, origin_url, provider_type=None): # TODO raise NotImplementedError('not yet supported for Cassandra') def metadata_provider_add(self, provider_name, provider_type, provider_url, metadata): # TODO raise NotImplementedError('not yet supported for Cassandra') def metadata_provider_get(self, provider_id): # TODO raise NotImplementedError('not yet supported for Cassandra') def metadata_provider_get_by(self, provider): # TODO raise NotImplementedError('not yet supported for Cassandra') diff --git a/swh/storage/db.py b/swh/storage/db.py index 3cc8cc06..770cb29c 100644 --- a/swh/storage/db.py +++ b/swh/storage/db.py @@ -1,969 +1,969 @@ # Copyright (C) 2015-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import random import select from swh.core.db import BaseDb from swh.core.db.db_utils import stored_procedure, jsonize from swh.core.db.db_utils import execute_values_generator from swh.model.model import SHA1_SIZE class Db(BaseDb): """Proxy to the SWH DB, with wrappers around stored procedures """ def mktemp_dir_entry(self, entry_type, cur=None): self._cursor(cur).execute('SELECT swh_mktemp_dir_entry(%s)', (('directory_entry_%s' % entry_type),)) @stored_procedure('swh_mktemp_revision') def mktemp_revision(self, cur=None): pass @stored_procedure('swh_mktemp_release') def mktemp_release(self, cur=None): pass @stored_procedure('swh_mktemp_snapshot_branch') def mktemp_snapshot_branch(self, cur=None): pass def register_listener(self, notify_queue, cur=None): """Register a listener for NOTIFY queue `notify_queue`""" self._cursor(cur).execute("LISTEN %s" % notify_queue) def listen_notifies(self, timeout): """Listen to notifications for `timeout` seconds""" if select.select([self.conn], [], [], timeout) == ([], [], []): return else: self.conn.poll() while self.conn.notifies: yield self.conn.notifies.pop(0) @stored_procedure('swh_content_add') def content_add_from_temp(self, cur=None): pass @stored_procedure('swh_directory_add') def directory_add_from_temp(self, cur=None): pass @stored_procedure('swh_skipped_content_add') def skipped_content_add_from_temp(self, cur=None): pass @stored_procedure('swh_revision_add') def revision_add_from_temp(self, cur=None): pass @stored_procedure('swh_release_add') def release_add_from_temp(self, cur=None): pass def content_update_from_temp(self, keys_to_update, cur=None): cur = self._cursor(cur) cur.execute("""select swh_content_update(ARRAY[%s] :: text[])""" % keys_to_update) content_get_metadata_keys = [ 'sha1', 'sha1_git', 'sha256', 'blake2s256', 'length', 'status'] content_add_keys = content_get_metadata_keys + ['ctime'] skipped_content_keys = [ 'sha1', 'sha1_git', 'sha256', 'blake2s256', 'length', 'reason', 'status', 'origin'] def content_get_metadata_from_sha1s(self, sha1s, cur=None): cur = self._cursor(cur) yield from execute_values_generator( cur, """ select t.sha1, %s from (values %%s) as t (sha1) inner join content using (sha1) """ % ', '.join(self.content_get_metadata_keys[1:]), ((sha1,) for sha1 in sha1s), ) def content_get_range(self, start, end, limit=None, cur=None): """Retrieve contents within range [start, end]. """ cur = self._cursor(cur) query = """select %s from content where %%s <= sha1 and sha1 <= %%s order by sha1 limit %%s""" % ', '.join(self.content_get_metadata_keys) cur.execute(query, (start, end, limit)) yield from cur content_hash_keys = ['sha1', 'sha1_git', 'sha256', 'blake2s256'] def content_missing_from_list(self, contents, cur=None): cur = self._cursor(cur) keys = ', '.join(self.content_hash_keys) equality = ' AND '.join( ('t.%s = c.%s' % (key, key)) for key in self.content_hash_keys ) yield from execute_values_generator( cur, """ SELECT %s FROM (VALUES %%s) as t(%s) WHERE NOT EXISTS ( SELECT 1 FROM content c WHERE %s ) """ % (keys, keys, equality), (tuple(c[key] for key in self.content_hash_keys) for c in contents) ) def content_missing_per_sha1(self, sha1s, cur=None): cur = self._cursor(cur) yield from execute_values_generator(cur, """ SELECT t.sha1 FROM (VALUES %s) AS t(sha1) WHERE NOT EXISTS ( SELECT 1 FROM content c WHERE c.sha1 = t.sha1 )""", ((sha1,) for sha1 in sha1s)) def content_missing_per_sha1_git(self, contents, cur=None): cur = self._cursor(cur) yield from execute_values_generator(cur, """ SELECT t.sha1_git FROM (VALUES %s) AS t(sha1_git) WHERE NOT EXISTS ( SELECT 1 FROM content c WHERE c.sha1_git = t.sha1_git )""", ((sha1,) for sha1 in contents)) def skipped_content_missing(self, contents, cur=None): if not contents: return [] cur = self._cursor(cur) query = """SELECT * FROM (VALUES %s) AS t (%s) WHERE not exists (SELECT 1 FROM skipped_content s WHERE - s.sha1 is not distinct from t.sha1 and - s.sha1_git is not distinct from t.sha1_git and - s.sha256 is not distinct from t.sha256);""" % \ + s.sha1 is not distinct from t.sha1::sha1 and + s.sha1_git is not distinct from t.sha1_git::sha1 and + s.sha256 is not distinct from t.sha256::bytea);""" % \ ((', '.join('%s' for _ in contents)), ', '.join(self.content_hash_keys)) cur.execute(query, [tuple(cont[key] for key in self.content_hash_keys) for cont in contents]) yield from cur def snapshot_exists(self, snapshot_id, cur=None): """Check whether a snapshot with the given id exists""" cur = self._cursor(cur) cur.execute("""SELECT 1 FROM snapshot where id=%s""", (snapshot_id,)) return bool(cur.fetchone()) def snapshot_missing_from_list(self, snapshots, cur=None): cur = self._cursor(cur) yield from execute_values_generator( cur, """ SELECT id FROM (VALUES %s) as t(id) WHERE NOT EXISTS ( SELECT 1 FROM snapshot d WHERE d.id = t.id ) """, ((id,) for id in snapshots)) def snapshot_add(self, snapshot_id, cur=None): """Add a snapshot from the temporary table""" cur = self._cursor(cur) cur.execute("""SELECT swh_snapshot_add(%s)""", (snapshot_id,)) snapshot_count_cols = ['target_type', 'count'] def snapshot_count_branches(self, snapshot_id, cur=None): cur = self._cursor(cur) query = """\ SELECT %s FROM swh_snapshot_count_branches(%%s) """ % ', '.join(self.snapshot_count_cols) cur.execute(query, (snapshot_id,)) yield from cur snapshot_get_cols = ['snapshot_id', 'name', 'target', 'target_type'] def snapshot_get_by_id(self, snapshot_id, branches_from=b'', branches_count=None, target_types=None, cur=None): cur = self._cursor(cur) query = """\ SELECT %s FROM swh_snapshot_get_by_id(%%s, %%s, %%s, %%s :: snapshot_target[]) """ % ', '.join(self.snapshot_get_cols) cur.execute(query, (snapshot_id, branches_from, branches_count, target_types)) yield from cur def snapshot_get_by_origin_visit(self, origin_url, visit_id, cur=None): cur = self._cursor(cur) query = """\ SELECT snapshot FROM origin_visit INNER JOIN origin ON origin.id = origin_visit.origin WHERE origin.url=%s AND origin_visit.visit=%s; """ cur.execute(query, (origin_url, visit_id)) ret = cur.fetchone() if ret: return ret[0] def snapshot_get_random(self, cur=None): return self._get_random_row_from_table( 'snapshot', ['id'], 'id', cur) content_find_cols = ['sha1', 'sha1_git', 'sha256', 'blake2s256', 'length', 'ctime', 'status'] def content_find(self, sha1=None, sha1_git=None, sha256=None, blake2s256=None, cur=None): """Find the content optionally on a combination of the following checksums sha1, sha1_git, sha256 or blake2s256. Args: sha1: sha1 content git_sha1: the sha1 computed `a la git` sha1 of the content sha256: sha256 content blake2s256: blake2s256 content Returns: The tuple (sha1, sha1_git, sha256, blake2s256) if found or None. """ cur = self._cursor(cur) checksum_dict = {'sha1': sha1, 'sha1_git': sha1_git, 'sha256': sha256, 'blake2s256': blake2s256} where_parts = [] args = [] # Adds only those keys which have value other than None for algorithm in checksum_dict: if checksum_dict[algorithm] is not None: args.append(checksum_dict[algorithm]) where_parts.append(algorithm + '= %s') query = ' AND '.join(where_parts) cur.execute("""SELECT %s FROM content WHERE %s """ % (','.join(self.content_find_cols), query), args) content = cur.fetchall() return content def content_get_random(self, cur=None): return self._get_random_row_from_table( 'content', ['sha1_git'], 'sha1_git', cur) def directory_missing_from_list(self, directories, cur=None): cur = self._cursor(cur) yield from execute_values_generator( cur, """ SELECT id FROM (VALUES %s) as t(id) WHERE NOT EXISTS ( SELECT 1 FROM directory d WHERE d.id = t.id ) """, ((id,) for id in directories)) directory_ls_cols = ['dir_id', 'type', 'target', 'name', 'perms', 'status', 'sha1', 'sha1_git', 'sha256', 'length'] def directory_walk_one(self, directory, cur=None): cur = self._cursor(cur) cols = ', '.join(self.directory_ls_cols) query = 'SELECT %s FROM swh_directory_walk_one(%%s)' % cols cur.execute(query, (directory,)) yield from cur def directory_walk(self, directory, cur=None): cur = self._cursor(cur) cols = ', '.join(self.directory_ls_cols) query = 'SELECT %s FROM swh_directory_walk(%%s)' % cols cur.execute(query, (directory,)) yield from cur def directory_entry_get_by_path(self, directory, paths, cur=None): """Retrieve a directory entry by path. """ cur = self._cursor(cur) cols = ', '.join(self.directory_ls_cols) query = ( 'SELECT %s FROM swh_find_directory_entry_by_path(%%s, %%s)' % cols) cur.execute(query, (directory, paths)) data = cur.fetchone() if set(data) == {None}: return None return data def directory_get_random(self, cur=None): return self._get_random_row_from_table( 'directory', ['id'], 'id', cur) def revision_missing_from_list(self, revisions, cur=None): cur = self._cursor(cur) yield from execute_values_generator( cur, """ SELECT id FROM (VALUES %s) as t(id) WHERE NOT EXISTS ( SELECT 1 FROM revision r WHERE r.id = t.id ) """, ((id,) for id in revisions)) revision_add_cols = [ 'id', 'date', 'date_offset', 'date_neg_utc_offset', 'committer_date', 'committer_date_offset', 'committer_date_neg_utc_offset', 'type', 'directory', 'message', 'author_fullname', 'author_name', 'author_email', 'committer_fullname', 'committer_name', 'committer_email', 'metadata', 'synthetic', ] revision_get_cols = revision_add_cols + ['parents'] def origin_visit_add(self, origin, ts, type, cur=None): """Add a new origin_visit for origin origin at timestamp ts with status 'ongoing'. Args: origin: origin concerned by the visit ts: the date of the visit type: type of loader for the visit Returns: The new visit index step for that origin """ cur = self._cursor(cur) self._cursor(cur).execute('SELECT swh_origin_visit_add(%s, %s, %s)', (origin, ts, type)) return cur.fetchone()[0] def origin_visit_update(self, origin_id, visit_id, updates, cur=None): """Update origin_visit's status.""" cur = self._cursor(cur) update_cols = [] values = [] where = ['origin.id = origin_visit.origin', 'origin.url=%s', 'visit=%s'] where_values = [origin_id, visit_id] if 'status' in updates: update_cols.append('status=%s') values.append(updates.pop('status')) if 'metadata' in updates: update_cols.append('metadata=%s') values.append(jsonize(updates.pop('metadata'))) if 'snapshot' in updates: update_cols.append('snapshot=%s') values.append(updates.pop('snapshot')) assert not updates, 'Unknown fields: %r' % updates query = """UPDATE origin_visit SET {update_cols} FROM origin WHERE {where}""".format(**{ 'update_cols': ', '.join(update_cols), 'where': ' AND '.join(where) }) cur.execute(query, (*values, *where_values)) def origin_visit_upsert(self, origin, visit, date, type, status, metadata, snapshot, cur=None): # doing an extra query like this is way simpler than trying to join # the origin id in the query below origin_id = next(self.origin_id_get_by_url([origin])) cur = self._cursor(cur) query = """INSERT INTO origin_visit ({cols}) VALUES ({values}) ON CONFLICT ON CONSTRAINT origin_visit_pkey DO UPDATE SET {updates}""".format( cols=', '.join(self.origin_visit_get_cols), values=', '.join('%s' for col in self.origin_visit_get_cols), updates=', '.join('{0}=excluded.{0}'.format(col) for col in self.origin_visit_get_cols)) cur.execute( query, (origin_id, visit, date, type, status, metadata, snapshot)) origin_visit_get_cols = [ 'origin', 'visit', 'date', 'type', 'status', 'metadata', 'snapshot'] origin_visit_select_cols = [ 'origin.url AS origin', 'visit', 'date', 'origin_visit.type AS type', 'status', 'metadata', 'snapshot'] def origin_visit_get_all(self, origin_id, last_visit=None, limit=None, cur=None): """Retrieve all visits for origin with id origin_id. Args: origin_id: The occurrence's origin Yields: The occurrence's history visits """ cur = self._cursor(cur) if last_visit: extra_condition = 'and visit > %s' args = (origin_id, last_visit, limit) else: extra_condition = '' args = (origin_id, limit) query = """\ SELECT %s FROM origin_visit INNER JOIN origin ON origin.id = origin_visit.origin WHERE origin.url=%%s %s order by visit asc limit %%s""" % ( ', '.join(self.origin_visit_select_cols), extra_condition ) cur.execute(query, args) yield from cur def origin_visit_get(self, origin_id, visit_id, cur=None): """Retrieve information on visit visit_id of origin origin_id. Args: origin_id: the origin concerned visit_id: The visit step for that origin Returns: The origin_visit information """ cur = self._cursor(cur) query = """\ SELECT %s FROM origin_visit INNER JOIN origin ON origin.id = origin_visit.origin WHERE origin.url = %%s AND visit = %%s """ % (', '.join(self.origin_visit_select_cols)) cur.execute(query, (origin_id, visit_id)) r = cur.fetchall() if not r: return None return r[0] def origin_visit_find_by_date(self, origin, visit_date, cur=None): cur = self._cursor(cur) cur.execute( 'SELECT * FROM swh_visit_find_by_date(%s, %s)', (origin, visit_date)) r = cur.fetchall() if r: return r[0] def origin_visit_exists(self, origin_id, visit_id, cur=None): """Check whether an origin visit with the given ids exists""" cur = self._cursor(cur) query = "SELECT 1 FROM origin_visit where origin = %s AND visit = %s" cur.execute(query, (origin_id, visit_id)) return bool(cur.fetchone()) def origin_visit_get_latest( self, origin_id, allowed_statuses=None, require_snapshot=False, cur=None): """Retrieve the most recent origin_visit of the given origin, with optional filters. Args: origin_id: the origin concerned allowed_statuses: the visit statuses allowed for the returned visit require_snapshot (bool): If True, only a visit with a known snapshot will be returned. Returns: The origin_visit information, or None if no visit matches. """ cur = self._cursor(cur) query_parts = [ 'SELECT %s' % ', '.join(self.origin_visit_select_cols), 'FROM origin_visit', 'INNER JOIN origin ON origin.id = origin_visit.origin'] query_parts.append('WHERE origin.url = %s') if require_snapshot: query_parts.append('AND snapshot is not null') if allowed_statuses: query_parts.append( cur.mogrify('AND status IN %s', (tuple(allowed_statuses),)).decode()) query_parts.append('ORDER BY date DESC, visit DESC LIMIT 1') query = '\n'.join(query_parts) cur.execute(query, (origin_id,)) r = cur.fetchone() if not r: return None return r def origin_visit_get_random(self, type, cur=None): """Randomly select one origin visit that was full and in the last 3 months """ cur = self._cursor(cur) columns = ','.join(self.origin_visit_select_cols) query = f"""with visits as ( select * from origin_visit where origin_visit.status='full' and origin_visit.type=%s and origin_visit.date > now() - '3 months'::interval ) select {columns} from visits as origin_visit inner join origin on origin_visit.origin=origin.id where random() < 0.1 limit 1 """ cur.execute(query, (type, )) return cur.fetchone() @staticmethod def mangle_query_key(key, main_table): if key == 'id': return 't.id' if key == 'parents': return ''' ARRAY( SELECT rh.parent_id::bytea FROM revision_history rh WHERE rh.id = t.id ORDER BY rh.parent_rank )''' if '_' not in key: return '%s.%s' % (main_table, key) head, tail = key.split('_', 1) if (head in ('author', 'committer') and tail in ('name', 'email', 'id', 'fullname')): return '%s.%s' % (head, tail) return '%s.%s' % (main_table, key) def revision_get_from_list(self, revisions, cur=None): cur = self._cursor(cur) query_keys = ', '.join( self.mangle_query_key(k, 'revision') for k in self.revision_get_cols ) yield from execute_values_generator( cur, """ SELECT %s FROM (VALUES %%s) as t(id) LEFT JOIN revision ON t.id = revision.id LEFT JOIN person author ON revision.author = author.id LEFT JOIN person committer ON revision.committer = committer.id """ % query_keys, ((id,) for id in revisions)) def revision_log(self, root_revisions, limit=None, cur=None): cur = self._cursor(cur) query = """SELECT %s FROM swh_revision_log(%%s, %%s) """ % ', '.join(self.revision_get_cols) cur.execute(query, (root_revisions, limit)) yield from cur revision_shortlog_cols = ['id', 'parents'] def revision_shortlog(self, root_revisions, limit=None, cur=None): cur = self._cursor(cur) query = """SELECT %s FROM swh_revision_list(%%s, %%s) """ % ', '.join(self.revision_shortlog_cols) cur.execute(query, (root_revisions, limit)) yield from cur def revision_get_random(self, cur=None): return self._get_random_row_from_table( 'revision', ['id'], 'id', cur) def release_missing_from_list(self, releases, cur=None): cur = self._cursor(cur) yield from execute_values_generator( cur, """ SELECT id FROM (VALUES %s) as t(id) WHERE NOT EXISTS ( SELECT 1 FROM release r WHERE r.id = t.id ) """, ((id,) for id in releases)) object_find_by_sha1_git_cols = ['sha1_git', 'type'] def object_find_by_sha1_git(self, ids, cur=None): cur = self._cursor(cur) yield from execute_values_generator( cur, """ WITH t (sha1_git) AS (VALUES %s), known_objects as (( select id as sha1_git, 'release'::object_type as type, object_id from release r where exists (select 1 from t where t.sha1_git = r.id) ) union all ( select id as sha1_git, 'revision'::object_type as type, object_id from revision r where exists (select 1 from t where t.sha1_git = r.id) ) union all ( select id as sha1_git, 'directory'::object_type as type, object_id from directory d where exists (select 1 from t where t.sha1_git = d.id) ) union all ( select sha1_git as sha1_git, 'content'::object_type as type, object_id from content c where exists (select 1 from t where t.sha1_git = c.sha1_git) )) select t.sha1_git as sha1_git, k.type from t left join known_objects k on t.sha1_git = k.sha1_git """, ((id,) for id in ids) ) def stat_counters(self, cur=None): cur = self._cursor(cur) cur.execute('SELECT * FROM swh_stat_counters()') yield from cur def origin_add(self, url, cur=None): """Insert a new origin and return the new identifier.""" insert = """INSERT INTO origin (url) values (%s) RETURNING url""" cur.execute(insert, (url,)) return cur.fetchone()[0] origin_cols = ['url'] def origin_get_by_url(self, origins, cur=None): """Retrieve origin `(type, url)` from urls if found.""" cur = self._cursor(cur) query = """SELECT %s FROM (VALUES %%s) as t(url) LEFT JOIN origin ON t.url = origin.url """ % ','.join('origin.' + col for col in self.origin_cols) yield from execute_values_generator( cur, query, ((url,) for url in origins)) def origin_get_by_sha1(self, sha1s, cur=None): """Retrieve origin urls from sha1s if found.""" cur = self._cursor(cur) query = """SELECT %s FROM (VALUES %%s) as t(sha1) LEFT JOIN origin ON t.sha1 = digest(origin.url, 'sha1') """ % ','.join('origin.' + col for col in self.origin_cols) yield from execute_values_generator( cur, query, ((sha1,) for sha1 in sha1s)) def origin_id_get_by_url(self, origins, cur=None): """Retrieve origin `(type, url)` from urls if found.""" cur = self._cursor(cur) query = """SELECT id FROM (VALUES %s) as t(url) LEFT JOIN origin ON t.url = origin.url """ for row in execute_values_generator( cur, query, ((url,) for url in origins)): yield row[0] origin_get_range_cols = ['id', 'url'] def origin_get_range(self, origin_from=1, origin_count=100, cur=None): """Retrieve ``origin_count`` origins whose ids are greater or equal than ``origin_from``. Origins are sorted by id before retrieving them. Args: origin_from (int): the minimum id of origins to retrieve origin_count (int): the maximum number of origins to retrieve """ cur = self._cursor(cur) query = """SELECT %s FROM origin WHERE id >= %%s ORDER BY id LIMIT %%s """ % ','.join(self.origin_get_range_cols) cur.execute(query, (origin_from, origin_count)) yield from cur def _origin_query(self, url_pattern, count=False, offset=0, limit=50, regexp=False, with_visit=False, cur=None): """ Method factorizing query creation for searching and counting origins. """ cur = self._cursor(cur) if count: origin_cols = 'COUNT(*)' else: origin_cols = ','.join(self.origin_cols) query = """SELECT %s FROM origin WHERE """ if with_visit: query += """ EXISTS ( SELECT 1 FROM origin_visit INNER JOIN snapshot ON snapshot=snapshot.id WHERE origin=origin.id ) AND """ query += 'url %s %%s ' if not count: query += 'ORDER BY id OFFSET %%s LIMIT %%s' if not regexp: query = query % (origin_cols, 'ILIKE') query_params = ('%'+url_pattern+'%', offset, limit) else: query = query % (origin_cols, '~*') query_params = (url_pattern, offset, limit) if count: query_params = (query_params[0],) cur.execute(query, query_params) def origin_search(self, url_pattern, offset=0, limit=50, regexp=False, with_visit=False, cur=None): """Search for origins whose urls contain a provided string pattern or match a provided regular expression. The search is performed in a case insensitive way. Args: url_pattern (str): the string pattern to search for in origin urls offset (int): number of found origins to skip before returning results limit (int): the maximum number of found origins to return regexp (bool): if True, consider the provided pattern as a regular expression and returns origins whose urls match it with_visit (bool): if True, filter out origins with no visit """ self._origin_query(url_pattern, offset=offset, limit=limit, regexp=regexp, with_visit=with_visit, cur=cur) yield from cur def origin_count(self, url_pattern, regexp=False, with_visit=False, cur=None): """Count origins whose urls contain a provided string pattern or match a provided regular expression. The pattern search in origin urls is performed in a case insensitive way. Args: url_pattern (str): the string pattern to search for in origin urls regexp (bool): if True, consider the provided pattern as a regular expression and returns origins whose urls match it with_visit (bool): if True, filter out origins with no visit """ self._origin_query(url_pattern, count=True, regexp=regexp, with_visit=with_visit, cur=cur) return cur.fetchone()[0] release_add_cols = [ 'id', 'target', 'target_type', 'date', 'date_offset', 'date_neg_utc_offset', 'name', 'comment', 'synthetic', 'author_fullname', 'author_name', 'author_email', ] release_get_cols = release_add_cols def release_get_from_list(self, releases, cur=None): cur = self._cursor(cur) query_keys = ', '.join( self.mangle_query_key(k, 'release') for k in self.release_get_cols ) yield from execute_values_generator( cur, """ SELECT %s FROM (VALUES %%s) as t(id) LEFT JOIN release ON t.id = release.id LEFT JOIN person author ON release.author = author.id """ % query_keys, ((id,) for id in releases)) def release_get_random(self, cur=None): return self._get_random_row_from_table( 'release', ['id'], 'id', cur) def origin_metadata_add(self, origin, ts, provider, tool, metadata, cur=None): """ Add an origin_metadata for the origin at ts with provider, tool and metadata. Args: origin (int): the origin's id for which the metadata is added ts (datetime): time when the metadata was found provider (int): the metadata provider identifier tool (int): the tool's identifier used to extract metadata metadata (jsonb): the metadata retrieved at the time and location Returns: id (int): the origin_metadata unique id """ cur = self._cursor(cur) insert = """INSERT INTO origin_metadata (origin_id, discovery_date, provider_id, tool_id, metadata) SELECT id, %s, %s, %s, %s FROM origin WHERE url = %s""" cur.execute(insert, (ts, provider, tool, jsonize(metadata), origin)) origin_metadata_get_cols = ['origin_url', 'discovery_date', 'tool_id', 'metadata', 'provider_id', 'provider_name', 'provider_type', 'provider_url'] def origin_metadata_get_by(self, origin_url, provider_type=None, cur=None): """Retrieve all origin_metadata entries for one origin_url """ cur = self._cursor(cur) if not provider_type: query = '''SELECT %s FROM swh_origin_metadata_get_by_origin( %%s)''' % (','.join( self.origin_metadata_get_cols)) cur.execute(query, (origin_url, )) else: query = '''SELECT %s FROM swh_origin_metadata_get_by_provider_type( %%s, %%s)''' % (','.join( self.origin_metadata_get_cols)) cur.execute(query, (origin_url, provider_type)) yield from cur tool_cols = ['id', 'name', 'version', 'configuration'] @stored_procedure('swh_mktemp_tool') def mktemp_tool(self, cur=None): pass def tool_add_from_temp(self, cur=None): cur = self._cursor(cur) cur.execute("SELECT %s from swh_tool_add()" % ( ','.join(self.tool_cols), )) yield from cur def tool_get(self, name, version, configuration, cur=None): cur = self._cursor(cur) cur.execute('''select %s from tool where name=%%s and version=%%s and configuration=%%s''' % ( ','.join(self.tool_cols)), (name, version, configuration)) return cur.fetchone() metadata_provider_cols = ['id', 'provider_name', 'provider_type', 'provider_url', 'metadata'] def metadata_provider_add(self, provider_name, provider_type, provider_url, metadata, cur=None): """Insert a new provider and return the new identifier.""" cur = self._cursor(cur) insert = """INSERT INTO metadata_provider (provider_name, provider_type, provider_url, metadata) values (%s, %s, %s, %s) RETURNING id""" cur.execute(insert, (provider_name, provider_type, provider_url, jsonize(metadata))) return cur.fetchone()[0] def metadata_provider_get(self, provider_id, cur=None): cur = self._cursor(cur) cur.execute('''select %s from metadata_provider where id=%%s ''' % ( ','.join(self.metadata_provider_cols)), (provider_id, )) return cur.fetchone() def metadata_provider_get_by(self, provider_name, provider_url, cur=None): cur = self._cursor(cur) cur.execute('''select %s from metadata_provider where provider_name=%%s and provider_url=%%s''' % ( ','.join(self.metadata_provider_cols)), (provider_name, provider_url)) return cur.fetchone() def _get_random_row_from_table(self, table_name, cols, id_col, cur=None): random_sha1 = bytes(random.randint(0, 255) for _ in range(SHA1_SIZE)) cur = self._cursor(cur) query = ''' (SELECT {cols} FROM {table} WHERE {id_col} >= %s ORDER BY {id_col} LIMIT 1) UNION (SELECT {cols} FROM {table} WHERE {id_col} < %s ORDER BY {id_col} DESC LIMIT 1) LIMIT 1 '''.format(cols=', '.join(cols), table=table_name, id_col=id_col) cur.execute(query, (random_sha1, random_sha1)) row = cur.fetchone() if row: return row[0] diff --git a/swh/storage/filter.py b/swh/storage/filter.py index 9531233d..15893103 100644 --- a/swh/storage/filter.py +++ b/swh/storage/filter.py @@ -1,115 +1,143 @@ # Copyright (C) 2019 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information -from typing import Dict, Generator, Sequence, Set +from typing import Dict, Generator, Iterable, Set from swh.storage import get_storage class FilteringProxyStorage: """Filtering Storage implementation. This is in charge of transparently filtering out known objects prior to adding them to storage. Sample configuration use case for filtering storage: .. code-block: yaml storage: cls: filter storage: cls: remote url: http://storage.internal.staging.swh.network:5002/ """ def __init__(self, storage): self.storage = get_storage(**storage) self.objects_seen = { - 'content': set(), # set of content hashes (sha256) seen - 'directory': set(), - 'revision': set(), + 'content': set(), # sha256 + 'skipped_content': set(), # sha1_git + 'directory': set(), # sha1_git + 'revision': set(), # sha1_git } def __getattr__(self, key): return getattr(self.storage, key) - def content_add(self, content: Sequence[Dict]) -> Dict: + def content_add(self, content: Iterable[Dict]) -> Dict: contents = list(content) contents_to_add = self._filter_missing_contents(contents) return self.storage.content_add( x for x in contents if x['sha256'] in contents_to_add ) - def directory_add(self, directories: Sequence[Dict]) -> Dict: + def skipped_content_add(self, content: Iterable[Dict]) -> Dict: + contents = list(content) + contents_to_add = self._filter_missing_skipped_contents(contents) + return self.storage.skipped_content_add( + x for x in contents if x['sha1_git'] in contents_to_add + ) + + def directory_add(self, directories: Iterable[Dict]) -> Dict: directories = list(directories) missing_ids = self._filter_missing_ids( 'directory', (d['id'] for d in directories) ) return self.storage.directory_add( d for d in directories if d['id'] in missing_ids ) def revision_add(self, revisions): revisions = list(revisions) missing_ids = self._filter_missing_ids( 'revision', (d['id'] for d in revisions) ) return self.storage.revision_add( r for r in revisions if r['id'] in missing_ids ) def _filter_missing_contents( - self, content_hashes: Sequence[Dict]) -> Set[bytes]: + self, content_hashes: Iterable[Dict]) -> Set[bytes]: """Return only the content keys missing from swh Args: content_hashes: List of sha256 to check for existence in swh storage """ objects_seen = self.objects_seen['content'] missing_hashes = [] for hashes in content_hashes: if hashes['sha256'] in objects_seen: continue objects_seen.add(hashes['sha256']) missing_hashes.append(hashes) return set(self.storage.content_missing( missing_hashes, key_hash='sha256', )) + def _filter_missing_skipped_contents( + self, content_hashes: Iterable[Dict]) -> Set[bytes]: + """Return only the content keys missing from swh + + Args: + content_hashes: List of sha1_git to check for existence in swh + storage + + """ + objects_seen = self.objects_seen['skipped_content'] + missing_hashes = [] + for hashes in content_hashes: + if hashes['sha1_git'] in objects_seen: + continue + objects_seen.add(hashes['sha1_git']) + missing_hashes.append(hashes) + + return {c['sha1_git'] + for c in self.storage.skipped_content_missing(missing_hashes)} + def _filter_missing_ids( self, object_type: str, ids: Generator[bytes, None, None]) -> Set[bytes]: """Filter missing ids from the storage for a given object type. Args: object_type: object type to use {revision, directory} - ids: Sequence of object_type ids + ids: Iterable of object_type ids Returns: Missing ids from the storage for object_type """ objects_seen = self.objects_seen[object_type] missing_ids = [] for id in ids: if id in objects_seen: continue objects_seen.add(id) missing_ids.append(id) fn_by_object_type = { 'revision': self.storage.revision_missing, 'directory': self.storage.directory_missing, } fn = fn_by_object_type[object_type] return set(fn(missing_ids)) diff --git a/swh/storage/in_memory.py b/swh/storage/in_memory.py index 6e260138..78ffb39a 100644 --- a/swh/storage/in_memory.py +++ b/swh/storage/in_memory.py @@ -1,1053 +1,1052 @@ # Copyright (C) 2015-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import re import bisect import dateutil import collections import copy import datetime import itertools import random from collections import defaultdict from datetime import timedelta from typing import Any, Dict, List, Optional import attr from swh.model.model import ( - Content, Directory, Revision, Release, Snapshot, OriginVisit, Origin, - SHA1_SIZE) + BaseContent, Content, SkippedContent, Directory, Revision, Release, + Snapshot, OriginVisit, Origin, SHA1_SIZE) from swh.model.hashutil import DEFAULT_ALGORITHMS, hash_to_bytes, hash_to_hex from swh.objstorage import get_objstorage from swh.objstorage.exc import ObjNotFoundError from .storage import get_journal_writer from .converters import origin_url_to_sha1 from .utils import get_partition_bounds_bytes # Max block size of contents to return BULK_BLOCK_CONTENT_LEN_MAX = 10000 def now(): return datetime.datetime.now(tz=datetime.timezone.utc) class InMemoryStorage: def __init__(self, journal_writer=None): self._contents = {} self._content_indexes = defaultdict(lambda: defaultdict(set)) self._skipped_contents = {} self._skipped_content_indexes = defaultdict(lambda: defaultdict(set)) self.reset() if journal_writer: self.journal_writer = get_journal_writer(**journal_writer) else: self.journal_writer = None def reset(self): self._directories = {} self._revisions = {} self._releases = {} self._snapshots = {} self._origins = {} self._origins_by_id = [] self._origins_by_sha1 = {} self._origin_visits = {} self._persons = [] self._origin_metadata = defaultdict(list) self._tools = {} self._metadata_providers = {} self._objects = defaultdict(list) # ideally we would want a skip list for both fast inserts and searches self._sorted_sha1s = [] self.objstorage = get_objstorage('memory', {}) def check_config(self, *, check_write): return True def _content_add(self, contents, with_data): - content_with_data = [] - content_without_data = [] for content in contents: if content.status is None: content.status = 'visible' + if content.status == 'absent': + raise ValueError('content with status=absent') if content.length is None: - content.length = -1 - if content.status != 'absent': - if self._content_key(content) not in self._contents: - content_with_data.append(content) - else: - if self._content_key(content) not in self._skipped_contents: - content_without_data.append(content) + raise ValueError('content with length=None') if self.journal_writer: - for content in content_with_data: + for content in contents: content = attr.evolve(content, data=None) self.journal_writer.write_addition('content', content) - for content in content_without_data: - self.journal_writer.write_addition('content', content) - - count_content_added, count_content_bytes_added = \ - self._content_add_present(content_with_data, with_data) - - count_skipped_content_added = self._content_add_absent( - content_without_data - ) summary = { - 'content:add': count_content_added, - 'skipped_content:add': count_skipped_content_added, + 'content:add': 0, } if with_data: - summary['content:add:bytes'] = count_content_bytes_added + summary['content:add:bytes'] = 0 - return summary - - def _content_add_present(self, contents, with_data): - count_content_added = 0 - count_content_bytes_added = 0 for content in contents: key = self._content_key(content) if key in self._contents: continue for algorithm in DEFAULT_ALGORITHMS: hash_ = content.get_hash(algorithm) if hash_ in self._content_indexes[algorithm]\ and (algorithm not in {'blake2s256', 'sha256'}): from . import HashCollision raise HashCollision(algorithm, hash_, key) for algorithm in DEFAULT_ALGORITHMS: hash_ = content.get_hash(algorithm) self._content_indexes[algorithm][hash_].add(key) self._objects[content.sha1_git].append( ('content', content.sha1)) self._contents[key] = content bisect.insort(self._sorted_sha1s, content.sha1) - count_content_added += 1 + summary['content:add'] += 1 if with_data: content_data = self._contents[key].data self._contents[key] = attr.evolve( self._contents[key], data=None) - count_content_bytes_added += len(content_data) + summary['content:add:bytes'] += len(content_data) self.objstorage.add(content_data, content.sha1) - return (count_content_added, count_content_bytes_added) - - def _content_add_absent(self, contents): - count = 0 - skipped_content_missing = self.skipped_content_missing(contents) - for content in skipped_content_missing: - key = self._content_key(content) - for algo in DEFAULT_ALGORITHMS: - self._skipped_content_indexes[algo][content.get_hash(algo)] \ - .add(key) - self._skipped_contents[key] = content - count += 1 - - return count - - def _content_to_model(self, contents): - for content in contents: - content = content.copy() - content.pop('origin', None) - yield Content.from_dict(content) + return summary def content_add(self, content): now = datetime.datetime.now(tz=datetime.timezone.utc) - content = [attr.evolve(c, ctime=now) - for c in self._content_to_model(content)] + content = [attr.evolve(Content.from_dict(c), ctime=now) + for c in content] return self._content_add(content, with_data=True) def content_update(self, content, keys=[]): if self.journal_writer: raise NotImplementedError( 'content_update is not yet supported with a journal_writer.') for cont_update in content: cont_update = cont_update.copy() sha1 = cont_update.pop('sha1') for old_key in self._content_indexes['sha1'][sha1]: old_cont = self._contents.pop(old_key) for algorithm in DEFAULT_ALGORITHMS: hash_ = old_cont.get_hash(algorithm) self._content_indexes[algorithm][hash_].remove(old_key) new_cont = attr.evolve(old_cont, **cont_update) new_key = self._content_key(new_cont) self._contents[new_key] = new_cont for algorithm in DEFAULT_ALGORITHMS: hash_ = new_cont.get_hash(algorithm) self._content_indexes[algorithm][hash_].add(new_key) def content_add_metadata(self, content): - content = list(self._content_to_model(content)) + content = [Content.from_dict(c) for c in content] return self._content_add(content, with_data=False) def content_get(self, content): # FIXME: Make this method support slicing the `data`. if len(content) > BULK_BLOCK_CONTENT_LEN_MAX: raise ValueError( "Sending at most %s contents." % BULK_BLOCK_CONTENT_LEN_MAX) for obj_id in content: try: data = self.objstorage.get(obj_id) except ObjNotFoundError: yield None continue yield {'sha1': obj_id, 'data': data} def content_get_range(self, start, end, limit=1000): if limit is None: raise ValueError('Development error: limit should not be None') from_index = bisect.bisect_left(self._sorted_sha1s, start) sha1s = itertools.islice(self._sorted_sha1s, from_index, None) sha1s = ((sha1, content_key) for sha1 in sha1s for content_key in self._content_indexes['sha1'][sha1]) matched = [] next_content = None for sha1, key in sha1s: if sha1 > end: break if len(matched) >= limit: next_content = sha1 break matched.append(self._contents[key].to_dict()) return { 'contents': matched, 'next': next_content, } def content_get_partition( self, partition_id: int, nb_partitions: int, limit: int = 1000, page_token: str = None): if limit is None: raise ValueError('Development error: limit should not be None') (start, end) = get_partition_bounds_bytes( partition_id, nb_partitions, SHA1_SIZE) if page_token: start = hash_to_bytes(page_token) if end is None: end = b'\xff'*SHA1_SIZE result = self.content_get_range(start, end, limit) result2 = { 'contents': result['contents'], 'next_page_token': None, } if result['next']: result2['next_page_token'] = hash_to_hex(result['next']) return result2 def content_get_metadata( self, contents: List[bytes]) -> Dict[bytes, List[Dict]]: result: Dict = {sha1: [] for sha1 in contents} for sha1 in contents: if sha1 in self._content_indexes['sha1']: objs = self._content_indexes['sha1'][sha1] # only 1 element as content_add_metadata would have raised a # hash collision otherwise for key in objs: d = self._contents[key].to_dict() del d['ctime'] if 'data' in d: del d['data'] result[sha1].append(d) return result def content_find(self, content): if not set(content).intersection(DEFAULT_ALGORITHMS): raise ValueError('content keys must contain at least one of: ' '%s' % ', '.join(sorted(DEFAULT_ALGORITHMS))) found = [] for algo in DEFAULT_ALGORITHMS: hash = content.get(algo) if hash and hash in self._content_indexes[algo]: found.append(self._content_indexes[algo][hash]) if not found: return [] keys = list(set.intersection(*found)) return [self._contents[key].to_dict() for key in keys] def content_missing(self, content, key_hash='sha1'): for cont in content: for (algo, hash_) in cont.items(): if algo not in DEFAULT_ALGORITHMS: continue if hash_ not in self._content_indexes.get(algo, []): yield cont[key_hash] break else: for result in self.content_find(cont): if result['status'] == 'missing': yield cont[key_hash] def content_missing_per_sha1(self, contents): for content in contents: if content not in self._content_indexes['sha1']: yield content def content_missing_per_sha1_git(self, contents): for content in contents: if content not in self._content_indexes['sha1_git']: yield content + def content_get_random(self): + return random.choice(list(self._content_indexes['sha1_git'])) + + def _skipped_content_add(self, contents): + for content in contents: + if content.status is None: + content = attr.evolve(content, status='absent') + if content.length is None: + content = attr.evolve(content, length=-1) + if content.status != 'absent': + raise ValueError(f'Content with status={content.status}') + + if self.journal_writer: + for content in contents: + self.journal_writer.write_addition('content', content) + + summary = { + 'skipped_content:add': 0 + } + + skipped_content_missing = self.skipped_content_missing( + [c.to_dict() for c in contents]) + for content in skipped_content_missing: + key = self._content_key(content, allow_missing=True) + for algo in DEFAULT_ALGORITHMS: + if algo in content: + self._skipped_content_indexes[algo][content[algo]] \ + .add(key) + self._skipped_contents[key] = content + summary['skipped_content:add'] += 1 + + return summary + def skipped_content_missing(self, contents): for content in contents: for (key, algorithm) in self._content_key_algorithm(content): if algorithm == 'blake2s256': continue if key not in self._skipped_content_indexes[algorithm]: # index must contain hashes of algos except blake2s256 # else the content is considered skipped - yield content + yield {algo: content[algo] + for algo in DEFAULT_ALGORITHMS + if content[algo] is not None} break - def content_get_random(self): - return random.choice(list(self._content_indexes['sha1_git'])) + def skipped_content_add(self, content): + content = list(content) + now = datetime.datetime.now(tz=datetime.timezone.utc) + content = [attr.evolve(SkippedContent.from_dict(c), ctime=now) + for c in content] + return self._skipped_content_add(content) def directory_add(self, directories): directories = list(directories) if self.journal_writer: self.journal_writer.write_additions( 'directory', (dir_ for dir_ in directories if dir_['id'] not in self._directories)) directories = [Directory.from_dict(d) for d in directories] count = 0 for directory in directories: if directory.id not in self._directories: count += 1 self._directories[directory.id] = directory self._objects[directory.id].append( ('directory', directory.id)) return {'directory:add': count} def directory_missing(self, directories): for id in directories: if id not in self._directories: yield id def _join_dentry_to_content(self, dentry): keys = ( 'status', 'sha1', 'sha1_git', 'sha256', 'length', ) ret = dict.fromkeys(keys) ret.update(dentry) if ret['type'] == 'file': # TODO: Make it able to handle more than one content content = self.content_find({'sha1_git': ret['target']}) if content: content = content[0] for key in keys: ret[key] = content[key] return ret def _directory_ls(self, directory_id, recursive, prefix=b''): if directory_id in self._directories: for entry in self._directories[directory_id].entries: ret = self._join_dentry_to_content(entry.to_dict()) ret['name'] = prefix + ret['name'] ret['dir_id'] = directory_id yield ret if recursive and ret['type'] == 'dir': yield from self._directory_ls( ret['target'], True, prefix + ret['name'] + b'/') def directory_ls(self, directory, recursive=False): yield from self._directory_ls(directory, recursive) def directory_entry_get_by_path(self, directory, paths): return self._directory_entry_get_by_path(directory, paths, b'') def directory_get_random(self): if not self._directories: return None return random.choice(list(self._directories)) def _directory_entry_get_by_path(self, directory, paths, prefix): if not paths: return contents = list(self.directory_ls(directory)) if not contents: return def _get_entry(entries, name): for entry in entries: if entry['name'] == name: entry = entry.copy() entry['name'] = prefix + entry['name'] return entry first_item = _get_entry(contents, paths[0]) if len(paths) == 1: return first_item if not first_item or first_item['type'] != 'dir': return return self._directory_entry_get_by_path( first_item['target'], paths[1:], prefix + paths[0] + b'/') def revision_add(self, revisions): revisions = list(revisions) if self.journal_writer: self.journal_writer.write_additions( 'revision', (rev for rev in revisions if rev['id'] not in self._revisions)) revisions = [Revision.from_dict(rev) for rev in revisions] count = 0 for revision in revisions: if revision.id not in self._revisions: revision = attr.evolve( revision, committer=self._person_add(revision.committer), author=self._person_add(revision.author)) self._revisions[revision.id] = revision self._objects[revision.id].append( ('revision', revision.id)) count += 1 return {'revision:add': count} def revision_missing(self, revisions): for id in revisions: if id not in self._revisions: yield id def revision_get(self, revisions): for id in revisions: if id in self._revisions: yield self._revisions.get(id).to_dict() else: yield None def _get_parent_revs(self, rev_id, seen, limit): if limit and len(seen) >= limit: return if rev_id in seen or rev_id not in self._revisions: return seen.add(rev_id) yield self._revisions[rev_id].to_dict() for parent in self._revisions[rev_id].parents: yield from self._get_parent_revs(parent, seen, limit) def revision_log(self, revisions, limit=None): seen = set() for rev_id in revisions: yield from self._get_parent_revs(rev_id, seen, limit) def revision_shortlog(self, revisions, limit=None): yield from ((rev['id'], rev['parents']) for rev in self.revision_log(revisions, limit)) def revision_get_random(self): return random.choice(list(self._revisions)) def release_add(self, releases): releases = list(releases) if self.journal_writer: self.journal_writer.write_additions( 'release', (rel for rel in releases if rel['id'] not in self._releases)) releases = [Release.from_dict(rel) for rel in releases] count = 0 for rel in releases: if rel.id not in self._releases: if rel.author: self._person_add(rel.author) self._objects[rel.id].append( ('release', rel.id)) self._releases[rel.id] = rel count += 1 return {'release:add': count} def release_missing(self, releases): yield from (rel for rel in releases if rel not in self._releases) def release_get(self, releases): for rel_id in releases: if rel_id in self._releases: yield self._releases[rel_id].to_dict() else: yield None def release_get_random(self): return random.choice(list(self._releases)) def snapshot_add(self, snapshots): count = 0 snapshots = (Snapshot.from_dict(d) for d in snapshots) snapshots = (snap for snap in snapshots if snap.id not in self._snapshots) for snapshot in snapshots: if self.journal_writer: self.journal_writer.write_addition('snapshot', snapshot) sorted_branch_names = sorted(snapshot.branches) self._snapshots[snapshot.id] = (snapshot, sorted_branch_names) self._objects[snapshot.id].append(('snapshot', snapshot.id)) count += 1 return {'snapshot:add': count} def snapshot_missing(self, snapshots): for id in snapshots: if id not in self._snapshots: yield id def snapshot_get(self, snapshot_id): return self.snapshot_get_branches(snapshot_id) def snapshot_get_by_origin_visit(self, origin, visit): origin_url = self._get_origin_url(origin) if not origin_url: return if origin_url not in self._origins or \ visit > len(self._origin_visits[origin_url]): return None snapshot_id = self._origin_visits[origin_url][visit-1].snapshot if snapshot_id: return self.snapshot_get(snapshot_id) else: return None def snapshot_get_latest(self, origin, allowed_statuses=None): origin_url = self._get_origin_url(origin) if not origin_url: return visit = self.origin_visit_get_latest( origin_url, allowed_statuses=allowed_statuses, require_snapshot=True) if visit and visit['snapshot']: snapshot = self.snapshot_get(visit['snapshot']) if not snapshot: raise ValueError( 'last origin visit references an unknown snapshot') return snapshot def snapshot_count_branches(self, snapshot_id): (snapshot, _) = self._snapshots[snapshot_id] return collections.Counter(branch.target_type.value if branch else None for branch in snapshot.branches.values()) def snapshot_get_branches(self, snapshot_id, branches_from=b'', branches_count=1000, target_types=None): res = self._snapshots.get(snapshot_id) if res is None: return None (snapshot, sorted_branch_names) = res from_index = bisect.bisect_left( sorted_branch_names, branches_from) if target_types: next_branch = None branches = {} for branch_name in sorted_branch_names[from_index:]: branch = snapshot.branches[branch_name] if branch and branch.target_type.value in target_types: if len(branches) < branches_count: branches[branch_name] = branch else: next_branch = branch_name break else: # As there is no 'target_types', we can do that much faster to_index = from_index + branches_count returned_branch_names = sorted_branch_names[from_index:to_index] branches = {branch_name: snapshot.branches[branch_name] for branch_name in returned_branch_names} if to_index >= len(sorted_branch_names): next_branch = None else: next_branch = sorted_branch_names[to_index] branches = {name: branch.to_dict() if branch else None for (name, branch) in branches.items()} return { 'id': snapshot_id, 'branches': branches, 'next_branch': next_branch, } def snapshot_get_random(self): return random.choice(list(self._snapshots)) def object_find_by_sha1_git(self, ids): ret = {} for id_ in ids: objs = self._objects.get(id_, []) ret[id_] = [{ 'sha1_git': id_, 'type': obj[0], } for obj in objs] return ret def _convert_origin(self, t): if t is None: return None return t.to_dict() def origin_get(self, origins): if isinstance(origins, dict): # Old API return_single = True origins = [origins] else: return_single = False # Sanity check to be error-compatible with the pgsql backend if any('id' in origin for origin in origins) \ and not all('id' in origin for origin in origins): raise ValueError( 'Either all origins or none at all should have an "id".') if any('url' in origin for origin in origins) \ and not all('url' in origin for origin in origins): raise ValueError( 'Either all origins or none at all should have ' 'an "url" key.') results = [] for origin in origins: result = None if 'url' in origin: if origin['url'] in self._origins: result = self._origins[origin['url']] else: raise ValueError( 'Origin must have an url.') results.append(self._convert_origin(result)) if return_single: assert len(results) == 1 return results[0] else: return results def origin_get_by_sha1(self, sha1s): return [ self._convert_origin(self._origins_by_sha1.get(sha1)) for sha1 in sha1s ] def origin_get_range(self, origin_from=1, origin_count=100): origin_from = max(origin_from, 1) if origin_from <= len(self._origins_by_id): max_idx = origin_from + origin_count - 1 if max_idx > len(self._origins_by_id): max_idx = len(self._origins_by_id) for idx in range(origin_from-1, max_idx): origin = self._convert_origin( self._origins[self._origins_by_id[idx]]) yield {'id': idx+1, **origin} def origin_list(self, page_token: Optional[str] = None, limit: int = 100 ) -> dict: origin_urls = sorted(self._origins) if page_token: from_ = bisect.bisect_left(origin_urls, page_token) else: from_ = 0 result = { 'origins': [{'url': origin_url} for origin_url in origin_urls[from_:from_+limit]] } if from_+limit < len(origin_urls): result['next_page_token'] = origin_urls[from_+limit] return result def origin_search(self, url_pattern, offset=0, limit=50, regexp=False, with_visit=False): origins = map(self._convert_origin, self._origins.values()) if regexp: pat = re.compile(url_pattern) origins = [orig for orig in origins if pat.search(orig['url'])] else: origins = [orig for orig in origins if url_pattern in orig['url']] if with_visit: origins = [ orig for orig in origins if len(self._origin_visits[orig['url']]) > 0 and set(ov.snapshot for ov in self._origin_visits[orig['url']] if ov.snapshot) & set(self._snapshots)] return origins[offset:offset+limit] def origin_count(self, url_pattern, regexp=False, with_visit=False): return len(self.origin_search(url_pattern, regexp=regexp, with_visit=with_visit, limit=len(self._origins))) def origin_add(self, origins): origins = copy.deepcopy(list(origins)) for origin in origins: self.origin_add_one(origin) return origins def origin_add_one(self, origin): origin = Origin.from_dict(origin) if origin.url not in self._origins: if self.journal_writer: self.journal_writer.write_addition('origin', origin) # generate an origin_id because it is needed by origin_get_range. # TODO: remove this when we remove origin_get_range origin_id = len(self._origins) + 1 self._origins_by_id.append(origin.url) assert len(self._origins_by_id) == origin_id self._origins[origin.url] = origin self._origins_by_sha1[origin_url_to_sha1(origin.url)] = origin self._origin_visits[origin.url] = [] self._objects[origin.url].append(('origin', origin.url)) return origin.url def origin_visit_add(self, origin, date, type): origin_url = origin if origin_url is None: raise ValueError('Unknown origin.') if isinstance(date, str): # FIXME: Converge on iso8601 at some point date = dateutil.parser.parse(date) elif not isinstance(date, datetime.datetime): raise TypeError('date must be a datetime or a string.') visit_ret = None if origin_url in self._origins: origin = self._origins[origin_url] # visit ids are in the range [1, +inf[ visit_id = len(self._origin_visits[origin_url]) + 1 status = 'ongoing' visit = OriginVisit( origin=origin.url, date=date, type=type, status=status, snapshot=None, metadata=None, visit=visit_id, ) self._origin_visits[origin_url].append(visit) visit_ret = { 'origin': origin.url, 'visit': visit_id, } self._objects[(origin_url, visit_id)].append( ('origin_visit', None)) if self.journal_writer: self.journal_writer.write_addition('origin_visit', visit) return visit_ret def origin_visit_update(self, origin, visit_id, status=None, metadata=None, snapshot=None): if not isinstance(origin, str): raise TypeError('origin must be a string, not %r' % (origin,)) origin_url = self._get_origin_url(origin) if origin_url is None: raise ValueError('Unknown origin.') try: visit = self._origin_visits[origin_url][visit_id-1] except IndexError: raise ValueError('Unknown visit_id for this origin') \ from None updates = {} if status: updates['status'] = status if metadata: updates['metadata'] = metadata if snapshot: updates['snapshot'] = snapshot visit = attr.evolve(visit, **updates) if self.journal_writer: self.journal_writer.write_update('origin_visit', visit) self._origin_visits[origin_url][visit_id-1] = visit def origin_visit_upsert(self, visits): for visit in visits: if not isinstance(visit['origin'], str): raise TypeError("visit['origin'] must be a string, not %r" % (visit['origin'],)) visits = [OriginVisit.from_dict(d) for d in visits] if self.journal_writer: for visit in visits: self.journal_writer.write_addition('origin_visit', visit) for visit in visits: visit_id = visit.visit origin_url = visit.origin visit = attr.evolve(visit, origin=origin_url) self._objects[(origin_url, visit_id)].append( ('origin_visit', None)) while len(self._origin_visits[origin_url]) <= visit_id: self._origin_visits[origin_url].append(None) self._origin_visits[origin_url][visit_id-1] = visit def _convert_visit(self, visit): if visit is None: return visit = visit.to_dict() return visit def origin_visit_get(self, origin, last_visit=None, limit=None): origin_url = self._get_origin_url(origin) if origin_url in self._origin_visits: visits = self._origin_visits[origin_url] if last_visit is not None: visits = visits[last_visit:] if limit is not None: visits = visits[:limit] for visit in visits: if not visit: continue visit_id = visit.visit yield self._convert_visit( self._origin_visits[origin_url][visit_id-1]) def origin_visit_find_by_date(self, origin, visit_date): origin_url = self._get_origin_url(origin) if origin_url in self._origin_visits: visits = self._origin_visits[origin_url] visit = min( visits, key=lambda v: (abs(v.date - visit_date), -v.visit)) return self._convert_visit(visit) def origin_visit_get_by(self, origin, visit): origin_url = self._get_origin_url(origin) if origin_url in self._origin_visits and \ visit <= len(self._origin_visits[origin_url]): return self._convert_visit( self._origin_visits[origin_url][visit-1]) def origin_visit_get_latest( self, origin, allowed_statuses=None, require_snapshot=False): origin = self._origins.get(origin) if not origin: return visits = self._origin_visits[origin.url] if allowed_statuses is not None: visits = [visit for visit in visits if visit.status in allowed_statuses] if require_snapshot: visits = [visit for visit in visits if visit.snapshot] visit = max( visits, key=lambda v: (v.date, v.visit), default=None) return self._convert_visit(visit) def _select_random_origin_visit_by_type(self, type: str) -> str: while True: url = random.choice(list(self._origin_visits.keys())) random_origin_visits = self._origin_visits[url] if random_origin_visits[0].type == type: return url def origin_visit_get_random(self, type: str) -> Optional[Dict[str, Any]]: url = self._select_random_origin_visit_by_type(type) random_origin_visits = copy.deepcopy(self._origin_visits[url]) random_origin_visits.reverse() back_in_the_day = now() - timedelta(weeks=12) # 3 months back # This should be enough for tests for visit in random_origin_visits: if visit.date > back_in_the_day and visit.status == 'full': return visit.to_dict() else: return None def stat_counters(self): keys = ( 'content', 'directory', 'origin', 'origin_visit', 'person', 'release', 'revision', 'skipped_content', 'snapshot' ) stats = {key: 0 for key in keys} stats.update(collections.Counter( obj_type for (obj_type, obj_id) in itertools.chain(*self._objects.values()))) return stats def refresh_stat_counters(self): pass def origin_metadata_add(self, origin_url, ts, provider, tool, metadata): if not isinstance(origin_url, str): raise TypeError('origin_id must be str, not %r' % (origin_url,)) if isinstance(ts, str): ts = dateutil.parser.parse(ts) origin_metadata = { 'origin_url': origin_url, 'discovery_date': ts, 'tool_id': tool, 'metadata': metadata, 'provider_id': provider, } self._origin_metadata[origin_url].append(origin_metadata) return None def origin_metadata_get_by(self, origin_url, provider_type=None): if not isinstance(origin_url, str): raise TypeError('origin_url must be str, not %r' % (origin_url,)) metadata = [] for item in self._origin_metadata[origin_url]: item = copy.deepcopy(item) provider = self.metadata_provider_get(item['provider_id']) for attr_name in ('name', 'type', 'url'): item['provider_' + attr_name] = \ provider['provider_' + attr_name] metadata.append(item) return metadata def tool_add(self, tools): inserted = [] for tool in tools: key = self._tool_key(tool) assert 'id' not in tool record = copy.deepcopy(tool) record['id'] = key # TODO: remove this if key not in self._tools: self._tools[key] = record inserted.append(copy.deepcopy(self._tools[key])) return inserted def tool_get(self, tool): return self._tools.get(self._tool_key(tool)) def metadata_provider_add(self, provider_name, provider_type, provider_url, metadata): provider = { 'provider_name': provider_name, 'provider_type': provider_type, 'provider_url': provider_url, 'metadata': metadata, } key = self._metadata_provider_key(provider) provider['id'] = key self._metadata_providers[key] = provider return key def metadata_provider_get(self, provider_id): return self._metadata_providers.get(provider_id) def metadata_provider_get_by(self, provider): key = self._metadata_provider_key(provider) return self._metadata_providers.get(key) def _get_origin_url(self, origin): if isinstance(origin, str): return origin else: raise TypeError('origin must be a string.') def _person_add(self, person): key = ('person', person.fullname) if key not in self._objects: person_id = len(self._persons) + 1 self._persons.append(person) self._objects[key].append(('person', person_id)) else: person_id = self._objects[key][0][1] person = self._persons[person_id-1] return person @staticmethod - def _content_key(content): + def _content_key(content, allow_missing=False): """A stable key for a content""" - return tuple(getattr(content, key) + return tuple(getattr(content, key, None) for key in sorted(DEFAULT_ALGORITHMS)) @staticmethod def _content_key_algorithm(content): """ A stable key and the algorithm for a content""" - if isinstance(content, Content): + if isinstance(content, BaseContent): content = content.to_dict() return tuple((content.get(key), key) for key in sorted(DEFAULT_ALGORITHMS)) @staticmethod def _tool_key(tool): return '%r %r %r' % (tool['name'], tool['version'], tuple(sorted(tool['configuration'].items()))) @staticmethod def _metadata_provider_key(provider): return '%r %r' % (provider['provider_name'], provider['provider_url']) def diff_directories(self, from_dir, to_dir, track_renaming=False): raise NotImplementedError('InMemoryStorage.diff_directories') def diff_revisions(self, from_rev, to_rev, track_renaming=False): raise NotImplementedError('InMemoryStorage.diff_revisions') def diff_revision(self, revision, track_renaming=False): raise NotImplementedError('InMemoryStorage.diff_revision') diff --git a/swh/storage/interface.py b/swh/storage/interface.py index 47a49c25..72ed9767 100644 --- a/swh/storage/interface.py +++ b/swh/storage/interface.py @@ -1,1235 +1,1268 @@ # Copyright (C) 2015-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information from typing import Any, Dict, List, Optional from swh.core.api import remote_api_endpoint def deprecated(f): f.deprecated_endpoint = True return f class StorageInterface: @remote_api_endpoint('check_config') def check_config(self, *, check_write): """Check that the storage is configured and ready to go.""" ... @remote_api_endpoint('content/add') def content_add(self, content): """Add content blobs to the storage Args: contents (iterable): iterable of dictionaries representing individual pieces of content to add. Each dictionary has the following keys: - data (bytes): the actual content - - length (int): content length (default: -1) + - length (int): content length - one key for each checksum algorithm in :data:`swh.model.hashutil.ALGORITHMS`, mapped to the corresponding checksum - - status (str): one of visible, hidden, absent - - reason (str): if status = absent, the reason why - - origin (int): if status = absent, the origin we saw the - content in + - status (str): one of visible, hidden Raises: The following exceptions can occur: - HashCollision in case of collision - Any other exceptions raise by the db In case of errors, some of the content may have been stored in the DB and in the objstorage. Since additions to both idempotent, that should not be a problem. Returns: - Summary dict with the following key and associated values: + Summary dict with the following keys and associated values: content:add: New contents added content:add:bytes: Sum of the contents' length data - skipped_content:add: New skipped contents (no data) added """ ... @remote_api_endpoint('content/update') def content_update(self, content, keys=[]): """Update content blobs to the storage. Does nothing for unknown contents or skipped ones. Args: content (iterable): iterable of dictionaries representing individual pieces of content to update. Each dictionary has the following keys: - data (bytes): the actual content - length (int): content length (default: -1) - one key for each checksum algorithm in :data:`swh.model.hashutil.ALGORITHMS`, mapped to the corresponding checksum - status (str): one of visible, hidden, absent keys (list): List of keys (str) whose values needs an update, e.g., new hash column """ ... @remote_api_endpoint('content/add_metadata') def content_add_metadata(self, content): """Add content metadata to the storage (like `content_add`, but without inserting to the objstorage). Args: content (iterable): iterable of dictionaries representing individual pieces of content to add. Each dictionary has the following keys: - length (int): content length (default: -1) - one key for each checksum algorithm in :data:`swh.model.hashutil.ALGORITHMS`, mapped to the corresponding checksum - status (str): one of visible, hidden, absent - reason (str): if status = absent, the reason why - origin (int): if status = absent, the origin we saw the content in - ctime (datetime): time of insertion in the archive Returns: Summary dict with the following key and associated values: content:add: New contents added skipped_content:add: New skipped contents (no data) added """ ... @remote_api_endpoint('content/data') def content_get(self, content): """Retrieve in bulk contents and their data. This generator yields exactly as many items than provided sha1 identifiers, but callers should not assume this will always be true. It may also yield `None` values in case an object was not found. Args: content: iterables of sha1 Yields: Dict[str, bytes]: Generates streams of contents as dict with their raw data: - sha1 (bytes): content id - data (bytes): content's raw data Raises: ValueError in case of too much contents are required. cf. BULK_BLOCK_CONTENT_LEN_MAX """ ... @deprecated @remote_api_endpoint('content/range') def content_get_range(self, start, end, limit=1000): """Retrieve contents within range [start, end] bound by limit. Note that this function may return more than one blob per hash. The limit is enforced with multiplicity (ie. two blobs with the same hash will count twice toward the limit). Args: **start** (bytes): Starting identifier range (expected smaller than end) **end** (bytes): Ending identifier range (expected larger than start) **limit** (int): Limit result (default to 1000) Returns: a dict with keys: - contents [dict]: iterable of contents in between the range. - next (bytes): There remains content in the range starting from this next sha1 """ ... @remote_api_endpoint('content/partition') def content_get_partition( self, partition_id: int, nb_partitions: int, limit: int = 1000, page_token: str = None): """Splits contents into nb_partitions, and returns one of these based on partition_id (which must be in [0, nb_partitions-1]) There is no guarantee on how the partitioning is done, or the result order. Args: partition_id (int): index of the partition to fetch nb_partitions (int): total number of partitions to split into limit (int): Limit result (default to 1000) page_token (Optional[str]): opaque token used for pagination. Returns: a dict with keys: - contents (List[dict]): iterable of contents in the partition. - **next_page_token** (Optional[str]): opaque token to be used as `page_token` for retrieving the next page. if absent, there is no more pages to gather. """ ... @remote_api_endpoint('content/metadata') def content_get_metadata( self, contents: List[bytes]) -> Dict[bytes, List[Dict]]: """Retrieve content metadata in bulk Args: content: iterable of content identifiers (sha1) Returns: a dict with keys the content's sha1 and the associated value either the existing content's metadata or None if the content does not exist. """ ... @remote_api_endpoint('content/missing') def content_missing(self, content, key_hash='sha1'): """List content missing from storage Args: content ([dict]): iterable of dictionaries whose keys are either 'length' or an item of :data:`swh.model.hashutil.ALGORITHMS`; mapped to the corresponding checksum (or length). key_hash (str): name of the column to use as hash id result (default: 'sha1') Returns: iterable ([bytes]): missing content ids (as per the key_hash column) Raises: TODO: an exception when we get a hash collision. """ ... @remote_api_endpoint('content/missing/sha1') def content_missing_per_sha1(self, contents): """List content missing from storage based only on sha1. Args: contents: Iterable of sha1 to check for absence. Returns: iterable: missing ids Raises: TODO: an exception when we get a hash collision. """ ... @remote_api_endpoint('content/missing/sha1_git') def content_missing_per_sha1_git(self, contents): """List content missing from storage based only on sha1_git. Args: contents (Iterable): An iterable of content id (sha1_git) Yields: missing contents sha1_git """ ... - @remote_api_endpoint('content/skipped/missing') - def skipped_content_missing(self, contents): - """List skipped_content missing from storage - - Args: - content: iterable of dictionaries containing the data for each - checksum algorithm. - - Returns: - iterable: missing signatures - - """ - ... - @remote_api_endpoint('content/present') def content_find(self, content): """Find a content hash in db. Args: content: a dictionary representing one content hash, mapping checksum algorithm names (see swh.model.hashutil.ALGORITHMS) to checksum values Returns: a triplet (sha1, sha1_git, sha256) if the content exist or None otherwise. Raises: ValueError: in case the key of the dictionary is not sha1, sha1_git nor sha256. """ ... @remote_api_endpoint('content/get_random') def content_get_random(self): """Finds a random content id. Returns: a sha1_git """ ... + @remote_api_endpoint('content/skipped/add') + def skipped_content_add(self, content): + """Add contents to the skipped_content list, which contains + (partial) information about content missing from the archive. + + Args: + contents (iterable): iterable of dictionaries representing + individual pieces of content to add. Each dictionary has the + following keys: + + - length (Optional[int]): content length (default: -1) + - one key for each checksum algorithm in + :data:`swh.model.hashutil.ALGORITHMS`, mapped to the + corresponding checksum; each is optional + - status (str): must be "absent" + - reason (str): the reason why the content is absent + - origin (int): if status = absent, the origin we saw the + content in + + Raises: + + The following exceptions can occur: + + - HashCollision in case of collision + - Any other exceptions raise by the backend + + In case of errors, some content may have been stored in + the DB and in the objstorage. + Since additions to both idempotent, that should not be a problem. + + Returns: + Summary dict with the following key and associated values: + + skipped_content:add: New skipped contents (no data) added + """ + ... + + @remote_api_endpoint('content/skipped/missing') + def skipped_content_missing(self, contents): + """List skipped_content missing from storage + + Args: + content: iterable of dictionaries containing the data for each + checksum algorithm. + + Returns: + iterable: missing signatures + + """ + ... + @remote_api_endpoint('directory/add') def directory_add(self, directories): """Add directories to the storage Args: directories (iterable): iterable of dictionaries representing the individual directories to add. Each dict has the following keys: - id (sha1_git): the id of the directory to add - entries (list): list of dicts for each entry in the directory. Each dict has the following keys: - name (bytes) - type (one of 'file', 'dir', 'rev'): type of the directory entry (file, directory, revision) - target (sha1_git): id of the object pointed at by the directory entry - perms (int): entry permissions Returns: Summary dict of keys with associated count as values: directory:add: Number of directories actually added """ ... @remote_api_endpoint('directory/missing') def directory_missing(self, directories): """List directories missing from storage Args: directories (iterable): an iterable of directory ids Yields: missing directory ids """ ... @remote_api_endpoint('directory/ls') def directory_ls(self, directory, recursive=False): """Get entries for one directory. Args: - directory: the directory to list entries from. - recursive: if flag on, this list recursively from this directory. Returns: List of entries for such directory. If `recursive=True`, names in the path of a dir/file not at the root are concatenated with a slash (`/`). """ ... @remote_api_endpoint('directory/path') def directory_entry_get_by_path(self, directory, paths): """Get the directory entry (either file or dir) from directory with path. Args: - directory: sha1 of the top level directory - paths: path to lookup from the top level directory. From left (top) to right (bottom). Returns: The corresponding directory entry if found, None otherwise. """ ... @remote_api_endpoint('directory/get_random') def directory_get_random(self): """Finds a random directory id. Returns: a sha1_git """ ... @remote_api_endpoint('revision/add') def revision_add(self, revisions): """Add revisions to the storage Args: revisions (Iterable[dict]): iterable of dictionaries representing the individual revisions to add. Each dict has the following keys: - **id** (:class:`sha1_git`): id of the revision to add - **date** (:class:`dict`): date the revision was written - **committer_date** (:class:`dict`): date the revision got added to the origin - **type** (one of 'git', 'tar'): type of the revision added - **directory** (:class:`sha1_git`): the directory the revision points at - **message** (:class:`bytes`): the message associated with the revision - **author** (:class:`Dict[str, bytes]`): dictionary with keys: name, fullname, email - **committer** (:class:`Dict[str, bytes]`): dictionary with keys: name, fullname, email - **metadata** (:class:`jsonb`): extra information as dictionary - **synthetic** (:class:`bool`): revision's nature (tarball, directory creates synthetic revision`) - **parents** (:class:`list[sha1_git]`): the parents of this revision date dictionaries have the form defined in :mod:`swh.model`. Returns: Summary dict of keys with associated count as values revision:add: New objects actually stored in db """ ... @remote_api_endpoint('revision/missing') def revision_missing(self, revisions): """List revisions missing from storage Args: revisions (iterable): revision ids Yields: missing revision ids """ ... @remote_api_endpoint('revision') def revision_get(self, revisions): """Get all revisions from storage Args: revisions: an iterable of revision ids Returns: iterable: an iterable of revisions as dictionaries (or None if the revision doesn't exist) """ ... @remote_api_endpoint('revision/log') def revision_log(self, revisions, limit=None): """Fetch revision entry from the given root revisions. Args: revisions: array of root revision to lookup limit: limitation on the output result. Default to None. Yields: List of revision log from such revisions root. """ ... @remote_api_endpoint('revision/shortlog') def revision_shortlog(self, revisions, limit=None): """Fetch the shortlog for the given revisions Args: revisions: list of root revisions to lookup limit: depth limitation for the output Yields: a list of (id, parents) tuples. """ ... @remote_api_endpoint('revision/get_random') def revision_get_random(self): """Finds a random revision id. Returns: a sha1_git """ ... @remote_api_endpoint('release/add') def release_add(self, releases): """Add releases to the storage Args: releases (Iterable[dict]): iterable of dictionaries representing the individual releases to add. Each dict has the following keys: - **id** (:class:`sha1_git`): id of the release to add - **revision** (:class:`sha1_git`): id of the revision the release points to - **date** (:class:`dict`): the date the release was made - **name** (:class:`bytes`): the name of the release - **comment** (:class:`bytes`): the comment associated with the release - **author** (:class:`Dict[str, bytes]`): dictionary with keys: name, fullname, email the date dictionary has the form defined in :mod:`swh.model`. Returns: Summary dict of keys with associated count as values release:add: New objects contents actually stored in db """ ... @remote_api_endpoint('release/missing') def release_missing(self, releases): """List releases missing from storage Args: releases: an iterable of release ids Returns: a list of missing release ids """ ... @remote_api_endpoint('release') def release_get(self, releases): """Given a list of sha1, return the releases's information Args: releases: list of sha1s Yields: dicts with the same keys as those given to `release_add` (or ``None`` if a release does not exist) """ ... @remote_api_endpoint('release/get_random') def release_get_random(self): """Finds a random release id. Returns: a sha1_git """ ... @remote_api_endpoint('snapshot/add') def snapshot_add(self, snapshots): """Add snapshots to the storage. Args: snapshot ([dict]): the snapshots to add, containing the following keys: - **id** (:class:`bytes`): id of the snapshot - **branches** (:class:`dict`): branches the snapshot contains, mapping the branch name (:class:`bytes`) to the branch target, itself a :class:`dict` (or ``None`` if the branch points to an unknown object) - **target_type** (:class:`str`): one of ``content``, ``directory``, ``revision``, ``release``, ``snapshot``, ``alias`` - **target** (:class:`bytes`): identifier of the target (currently a ``sha1_git`` for all object kinds, or the name of the target branch for aliases) Raises: ValueError: if the origin or visit id does not exist. Returns: Summary dict of keys with associated count as values snapshot:add: Count of object actually stored in db """ ... @remote_api_endpoint('snapshot/missing') def snapshot_missing(self, snapshots): """List snapshots missing from storage Args: snapshots (iterable): an iterable of snapshot ids Yields: missing snapshot ids """ ... @remote_api_endpoint('snapshot') def snapshot_get(self, snapshot_id): """Get the content, possibly partial, of a snapshot with the given id The branches of the snapshot are iterated in the lexicographical order of their names. .. warning:: At most 1000 branches contained in the snapshot will be returned for performance reasons. In order to browse the whole set of branches, the method :meth:`snapshot_get_branches` should be used instead. Args: snapshot_id (bytes): identifier of the snapshot Returns: dict: a dict with three keys: * **id**: identifier of the snapshot * **branches**: a dict of branches contained in the snapshot whose keys are the branches' names. * **next_branch**: the name of the first branch not returned or :const:`None` if the snapshot has less than 1000 branches. """ ... @remote_api_endpoint('snapshot/by_origin_visit') def snapshot_get_by_origin_visit(self, origin, visit): """Get the content, possibly partial, of a snapshot for the given origin visit The branches of the snapshot are iterated in the lexicographical order of their names. .. warning:: At most 1000 branches contained in the snapshot will be returned for performance reasons. In order to browse the whole set of branches, the method :meth:`snapshot_get_branches` should be used instead. Args: origin (int): the origin identifier visit (int): the visit identifier Returns: dict: None if the snapshot does not exist; a dict with three keys otherwise: * **id**: identifier of the snapshot * **branches**: a dict of branches contained in the snapshot whose keys are the branches' names. * **next_branch**: the name of the first branch not returned or :const:`None` if the snapshot has less than 1000 branches. """ ... @remote_api_endpoint('snapshot/latest') def snapshot_get_latest(self, origin, allowed_statuses=None): """Get the content, possibly partial, of the latest snapshot for the given origin, optionally only from visits that have one of the given allowed_statuses The branches of the snapshot are iterated in the lexicographical order of their names. .. warning:: At most 1000 branches contained in the snapshot will be returned for performance reasons. In order to browse the whole set of branches, the method :meth:`snapshot_get_branches` should be used instead. Args: origin (str): the origin's URL allowed_statuses (list of str): list of visit statuses considered to find the latest snapshot for the visit. For instance, ``allowed_statuses=['full']`` will only consider visits that have successfully run to completion. Returns: dict: a dict with three keys: * **id**: identifier of the snapshot * **branches**: a dict of branches contained in the snapshot whose keys are the branches' names. * **next_branch**: the name of the first branch not returned or :const:`None` if the snapshot has less than 1000 branches. """ ... @remote_api_endpoint('snapshot/count_branches') def snapshot_count_branches(self, snapshot_id): """Count the number of branches in the snapshot with the given id Args: snapshot_id (bytes): identifier of the snapshot Returns: dict: A dict whose keys are the target types of branches and values their corresponding amount """ ... @remote_api_endpoint('snapshot/get_branches') def snapshot_get_branches(self, snapshot_id, branches_from=b'', branches_count=1000, target_types=None): """Get the content, possibly partial, of a snapshot with the given id The branches of the snapshot are iterated in the lexicographical order of their names. Args: snapshot_id (bytes): identifier of the snapshot branches_from (bytes): optional parameter used to skip branches whose name is lesser than it before returning them branches_count (int): optional parameter used to restrain the amount of returned branches target_types (list): optional parameter used to filter the target types of branch to return (possible values that can be contained in that list are `'content', 'directory', 'revision', 'release', 'snapshot', 'alias'`) Returns: dict: None if the snapshot does not exist; a dict with three keys otherwise: * **id**: identifier of the snapshot * **branches**: a dict of branches contained in the snapshot whose keys are the branches' names. * **next_branch**: the name of the first branch not returned or :const:`None` if the snapshot has less than `branches_count` branches after `branches_from` included. """ ... @remote_api_endpoint('snapshot/get_random') def snapshot_get_random(self): """Finds a random snapshot id. Returns: a sha1_git """ ... @remote_api_endpoint('origin/visit/add') def origin_visit_add(self, origin, date, type): """Add an origin_visit for the origin at ts with status 'ongoing'. Args: origin (str): visited origin's identifier or URL date (Union[str,datetime]): timestamp of such visit type (str): the type of loader used for the visit (hg, git, ...) Returns: dict: dictionary with keys origin and visit where: - origin: origin identifier - visit: the visit identifier for the new visit occurrence """ ... @remote_api_endpoint('origin/visit/update') def origin_visit_update(self, origin, visit_id, status=None, metadata=None, snapshot=None): """Update an origin_visit's status. Args: origin (str): visited origin's URL visit_id: Visit's id status: Visit's new status metadata: Data associated to the visit snapshot (sha1_git): identifier of the snapshot to add to the visit Returns: None """ ... @remote_api_endpoint('origin/visit/upsert') def origin_visit_upsert(self, visits): """Add a origin_visits with a specific id and with all its data. If there is already an origin_visit with the same `(origin_id, visit_id)`, overwrites it. Args: visits: iterable of dicts with keys: - **origin**: dict with keys either `id` or `url` - **visit**: origin visit id - **date**: timestamp of such visit - **status**: Visit's new status - **metadata**: Data associated to the visit - **snapshot**: identifier of the snapshot to add to the visit """ ... @remote_api_endpoint('origin/visit/get') def origin_visit_get(self, origin, last_visit=None, limit=None): """Retrieve all the origin's visit's information. Args: origin (str): The visited origin last_visit: Starting point from which listing the next visits Default to None limit (int): Number of results to return from the last visit. Default to None Yields: List of visits. """ ... @remote_api_endpoint('origin/visit/find_by_date') def origin_visit_find_by_date(self, origin, visit_date): """Retrieves the origin visit whose date is closest to the provided timestamp. In case of a tie, the visit with largest id is selected. Args: origin (str): The occurrence's origin (URL). target (datetime): target timestamp Returns: A visit. """ ... @remote_api_endpoint('origin/visit/getby') def origin_visit_get_by(self, origin, visit): """Retrieve origin visit's information. Args: origin: The occurrence's origin (identifier). Returns: The information on that particular (origin, visit) or None if it does not exist """ ... @remote_api_endpoint('origin/visit/get_latest') def origin_visit_get_latest( self, origin, allowed_statuses=None, require_snapshot=False): """Get the latest origin visit for the given origin, optionally looking only for those with one of the given allowed_statuses or for those with a known snapshot. Args: origin (str): the origin's URL allowed_statuses (list of str): list of visit statuses considered to find the latest visit. For instance, ``allowed_statuses=['full']`` will only consider visits that have successfully run to completion. require_snapshot (bool): If True, only a visit with a snapshot will be returned. Returns: dict: a dict with the following keys: - **origin**: the URL of the origin - **visit**: origin visit id - **type**: type of loader used for the visit - **date**: timestamp of such visit - **status**: Visit's new status - **metadata**: Data associated to the visit - **snapshot** (Optional[sha1_git]): identifier of the snapshot associated to the visit """ ... @remote_api_endpoint('origin/visit/get_random') def origin_visit_get_random( self, type: str) -> Optional[Dict[str, Any]]: """Randomly select one successful origin visit with made in the last 3 months. Returns: dict representing an origin visit, in the same format as :py:meth:`origin_visit_get`. """ ... @remote_api_endpoint('object/find_by_sha1_git') def object_find_by_sha1_git(self, ids): """Return the objects found with the given ids. Args: ids: a generator of sha1_gits Returns: dict: a mapping from id to the list of objects found. Each object found is itself a dict with keys: - sha1_git: the input id - type: the type of object found """ ... @remote_api_endpoint('origin/get') def origin_get(self, origins): """Return origins, either all identified by their ids or all identified by tuples (type, url). If the url is given and the type is omitted, one of the origins with that url is returned. Args: origin: a list of dictionaries representing the individual origins to find. These dicts have the key url: - url (bytes): the url the origin points to Returns: dict: the origin dictionary with the keys: - id: origin's id - url: origin's url Raises: ValueError: if the url or the id don't exist. """ ... @remote_api_endpoint('origin/get_sha1') def origin_get_by_sha1(self, sha1s): """Return origins, identified by the sha1 of their URLs. Args: sha1s (list[bytes]): a list of sha1s Yields: dicts containing origin information as returned by :meth:`swh.storage.storage.Storage.origin_get`, or None if an origin matching the sha1 is not found. """ ... @deprecated @remote_api_endpoint('origin/get_range') def origin_get_range(self, origin_from=1, origin_count=100): """Retrieve ``origin_count`` origins whose ids are greater or equal than ``origin_from``. Origins are sorted by id before retrieving them. Args: origin_from (int): the minimum id of origins to retrieve origin_count (int): the maximum number of origins to retrieve Yields: dicts containing origin information as returned by :meth:`swh.storage.storage.Storage.origin_get`. """ ... @remote_api_endpoint('origin/list') def origin_list( self, page_token: Optional[str] = None, limit: int = 100) -> dict: """Returns the list of origins Args: page_token: opaque token used for pagination. limit: the maximum number of results to return Returns: dict: dict with the following keys: - **next_page_token** (str, optional): opaque token to be used as `page_token` for retrieving the next page. if absent, there is no more pages to gather. - **origins** (List[dict]): list of origins, as returned by `origin_get`. """ ... @remote_api_endpoint('origin/search') def origin_search(self, url_pattern, offset=0, limit=50, regexp=False, with_visit=False): """Search for origins whose urls contain a provided string pattern or match a provided regular expression. The search is performed in a case insensitive way. Args: url_pattern (str): the string pattern to search for in origin urls offset (int): number of found origins to skip before returning results limit (int): the maximum number of found origins to return regexp (bool): if True, consider the provided pattern as a regular expression and return origins whose urls match it with_visit (bool): if True, filter out origins with no visit Yields: dicts containing origin information as returned by :meth:`swh.storage.storage.Storage.origin_get`. """ ... @deprecated @remote_api_endpoint('origin/count') def origin_count(self, url_pattern, regexp=False, with_visit=False): """Count origins whose urls contain a provided string pattern or match a provided regular expression. The pattern search in origin urls is performed in a case insensitive way. Args: url_pattern (str): the string pattern to search for in origin urls regexp (bool): if True, consider the provided pattern as a regular expression and return origins whose urls match it with_visit (bool): if True, filter out origins with no visit Returns: int: The number of origins matching the search criterion. """ ... @remote_api_endpoint('origin/add_multi') def origin_add(self, origins): """Add origins to the storage Args: origins: list of dictionaries representing the individual origins, with the following keys: - type: the origin type ('git', 'svn', 'deb', ...) - url (bytes): the url the origin points to Returns: list: given origins as dict updated with their id """ ... @remote_api_endpoint('origin/add') def origin_add_one(self, origin): """Add origin to the storage Args: origin: dictionary representing the individual origin to add. This dict has the following keys: - type (FIXME: enum TBD): the origin type ('git', 'wget', ...) - url (bytes): the url the origin points to Returns: the id of the added origin, or of the identical one that already exists. """ ... def stat_counters(self): """compute statistics about the number of tuples in various tables Returns: dict: a dictionary mapping textual labels (e.g., content) to integer values (e.g., the number of tuples in table content) """ ... def refresh_stat_counters(self): """Recomputes the statistics for `stat_counters`.""" ... @remote_api_endpoint('origin/metadata/add') def origin_metadata_add(self, origin_url, ts, provider, tool, metadata): """ Add an origin_metadata for the origin at ts with provenance and metadata. Args: origin_url (str): the origin url for which the metadata is added ts (datetime): timestamp of the found metadata provider (int): the provider of metadata (ex:'hal') tool (int): tool used to extract metadata metadata (jsonb): the metadata retrieved at the time and location """ ... @remote_api_endpoint('origin/metadata/get') def origin_metadata_get_by(self, origin_url, provider_type=None): """Retrieve list of all origin_metadata entries for the origin_id Args: origin_url (str): the origin's URL provider_type (str): (optional) type of provider Returns: list of dicts: the origin_metadata dictionary with the keys: - origin_id (int): origin's id - discovery_date (datetime): timestamp of discovery - tool_id (int): metadata's extracting tool - metadata (jsonb) - provider_id (int): metadata's provider - provider_name (str) - provider_type (str) - provider_url (str) """ ... @remote_api_endpoint('tool/add') def tool_add(self, tools): """Add new tools to the storage. Args: tools (iterable of :class:`dict`): Tool information to add to storage. Each tool is a :class:`dict` with the following keys: - name (:class:`str`): name of the tool - version (:class:`str`): version of the tool - configuration (:class:`dict`): configuration of the tool, must be json-encodable Returns: :class:`dict`: All the tools inserted in storage (including the internal ``id``). The order of the list is not guaranteed to match the order of the initial list. """ ... @remote_api_endpoint('tool/data') def tool_get(self, tool): """Retrieve tool information. Args: tool (dict): Tool information we want to retrieve from storage. The dicts have the same keys as those used in :func:`tool_add`. Returns: dict: The full tool information if it exists (``id`` included), None otherwise. """ ... @remote_api_endpoint('provider/add') def metadata_provider_add(self, provider_name, provider_type, provider_url, metadata): """Add a metadata provider. Args: provider_name (str): Its name provider_type (str): Its type (eg. `'deposit-client'`) provider_url (str): Its URL metadata: JSON-encodable object Returns: int: an identifier of the provider """ ... @remote_api_endpoint('provider/get') def metadata_provider_get(self, provider_id): """Get a metadata provider Args: provider_id: Its identifier, as given by `metadata_provider_add`. Returns: dict: same as `metadata_provider_add`; or None if it does not exist. """ ... @remote_api_endpoint('provider/getby') def metadata_provider_get_by(self, provider): """Get a metadata provider Args: provider (dict): A dictionary with keys: * provider_name: Its name * provider_url: Its URL Returns: dict: same as `metadata_provider_add`; or None if it does not exist. """ ... @deprecated @remote_api_endpoint('algos/diff_directories') def diff_directories(self, from_dir, to_dir, track_renaming=False): """Compute the list of file changes introduced between two arbitrary directories (insertion / deletion / modification / renaming of files). Args: from_dir (bytes): identifier of the directory to compare from to_dir (bytes): identifier of the directory to compare to track_renaming (bool): whether or not to track files renaming Returns: A list of dict describing the introduced file changes (see :func:`swh.storage.algos.diff.diff_directories` for more details). """ ... @deprecated @remote_api_endpoint('algos/diff_revisions') def diff_revisions(self, from_rev, to_rev, track_renaming=False): """Compute the list of file changes introduced between two arbitrary revisions (insertion / deletion / modification / renaming of files). Args: from_rev (bytes): identifier of the revision to compare from to_rev (bytes): identifier of the revision to compare to track_renaming (bool): whether or not to track files renaming Returns: A list of dict describing the introduced file changes (see :func:`swh.storage.algos.diff.diff_directories` for more details). """ ... @deprecated @remote_api_endpoint('algos/diff_revision') def diff_revision(self, revision, track_renaming=False): """Compute the list of file changes introduced by a specific revision (insertion / deletion / modification / renaming of files) by comparing it against its first parent. Args: revision (bytes): identifier of the revision from which to compute the list of files changes track_renaming (bool): whether or not to track files renaming Returns: A list of dict describing the introduced file changes (see :func:`swh.storage.algos.diff.diff_directories` for more details). """ ... diff --git a/swh/storage/retry.py b/swh/storage/retry.py index 9aa29115..af72a49d 100644 --- a/swh/storage/retry.py +++ b/swh/storage/retry.py @@ -1,139 +1,144 @@ # Copyright (C) 2019-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import logging import psycopg2 import traceback from datetime import datetime from typing import Dict, Iterable, List, Optional, Union from requests.exceptions import ConnectionError from tenacity import ( retry, stop_after_attempt, wait_random_exponential, retry_if_exception_type ) from swh.storage import get_storage, HashCollision logger = logging.getLogger(__name__) RETRY_EXCEPTIONS = [ # raised when two parallel insertions insert the same data psycopg2.IntegrityError, HashCollision, # when the server is restarting ConnectionError, ] def should_retry_adding(error: Exception) -> bool: """Retry if the error/exception if one of the RETRY_EXCEPTIONS type. """ for exc in RETRY_EXCEPTIONS: if retry_if_exception_type(exc)(error): error_name = error.__module__ + '.' + error.__class__.__name__ logger.warning('Retry adding a batch', exc_info=False, extra={ 'swh_type': 'storage_retry', 'swh_exception_type': error_name, 'swh_exception': traceback.format_exc(), }) return True return False swh_retry = retry(retry=should_retry_adding, wait=wait_random_exponential(multiplier=1, max=10), stop=stop_after_attempt(3)) class RetryingProxyStorage: """Storage implementation which retries adding objects when it specifically fails (hash collision, integrity error). """ def __init__(self, storage): self.storage = get_storage(**storage) def __getattr__(self, key): return getattr(self.storage, key) @swh_retry - def content_add(self, content: List[Dict]) -> Dict: + def content_add(self, content: Iterable[Dict]) -> Dict: contents = list(content) return self.storage.content_add(contents) @swh_retry - def content_add_metadata(self, content: List[Dict]) -> Dict: + def content_add_metadata(self, content: Iterable[Dict]) -> Dict: contents = list(content) return self.storage.content_add_metadata(contents) + @swh_retry + def skipped_content_add(self, content: Iterable[Dict]) -> Dict: + contents = list(content) + return self.storage.skipped_content_add(contents) + @swh_retry def origin_add_one(self, origin: Dict) -> str: return self.storage.origin_add_one(origin) @swh_retry def origin_visit_add(self, origin: Dict, date: Union[datetime, str], type: str) -> Dict: return self.storage.origin_visit_add(origin, date, type) @swh_retry def origin_visit_update( self, origin: str, visit_id: int, status: Optional[str] = None, metadata: Optional[Dict] = None, snapshot: Optional[Dict] = None) -> Dict: return self.storage.origin_visit_update( origin, visit_id, status=status, metadata=metadata, snapshot=snapshot) @swh_retry - def tool_add(self, tools: List[Dict]) -> List[Dict]: + def tool_add(self, tools: Iterable[Dict]) -> List[Dict]: tools = list(tools) return self.storage.tool_add(tools) @swh_retry def metadata_provider_add( self, provider_name: str, provider_type: str, provider_url: str, metadata: Dict) -> Union[str, int]: return self.storage.metadata_provider_add( provider_name, provider_type, provider_url, metadata) @swh_retry def origin_metadata_add( self, origin_url: str, ts: Union[str, datetime], provider_id: int, tool_id: int, metadata: Dict) -> None: return self.storage.origin_metadata_add( origin_url, ts, provider_id, tool_id, metadata) @swh_retry - def directory_add(self, directories: List[Dict]) -> Dict: + def directory_add(self, directories: Iterable[Dict]) -> Dict: directories = list(directories) return self.storage.directory_add(directories) @swh_retry - def revision_add(self, revisions: List[Dict]) -> Dict: + def revision_add(self, revisions: Iterable[Dict]) -> Dict: revisions = list(revisions) return self.storage.revision_add(revisions) @swh_retry - def release_add(self, releases: List[Dict]) -> Dict: + def release_add(self, releases: Iterable[Dict]) -> Dict: releases = list(releases) return self.storage.release_add(releases) @swh_retry - def snapshot_add(self, snapshot: List[Dict]) -> Dict: + def snapshot_add(self, snapshot: Iterable[Dict]) -> Dict: snapshots = list(snapshot) return self.storage.snapshot_add(snapshots) @swh_retry def flush(self, object_types: Optional[Iterable[str]] = None) -> Dict: """Specific case for buffer proxy storage failing to flush data """ if hasattr(self.storage, 'flush'): return self.storage.flush(object_types) return {} diff --git a/swh/storage/storage.py b/swh/storage/storage.py index 5a919934..53bd9817 100644 --- a/swh/storage/storage.py +++ b/swh/storage/storage.py @@ -1,1162 +1,1180 @@ # Copyright (C) 2015-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import copy import datetime import itertools import json from collections import defaultdict from concurrent.futures import ThreadPoolExecutor from contextlib import contextmanager from typing import Any, Dict, List, Optional import dateutil.parser import psycopg2 import psycopg2.pool from swh.model.model import SHA1_SIZE from swh.model.hashutil import ALGORITHMS, hash_to_bytes, hash_to_hex from swh.objstorage import get_objstorage from swh.objstorage.exc import ObjNotFoundError try: from swh.journal.writer import get_journal_writer except ImportError: get_journal_writer = None # type: ignore # mypy limitation, see https://github.com/python/mypy/issues/1153 from . import converters from .common import db_transaction_generator, db_transaction from .db import Db from .exc import StorageDBError from .algos import diff from .metrics import timed, send_metric, process_metrics from .utils import get_partition_bounds_bytes # Max block size of contents to return BULK_BLOCK_CONTENT_LEN_MAX = 10000 EMPTY_SNAPSHOT_ID = hash_to_bytes('1a8893e6a86f444e8be8e7bda6cb34fb1735a00e') """Identifier for the empty snapshot""" class Storage(): """SWH storage proxy, encompassing DB and object storage """ def __init__(self, db, objstorage, min_pool_conns=1, max_pool_conns=10, journal_writer=None): """ Args: db_conn: either a libpq connection string, or a psycopg2 connection obj_root: path to the root of the object storage """ try: if isinstance(db, psycopg2.extensions.connection): self._pool = None self._db = Db(db) else: self._pool = psycopg2.pool.ThreadedConnectionPool( min_pool_conns, max_pool_conns, db ) self._db = None except psycopg2.OperationalError as e: raise StorageDBError(e) self.objstorage = get_objstorage(**objstorage) if journal_writer: if get_journal_writer is None: raise EnvironmentError( 'You need the swh.journal package to use the ' 'journal_writer feature') self.journal_writer = get_journal_writer(**journal_writer) else: self.journal_writer = None def get_db(self): if self._db: return self._db else: return Db.from_pool(self._pool) def put_db(self, db): if db is not self._db: db.put_conn() @contextmanager def db(self): db = None try: db = self.get_db() yield db finally: if db: self.put_db(db) @timed @db_transaction() def check_config(self, *, check_write, db=None, cur=None): if not self.objstorage.check_config(check_write=check_write): return False # Check permissions on one of the tables if check_write: check = 'INSERT' else: check = 'SELECT' cur.execute( "select has_table_privilege(current_user, 'content', %s)", (check,) ) return cur.fetchone()[0] def _content_unique_key(self, hash, db): """Given a hash (tuple or dict), return a unique key from the aggregation of keys. """ keys = db.content_hash_keys if isinstance(hash, tuple): return hash return tuple([hash[k] for k in keys]) @staticmethod - def _normalize_content(d): + def _content_normalize(d): d = d.copy() if 'status' not in d: d['status'] = 'visible' - if 'length' not in d: - d['length'] = -1 - return d @staticmethod - def _validate_content(d): + def _content_validate(d): """Sanity checks on status / reason / length, that postgresql doesn't enforce.""" - if d['status'] not in ('visible', 'absent', 'hidden'): + if d['status'] not in ('visible', 'hidden'): raise ValueError('Invalid content status: {}'.format(d['status'])) - if d['status'] != 'absent' and d.get('reason') is not None: + if d.get('reason') is not None: raise ValueError( - 'Must not provide a reason if content is not absent.') + 'Must not provide a reason if content is present.') - if d['length'] < -1: - raise ValueError('Content length must be positive or -1.') + if d['length'] is None or d['length'] < 0: + raise ValueError('Content length must be positive.') - def _filter_new_content(self, content, db=None, cur=None): - """Sort contents into buckets 'with data' and 'without data', - and filter out those already in the database.""" - content_by_status = defaultdict(list) - for d in content: - content_by_status[d['status']].append(d) - - content_with_data = content_by_status['visible'] \ - + content_by_status['hidden'] - content_without_data = content_by_status['absent'] - - missing_content = set(self.content_missing(content_with_data, - db=db, cur=cur)) - missing_skipped = set(self._content_unique_key(hashes, db) - for hashes in self.skipped_content_missing( - content_without_data, db=db, cur=cur)) - - content_with_data = [ - cont for cont in content_with_data - if cont['sha1'] in missing_content] - content_without_data = [ - cont for cont in content_without_data - if self._content_unique_key(cont, db) in missing_skipped] - - summary = { - 'content:add': len(missing_content), - 'skipped_content:add': len(missing_skipped), - } - - return (content_with_data, content_without_data, summary) - - def _content_add_metadata(self, db, cur, - content_with_data, content_without_data): + def _content_add_metadata(self, db, cur, content): """Add content to the postgresql database but not the object storage. """ - if content_with_data: - # create temporary table for metadata injection - db.mktemp('content', cur) + # create temporary table for metadata injection + db.mktemp('content', cur) - db.copy_to(content_with_data, 'tmp_content', - db.content_add_keys, cur) + db.copy_to(content, 'tmp_content', + db.content_add_keys, cur) - # move metadata in place - try: - db.content_add_from_temp(cur) - except psycopg2.IntegrityError as e: - from . import HashCollision - if e.diag.sqlstate == '23505' and \ - e.diag.table_name == 'content': - constraint_to_hash_name = { - 'content_pkey': 'sha1', - 'content_sha1_git_idx': 'sha1_git', - 'content_sha256_idx': 'sha256', - } - colliding_hash_name = constraint_to_hash_name \ - .get(e.diag.constraint_name) - raise HashCollision(colliding_hash_name) from None - else: - raise - - if content_without_data: - content_without_data = \ - [cont.copy() for cont in content_without_data] - origin_ids = db.origin_id_get_by_url( - [cont.get('origin') for cont in content_without_data], - cur=cur) - for (cont, origin_id) in zip(content_without_data, origin_ids): - if 'origin' in cont: - cont['origin'] = origin_id - db.mktemp('skipped_content', cur) - db.copy_to(content_without_data, 'tmp_skipped_content', - db.skipped_content_keys, cur) - - # move metadata in place - db.skipped_content_add_from_temp(cur) + # move metadata in place + try: + db.content_add_from_temp(cur) + except psycopg2.IntegrityError as e: + from . import HashCollision + if e.diag.sqlstate == '23505' and \ + e.diag.table_name == 'content': + constraint_to_hash_name = { + 'content_pkey': 'sha1', + 'content_sha1_git_idx': 'sha1_git', + 'content_sha256_idx': 'sha256', + } + colliding_hash_name = constraint_to_hash_name \ + .get(e.diag.constraint_name) + raise HashCollision(colliding_hash_name) from None + else: + raise @timed @process_metrics @db_transaction() def content_add(self, content, db=None, cur=None): content = [dict(c.items()) for c in content] # semi-shallow copy now = datetime.datetime.now(tz=datetime.timezone.utc) for item in content: item['ctime'] = now - content = [self._normalize_content(c) for c in content] + content = [self._content_normalize(c) for c in content] for c in content: - self._validate_content(c) + self._content_validate(c) - (content_with_data, content_without_data, summary) = \ - self._filter_new_content(content, db, cur) + missing = list(self.content_missing(content, key_hash='sha1_git')) + content = [c for c in content if c['sha1_git'] in missing] if self.journal_writer: - for item in content_with_data: + for item in content: if 'data' in item: item = item.copy() del item['data'] self.journal_writer.write_addition('content', item) - for item in content_without_data: - self.journal_writer.write_addition('content', item) def add_to_objstorage(): """Add to objstorage the new missing_content Returns: Sum of all the content's data length pushed to the objstorage. Content present twice is only sent once. """ content_bytes_added = 0 data = {} - for cont in content_with_data: + for cont in content: if cont['sha1'] not in data: data[cont['sha1']] = cont['data'] content_bytes_added += max(0, cont['length']) # FIXME: Since we do the filtering anyway now, we might as # well make the objstorage's add_batch call return what we # want here (real bytes added)... that'd simplify this... self.objstorage.add_batch(data) return content_bytes_added with ThreadPoolExecutor(max_workers=1) as executor: added_to_objstorage = executor.submit(add_to_objstorage) - self._content_add_metadata( - db, cur, content_with_data, content_without_data) + self._content_add_metadata(db, cur, content) # Wait for objstorage addition before returning from the # transaction, bubbling up any exception content_bytes_added = added_to_objstorage.result() - summary['content:add:bytes'] = content_bytes_added - return summary + return { + 'content:add': len(content), + 'content:add:bytes': content_bytes_added, + } @timed @db_transaction() def content_update(self, content, keys=[], db=None, cur=None): # TODO: Add a check on input keys. How to properly implement # this? We don't know yet the new columns. if self.journal_writer: raise NotImplementedError( 'content_update is not yet supported with a journal_writer.') db.mktemp('content', cur) select_keys = list(set(db.content_get_metadata_keys).union(set(keys))) db.copy_to(content, 'tmp_content', select_keys, cur) db.content_update_from_temp(keys_to_update=keys, cur=cur) @timed @process_metrics @db_transaction() def content_add_metadata(self, content, db=None, cur=None): - content = [self._normalize_content(c) for c in content] + content = [self._content_normalize(c) for c in content] for c in content: - self._validate_content(c) + self._content_validate(c) - (content_with_data, content_without_data, summary) = \ - self._filter_new_content(content, db, cur) + missing = self.content_missing(content, key_hash='sha1_git') + content = [c for c in content if c['sha1_git'] in missing] if self.journal_writer: - for item in itertools.chain(content_with_data, - content_without_data): + for item in itertools.chain(content): assert 'data' not in content self.journal_writer.write_addition('content', item) - self._content_add_metadata( - db, cur, content_with_data, content_without_data) + self._content_add_metadata(db, cur, content) - return summary + return { + 'content:add': len(content), + } @timed def content_get(self, content): # FIXME: Make this method support slicing the `data`. if len(content) > BULK_BLOCK_CONTENT_LEN_MAX: raise ValueError( "Send at maximum %s contents." % BULK_BLOCK_CONTENT_LEN_MAX) for obj_id in content: try: data = self.objstorage.get(obj_id) except ObjNotFoundError: yield None continue yield {'sha1': obj_id, 'data': data} @timed @db_transaction() def content_get_range(self, start, end, limit=1000, db=None, cur=None): if limit is None: raise ValueError('Development error: limit should not be None') contents = [] next_content = None for counter, content_row in enumerate( db.content_get_range(start, end, limit+1, cur)): content = dict(zip(db.content_get_metadata_keys, content_row)) if counter >= limit: # take the last commit for the next page starting from this next_content = content['sha1'] break contents.append(content) return { 'contents': contents, 'next': next_content, } @timed @db_transaction() def content_get_partition( self, partition_id: int, nb_partitions: int, limit: int = 1000, page_token: str = None, db=None, cur=None): if limit is None: raise ValueError('Development error: limit should not be None') (start, end) = get_partition_bounds_bytes( partition_id, nb_partitions, SHA1_SIZE) if page_token: start = hash_to_bytes(page_token) if end is None: end = b'\xff'*SHA1_SIZE result = self.content_get_range(start, end, limit) result2 = { 'contents': result['contents'], 'next_page_token': None, } if result['next']: result2['next_page_token'] = hash_to_hex(result['next']) return result2 @timed @db_transaction(statement_timeout=500) def content_get_metadata( self, contents: List[bytes], db=None, cur=None) -> Dict[bytes, List[Dict]]: result: Dict[bytes, List[Dict]] = {sha1: [] for sha1 in contents} for row in db.content_get_metadata_from_sha1s(contents, cur): content_meta = dict(zip(db.content_get_metadata_keys, row)) result[content_meta['sha1']].append(content_meta) return result @timed @db_transaction_generator() def content_missing(self, content, key_hash='sha1', db=None, cur=None): keys = db.content_hash_keys if key_hash not in keys: raise ValueError("key_hash should be one of %s" % keys) key_hash_idx = keys.index(key_hash) if not content: return for obj in db.content_missing_from_list(content, cur): yield obj[key_hash_idx] @timed @db_transaction_generator() def content_missing_per_sha1(self, contents, db=None, cur=None): for obj in db.content_missing_per_sha1(contents, cur): yield obj[0] @timed @db_transaction_generator() def content_missing_per_sha1_git(self, contents, db=None, cur=None): for obj in db.content_missing_per_sha1_git(contents, cur): yield obj[0] - @timed - @db_transaction_generator() - def skipped_content_missing(self, contents, db=None, cur=None): - for content in db.skipped_content_missing(contents, cur): - yield dict(zip(db.content_hash_keys, content)) - @timed @db_transaction() def content_find(self, content, db=None, cur=None): if not set(content).intersection(ALGORITHMS): raise ValueError('content keys must contain at least one of: ' 'sha1, sha1_git, sha256, blake2s256') contents = db.content_find(sha1=content.get('sha1'), sha1_git=content.get('sha1_git'), sha256=content.get('sha256'), blake2s256=content.get('blake2s256'), cur=cur) return [dict(zip(db.content_find_cols, content)) for content in contents] @timed @db_transaction() def content_get_random(self, db=None, cur=None): return db.content_get_random(cur) + @staticmethod + def _skipped_content_normalize(d): + d = d.copy() + + if d.get('status') is None: + d['status'] = 'absent' + + if d.get('length') is None: + d['length'] = -1 + + return d + + @staticmethod + def _skipped_content_validate(d): + """Sanity checks on status / reason / length, that postgresql + doesn't enforce.""" + if d['status'] != 'absent': + raise ValueError('Invalid content status: {}'.format(d['status'])) + + if d.get('reason') is None: + raise ValueError( + 'Must provide a reason if content is absent.') + + if d['length'] < -1: + raise ValueError('Content length must be positive or -1.') + + def _skipped_content_add_metadata(self, db, cur, content): + content = \ + [cont.copy() for cont in content] + origin_ids = db.origin_id_get_by_url( + [cont.get('origin') for cont in content], + cur=cur) + for (cont, origin_id) in zip(content, origin_ids): + if 'origin' in cont: + cont['origin'] = origin_id + db.mktemp('skipped_content', cur) + db.copy_to(content, 'tmp_skipped_content', + db.skipped_content_keys, cur) + + # move metadata in place + db.skipped_content_add_from_temp(cur) + + @timed + @process_metrics + @db_transaction() + def skipped_content_add(self, content, db=None, cur=None): + content = [dict(c.items()) for c in content] # semi-shallow copy + now = datetime.datetime.now(tz=datetime.timezone.utc) + for item in content: + item['ctime'] = now + + content = [self._skipped_content_normalize(c) for c in content] + for c in content: + self._skipped_content_validate(c) + + missing_contents = self.skipped_content_missing(content) + content = [c for c in content + if any(all(c.get(algo) == missing_content.get(algo) + for algo in ALGORITHMS) + for missing_content in missing_contents)] + + if self.journal_writer: + for item in content: + self.journal_writer.write_addition('content', item) + + self._skipped_content_add_metadata(db, cur, content) + + return { + 'skipped_content:add': len(content), + } + + @timed + @db_transaction_generator() + def skipped_content_missing(self, contents, db=None, cur=None): + for content in db.skipped_content_missing(contents, cur): + yield dict(zip(db.content_hash_keys, content)) + @timed @process_metrics @db_transaction() def directory_add(self, directories, db=None, cur=None): directories = list(directories) summary = {'directory:add': 0} dirs = set() dir_entries = { 'file': defaultdict(list), 'dir': defaultdict(list), 'rev': defaultdict(list), } for cur_dir in directories: dir_id = cur_dir['id'] dirs.add(dir_id) for src_entry in cur_dir['entries']: entry = src_entry.copy() entry['dir_id'] = dir_id if entry['type'] not in ('file', 'dir', 'rev'): raise ValueError( 'Entry type must be file, dir, or rev; not %s' % entry['type']) dir_entries[entry['type']][dir_id].append(entry) dirs_missing = set(self.directory_missing(dirs, db=db, cur=cur)) if not dirs_missing: return summary if self.journal_writer: self.journal_writer.write_additions( 'directory', (dir_ for dir_ in directories if dir_['id'] in dirs_missing)) # Copy directory ids dirs_missing_dict = ({'id': dir} for dir in dirs_missing) db.mktemp('directory', cur) db.copy_to(dirs_missing_dict, 'tmp_directory', ['id'], cur) # Copy entries for entry_type, entry_list in dir_entries.items(): entries = itertools.chain.from_iterable( entries_for_dir for dir_id, entries_for_dir in entry_list.items() if dir_id in dirs_missing) db.mktemp_dir_entry(entry_type) db.copy_to( entries, 'tmp_directory_entry_%s' % entry_type, ['target', 'name', 'perms', 'dir_id'], cur, ) # Do the final copy db.directory_add_from_temp(cur) summary['directory:add'] = len(dirs_missing) return summary @timed @db_transaction_generator() def directory_missing(self, directories, db=None, cur=None): for obj in db.directory_missing_from_list(directories, cur): yield obj[0] @timed @db_transaction_generator(statement_timeout=20000) def directory_ls(self, directory, recursive=False, db=None, cur=None): if recursive: res_gen = db.directory_walk(directory, cur=cur) else: res_gen = db.directory_walk_one(directory, cur=cur) for line in res_gen: yield dict(zip(db.directory_ls_cols, line)) @timed @db_transaction(statement_timeout=2000) def directory_entry_get_by_path(self, directory, paths, db=None, cur=None): res = db.directory_entry_get_by_path(directory, paths, cur) if res: return dict(zip(db.directory_ls_cols, res)) @timed @db_transaction() def directory_get_random(self, db=None, cur=None): return db.directory_get_random(cur) @timed @process_metrics @db_transaction() def revision_add(self, revisions, db=None, cur=None): revisions = list(revisions) summary = {'revision:add': 0} revisions_missing = set(self.revision_missing( set(revision['id'] for revision in revisions), db=db, cur=cur)) if not revisions_missing: return summary db.mktemp_revision(cur) revisions_filtered = [ revision for revision in revisions if revision['id'] in revisions_missing] if self.journal_writer: self.journal_writer.write_additions('revision', revisions_filtered) revisions_filtered = map(converters.revision_to_db, revisions_filtered) parents_filtered = [] db.copy_to( revisions_filtered, 'tmp_revision', db.revision_add_cols, cur, lambda rev: parents_filtered.extend(rev['parents'])) db.revision_add_from_temp(cur) db.copy_to(parents_filtered, 'revision_history', ['id', 'parent_id', 'parent_rank'], cur) return {'revision:add': len(revisions_missing)} @timed @db_transaction_generator() def revision_missing(self, revisions, db=None, cur=None): if not revisions: return for obj in db.revision_missing_from_list(revisions, cur): yield obj[0] @timed @db_transaction_generator(statement_timeout=1000) def revision_get(self, revisions, db=None, cur=None): for line in db.revision_get_from_list(revisions, cur): data = converters.db_to_revision( dict(zip(db.revision_get_cols, line)) ) if not data['type']: yield None continue yield data @timed @db_transaction_generator(statement_timeout=2000) def revision_log(self, revisions, limit=None, db=None, cur=None): for line in db.revision_log(revisions, limit, cur): data = converters.db_to_revision( dict(zip(db.revision_get_cols, line)) ) if not data['type']: yield None continue yield data @timed @db_transaction_generator(statement_timeout=2000) def revision_shortlog(self, revisions, limit=None, db=None, cur=None): yield from db.revision_shortlog(revisions, limit, cur) @timed @db_transaction() def revision_get_random(self, db=None, cur=None): return db.revision_get_random(cur) @timed @process_metrics @db_transaction() def release_add(self, releases, db=None, cur=None): releases = list(releases) summary = {'release:add': 0} release_ids = set(release['id'] for release in releases) releases_missing = set(self.release_missing(release_ids, db=db, cur=cur)) if not releases_missing: return summary db.mktemp_release(cur) releases_missing = list(releases_missing) releases_filtered = [ release for release in releases if release['id'] in releases_missing ] if self.journal_writer: self.journal_writer.write_additions('release', releases_filtered) releases_filtered = map(converters.release_to_db, releases_filtered) db.copy_to(releases_filtered, 'tmp_release', db.release_add_cols, cur) db.release_add_from_temp(cur) return {'release:add': len(releases_missing)} @timed @db_transaction_generator() def release_missing(self, releases, db=None, cur=None): if not releases: return for obj in db.release_missing_from_list(releases, cur): yield obj[0] @timed @db_transaction_generator(statement_timeout=500) def release_get(self, releases, db=None, cur=None): for release in db.release_get_from_list(releases, cur): data = converters.db_to_release( dict(zip(db.release_get_cols, release)) ) yield data if data['target_type'] else None @timed @db_transaction() def release_get_random(self, db=None, cur=None): return db.release_get_random(cur) @timed @process_metrics @db_transaction() def snapshot_add(self, snapshots, db=None, cur=None): created_temp_table = False count = 0 for snapshot in snapshots: if not db.snapshot_exists(snapshot['id'], cur): if not created_temp_table: db.mktemp_snapshot_branch(cur) created_temp_table = True db.copy_to( ( { 'name': name, 'target': info['target'] if info else None, 'target_type': (info['target_type'] if info else None), } for name, info in snapshot['branches'].items() ), 'tmp_snapshot_branch', ['name', 'target', 'target_type'], cur, ) if self.journal_writer: self.journal_writer.write_addition('snapshot', snapshot) db.snapshot_add(snapshot['id'], cur) count += 1 return {'snapshot:add': count} @timed @db_transaction_generator() def snapshot_missing(self, snapshots, db=None, cur=None): for obj in db.snapshot_missing_from_list(snapshots, cur): yield obj[0] @timed @db_transaction(statement_timeout=2000) def snapshot_get(self, snapshot_id, db=None, cur=None): return self.snapshot_get_branches(snapshot_id, db=db, cur=cur) @timed @db_transaction(statement_timeout=2000) def snapshot_get_by_origin_visit(self, origin, visit, db=None, cur=None): snapshot_id = db.snapshot_get_by_origin_visit(origin, visit, cur) if snapshot_id: return self.snapshot_get(snapshot_id, db=db, cur=cur) return None @timed @db_transaction(statement_timeout=4000) def snapshot_get_latest(self, origin, allowed_statuses=None, db=None, cur=None): if isinstance(origin, int): origin = self.origin_get({'id': origin}, db=db, cur=cur) if not origin: return origin = origin['url'] origin_visit = self.origin_visit_get_latest( origin, allowed_statuses=allowed_statuses, require_snapshot=True, db=db, cur=cur) if origin_visit and origin_visit['snapshot']: snapshot = self.snapshot_get( origin_visit['snapshot'], db=db, cur=cur) if not snapshot: raise ValueError( 'last origin visit references an unknown snapshot') return snapshot @timed @db_transaction(statement_timeout=2000) def snapshot_count_branches(self, snapshot_id, db=None, cur=None): return dict([bc for bc in db.snapshot_count_branches(snapshot_id, cur)]) @timed @db_transaction(statement_timeout=2000) def snapshot_get_branches(self, snapshot_id, branches_from=b'', branches_count=1000, target_types=None, db=None, cur=None): if snapshot_id == EMPTY_SNAPSHOT_ID: return { 'id': snapshot_id, 'branches': {}, 'next_branch': None, } branches = {} next_branch = None fetched_branches = list(db.snapshot_get_by_id( snapshot_id, branches_from=branches_from, branches_count=branches_count+1, target_types=target_types, cur=cur, )) for branch in fetched_branches[:branches_count]: branch = dict(zip(db.snapshot_get_cols, branch)) del branch['snapshot_id'] name = branch.pop('name') if branch == {'target': None, 'target_type': None}: branch = None branches[name] = branch if len(fetched_branches) > branches_count: branch = dict(zip(db.snapshot_get_cols, fetched_branches[-1])) next_branch = branch['name'] if branches: return { 'id': snapshot_id, 'branches': branches, 'next_branch': next_branch, } return None @timed @db_transaction() def snapshot_get_random(self, db=None, cur=None): return db.snapshot_get_random(cur) @timed @db_transaction() def origin_visit_add(self, origin, date, type, db=None, cur=None): origin_url = origin if isinstance(date, str): # FIXME: Converge on iso8601 at some point date = dateutil.parser.parse(date) visit_id = db.origin_visit_add(origin_url, date, type, cur) if self.journal_writer: # We can write to the journal only after inserting to the # DB, because we want the id of the visit self.journal_writer.write_addition('origin_visit', { 'origin': origin_url, 'date': date, 'type': type, 'visit': visit_id, 'status': 'ongoing', 'metadata': None, 'snapshot': None}) send_metric('origin_visit:add', count=1, method_name='origin_visit') return { 'origin': origin_url, 'visit': visit_id, } @timed @db_transaction() def origin_visit_update(self, origin, visit_id, status=None, metadata=None, snapshot=None, db=None, cur=None): if not isinstance(origin, str): raise TypeError('origin must be a string, not %r' % (origin,)) origin_url = origin visit = db.origin_visit_get(origin_url, visit_id, cur=cur) if not visit: raise ValueError('Invalid visit_id for this origin.') visit = dict(zip(db.origin_visit_get_cols, visit)) updates = {} if status and status != visit['status']: updates['status'] = status if metadata and metadata != visit['metadata']: updates['metadata'] = metadata if snapshot and snapshot != visit['snapshot']: updates['snapshot'] = snapshot if updates: if self.journal_writer: self.journal_writer.write_update('origin_visit', { **visit, **updates}) db.origin_visit_update(origin_url, visit_id, updates, cur) @timed @db_transaction() def origin_visit_upsert(self, visits, db=None, cur=None): visits = copy.deepcopy(visits) for visit in visits: if isinstance(visit['date'], str): visit['date'] = dateutil.parser.parse(visit['date']) if not isinstance(visit['origin'], str): raise TypeError("visit['origin'] must be a string, not %r" % (visit['origin'],)) if self.journal_writer: for visit in visits: self.journal_writer.write_addition('origin_visit', visit) for visit in visits: # TODO: upsert them all in a single query db.origin_visit_upsert(**visit, cur=cur) @timed @db_transaction_generator(statement_timeout=500) def origin_visit_get(self, origin, last_visit=None, limit=None, db=None, cur=None): for line in db.origin_visit_get_all( origin, last_visit=last_visit, limit=limit, cur=cur): data = dict(zip(db.origin_visit_get_cols, line)) yield data @timed @db_transaction(statement_timeout=500) def origin_visit_find_by_date(self, origin, visit_date, db=None, cur=None): line = db.origin_visit_find_by_date(origin, visit_date, cur=cur) if line: return dict(zip(db.origin_visit_get_cols, line)) @timed @db_transaction(statement_timeout=500) def origin_visit_get_by(self, origin, visit, db=None, cur=None): ori_visit = db.origin_visit_get(origin, visit, cur) if not ori_visit: return None return dict(zip(db.origin_visit_get_cols, ori_visit)) @timed @db_transaction(statement_timeout=4000) def origin_visit_get_latest( self, origin, allowed_statuses=None, require_snapshot=False, db=None, cur=None): origin_visit = db.origin_visit_get_latest( origin, allowed_statuses=allowed_statuses, require_snapshot=require_snapshot, cur=cur) if origin_visit: return dict(zip(db.origin_visit_get_cols, origin_visit)) @timed @db_transaction() def origin_visit_get_random( self, type: str, db=None, cur=None) -> Optional[Dict[str, Any]]: result = db.origin_visit_get_random(type, cur) if result: return dict(zip(db.origin_visit_get_cols, result)) else: return None @timed @db_transaction(statement_timeout=2000) def object_find_by_sha1_git(self, ids, db=None, cur=None): ret = {id: [] for id in ids} for retval in db.object_find_by_sha1_git(ids, cur=cur): if retval[1]: ret[retval[0]].append(dict(zip(db.object_find_by_sha1_git_cols, retval))) return ret @timed @db_transaction(statement_timeout=500) def origin_get(self, origins, db=None, cur=None): if isinstance(origins, dict): # Old API return_single = True origins = [origins] elif len(origins) == 0: return [] else: return_single = False origin_urls = [origin['url'] for origin in origins] results = db.origin_get_by_url(origin_urls, cur) results = [dict(zip(db.origin_cols, result)) for result in results] if return_single: assert len(results) == 1 if results[0]['url'] is not None: return results[0] else: return None else: return [None if res['url'] is None else res for res in results] @timed @db_transaction_generator(statement_timeout=500) def origin_get_by_sha1(self, sha1s, db=None, cur=None): for line in db.origin_get_by_sha1(sha1s, cur): if line[0] is not None: yield dict(zip(db.origin_cols, line)) else: yield None @timed @db_transaction_generator() def origin_get_range(self, origin_from=1, origin_count=100, db=None, cur=None): for origin in db.origin_get_range(origin_from, origin_count, cur): yield dict(zip(db.origin_get_range_cols, origin)) @timed @db_transaction() def origin_list(self, page_token: Optional[str] = None, limit: int = 100, *, db=None, cur=None) -> dict: page_token = page_token or '0' if not isinstance(page_token, str): raise TypeError('page_token must be a string.') origin_from = int(page_token) result: Dict[str, Any] = { 'origins': [ dict(zip(db.origin_get_range_cols, origin)) for origin in db.origin_get_range(origin_from, limit, cur) ], } assert len(result['origins']) <= limit if len(result['origins']) == limit: result['next_page_token'] = str(result['origins'][limit-1]['id']+1) for origin in result['origins']: del origin['id'] return result @timed @db_transaction_generator() def origin_search(self, url_pattern, offset=0, limit=50, regexp=False, with_visit=False, db=None, cur=None): for origin in db.origin_search(url_pattern, offset, limit, regexp, with_visit, cur): yield dict(zip(db.origin_cols, origin)) @timed @db_transaction() def origin_count(self, url_pattern, regexp=False, with_visit=False, db=None, cur=None): return db.origin_count(url_pattern, regexp, with_visit, cur) @timed @db_transaction() def origin_add(self, origins, db=None, cur=None): origins = copy.deepcopy(list(origins)) for origin in origins: self.origin_add_one(origin, db=db, cur=cur) send_metric('origin:add', count=len(origins), method_name='origin_add') return origins @timed @db_transaction() def origin_add_one(self, origin, db=None, cur=None): origin_row = list(db.origin_get_by_url([origin['url']], cur))[0] origin_url = dict(zip(db.origin_cols, origin_row))['url'] if origin_url: return origin_url if self.journal_writer: self.journal_writer.write_addition('origin', origin) origins = db.origin_add(origin['url'], cur) send_metric('origin:add', count=len(origins), method_name='origin_add') return origins @db_transaction(statement_timeout=500) def stat_counters(self, db=None, cur=None): return {k: v for (k, v) in db.stat_counters()} @db_transaction() def refresh_stat_counters(self, db=None, cur=None): keys = [ 'content', 'directory', 'directory_entry_dir', 'directory_entry_file', 'directory_entry_rev', 'origin', 'origin_visit', 'person', 'release', 'revision', 'revision_history', 'skipped_content', 'snapshot'] for key in keys: cur.execute('select * from swh_update_counter(%s)', (key,)) @timed @db_transaction() def origin_metadata_add(self, origin_url, ts, provider, tool, metadata, db=None, cur=None): if isinstance(ts, str): ts = dateutil.parser.parse(ts) db.origin_metadata_add(origin_url, ts, provider, tool, metadata, cur) send_metric( 'origin_metadata:add', count=1, method_name='origin_metadata_add') @timed @db_transaction_generator(statement_timeout=500) def origin_metadata_get_by(self, origin_url, provider_type=None, db=None, cur=None): for line in db.origin_metadata_get_by(origin_url, provider_type, cur): yield dict(zip(db.origin_metadata_get_cols, line)) @timed @db_transaction() def tool_add(self, tools, db=None, cur=None): db.mktemp_tool(cur) db.copy_to(tools, 'tmp_tool', ['name', 'version', 'configuration'], cur) tools = db.tool_add_from_temp(cur) results = [dict(zip(db.tool_cols, line)) for line in tools] send_metric('tool:add', count=len(results), method_name='tool_add') return results @timed @db_transaction(statement_timeout=500) def tool_get(self, tool, db=None, cur=None): tool_conf = tool['configuration'] if isinstance(tool_conf, dict): tool_conf = json.dumps(tool_conf) idx = db.tool_get(tool['name'], tool['version'], tool_conf) if not idx: return None return dict(zip(db.tool_cols, idx)) @timed @db_transaction() def metadata_provider_add(self, provider_name, provider_type, provider_url, metadata, db=None, cur=None): result = db.metadata_provider_add(provider_name, provider_type, provider_url, metadata, cur) send_metric( 'metadata_provider:add', count=1, method_name='metadata_provider') return result @timed @db_transaction() def metadata_provider_get(self, provider_id, db=None, cur=None): result = db.metadata_provider_get(provider_id) if not result: return None return dict(zip(db.metadata_provider_cols, result)) @timed @db_transaction() def metadata_provider_get_by(self, provider, db=None, cur=None): result = db.metadata_provider_get_by(provider['provider_name'], provider['provider_url']) if not result: return None return dict(zip(db.metadata_provider_cols, result)) @timed def diff_directories(self, from_dir, to_dir, track_renaming=False): return diff.diff_directories(self, from_dir, to_dir, track_renaming) @timed def diff_revisions(self, from_rev, to_rev, track_renaming=False): return diff.diff_revisions(self, from_rev, to_rev, track_renaming) @timed def diff_revision(self, revision, track_renaming=False): return diff.diff_revision(self, revision, track_renaming) diff --git a/swh/storage/tests/conftest.py b/swh/storage/tests/conftest.py index a3641156..9d4d58ba 100644 --- a/swh/storage/tests/conftest.py +++ b/swh/storage/tests/conftest.py @@ -1,230 +1,234 @@ # Copyright (C) 2019 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import glob import pytest from typing import Union from pytest_postgresql import factories from pytest_postgresql.janitor import DatabaseJanitor, psycopg2, Version from os import path, environ from hypothesis import settings from typing import Dict import swh.storage from swh.core.utils import numfile_sortkey as sortkey from swh.model.tests.generate_testdata import gen_contents, gen_origins SQL_DIR = path.join(path.dirname(swh.storage.__file__), 'sql') environ['LC_ALL'] = 'C.UTF-8' DUMP_FILES = path.join(SQL_DIR, '*.sql') # define tests profile. Full documentation is at: # https://hypothesis.readthedocs.io/en/latest/settings.html#settings-profiles settings.register_profile("fast", max_examples=5, deadline=5000) settings.register_profile("slow", max_examples=20, deadline=5000) @pytest.fixture def swh_storage(postgresql_proc, swh_storage_postgresql): storage_config = { 'cls': 'local', 'db': 'postgresql://{user}@{host}:{port}/{dbname}'.format( host=postgresql_proc.host, port=postgresql_proc.port, user='postgres', dbname='tests'), 'objstorage': { 'cls': 'memory', 'args': {} }, 'journal_writer': { 'cls': 'memory', }, } storage = swh.storage.get_storage(**storage_config) return storage @pytest.fixture def swh_contents(swh_storage): contents = gen_contents(n=20) - swh_storage.content_add(contents) + swh_storage.content_add( + [c for c in contents if c['status'] != 'absent']) + swh_storage.skipped_content_add( + [c for c in contents if c['status'] == 'absent']) return contents @pytest.fixture def swh_origins(swh_storage): origins = gen_origins(n=100) swh_storage.origin_add(origins) return origins # the postgres_fact factory fixture below is mostly a copy of the code # from pytest-postgresql. We need a custom version here to be able to # specify our version of the DBJanitor we use. def postgresql_fact(process_fixture_name, db_name=None, dump_files=DUMP_FILES): @pytest.fixture def postgresql_factory(request): """ Fixture factory for PostgreSQL. :param FixtureRequest request: fixture request object :rtype: psycopg2.connection :returns: postgresql client """ config = factories.get_config(request) if not psycopg2: raise ImportError( 'No module named psycopg2. Please install it.' ) proc_fixture = request.getfixturevalue(process_fixture_name) # _, config = try_import('psycopg2', request) pg_host = proc_fixture.host pg_port = proc_fixture.port pg_user = proc_fixture.user pg_options = proc_fixture.options pg_db = db_name or config['dbname'] with SwhDatabaseJanitor( pg_user, pg_host, pg_port, pg_db, proc_fixture.version, dump_files=dump_files ): connection = psycopg2.connect( dbname=pg_db, user=pg_user, host=pg_host, port=pg_port, options=pg_options ) yield connection connection.close() return postgresql_factory swh_storage_postgresql = postgresql_fact('postgresql_proc') # This version of the DatabaseJanitor implement a different setup/teardown # behavior than than the stock one: instead of dropping, creating and # initializing the database for each test, it create and initialize the db only # once, then it truncate the tables. This is needed to have acceptable test # performances. class SwhDatabaseJanitor(DatabaseJanitor): def __init__( self, user: str, host: str, port: str, db_name: str, version: Union[str, float, Version], dump_files: str = DUMP_FILES ) -> None: super().__init__(user, host, port, db_name, version) self.dump_files = sorted( glob.glob(dump_files), key=sortkey) def db_setup(self): with psycopg2.connect( dbname=self.db_name, user=self.user, host=self.host, port=self.port, ) as cnx: with cnx.cursor() as cur: for fname in self.dump_files: with open(fname) as fobj: sql = fobj.read().replace('concurrently', '').strip() if sql: cur.execute(sql) cnx.commit() def db_reset(self): with psycopg2.connect( dbname=self.db_name, user=self.user, host=self.host, port=self.port, ) as cnx: with cnx.cursor() as cur: cur.execute( "SELECT table_name FROM information_schema.tables " "WHERE table_schema = %s", ('public',)) tables = set(table for (table,) in cur.fetchall()) for table in tables: cur.execute('truncate table %s cascade' % table) cur.execute( "SELECT sequence_name FROM information_schema.sequences " "WHERE sequence_schema = %s", ('public',)) seqs = set(seq for (seq,) in cur.fetchall()) for seq in seqs: cur.execute('ALTER SEQUENCE %s RESTART;' % seq) cnx.commit() def init(self): with self.cursor() as cur: cur.execute( "SELECT COUNT(1) FROM pg_database WHERE datname=%s;", (self.db_name,)) db_exists = cur.fetchone()[0] == 1 if db_exists: cur.execute( 'UPDATE pg_database SET datallowconn=true ' 'WHERE datname = %s;', (self.db_name,)) if db_exists: self.db_reset() else: with self.cursor() as cur: cur.execute('CREATE DATABASE "{}";'.format(self.db_name)) self.db_setup() def drop(self): pid_column = 'pid' with self.cursor() as cur: cur.execute( 'UPDATE pg_database SET datallowconn=false ' 'WHERE datname = %s;', (self.db_name,)) cur.execute( 'SELECT pg_terminate_backend(pg_stat_activity.{})' 'FROM pg_stat_activity ' 'WHERE pg_stat_activity.datname = %s;'.format(pid_column), (self.db_name,)) @pytest.fixture def sample_data() -> Dict: """Pre-defined sample storage object data to manipulate Returns: Dict of data (keys: content, directory, revision, release, person, origin) """ from .storage_data import data return { 'content': [data.cont, data.cont2], 'content_metadata': [data.cont3], + 'skipped_content': [data.skipped_cont, data.skipped_cont2], 'person': [data.person], 'directory': [data.dir2, data.dir], 'revision': [data.revision], 'release': [data.release, data.release2, data.release3], 'snapshot': [data.snapshot], 'origin': [data.origin, data.origin2], 'tool': [data.metadata_tool], 'provider': [data.provider], 'origin_metadata': [data.origin_metadata, data.origin_metadata2], } diff --git a/swh/storage/tests/test_buffer.py b/swh/storage/tests/test_buffer.py index 2bfcb24a..f58cc182 100644 --- a/swh/storage/tests/test_buffer.py +++ b/swh/storage/tests/test_buffer.py @@ -1,231 +1,277 @@ # Copyright (C) 2019 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information from swh.storage.buffer import BufferingProxyStorage def test_buffering_proxy_storage_content_threshold_not_hit(sample_data): contents = sample_data['content'] storage = BufferingProxyStorage( storage={'cls': 'memory'}, min_batch_size={ 'content': 10, } ) s = storage.content_add([contents[0], contents[1]]) assert s == {} # contents have not been written to storage missing_contents = storage.content_missing( [contents[0], contents[1]]) assert set(missing_contents) == set( [contents[0]['sha1'], contents[1]['sha1']]) s = storage.flush() assert s == { 'content:add': 1 + 1, 'content:add:bytes': contents[0]['length'] + contents[1]['length'], - 'skipped_content:add': 0 } missing_contents = storage.content_missing( [contents[0], contents[1]]) assert list(missing_contents) == [] def test_buffering_proxy_storage_content_threshold_nb_hit(sample_data): contents = sample_data['content'] storage = BufferingProxyStorage( storage={'cls': 'memory'}, min_batch_size={ 'content': 1, } ) s = storage.content_add([contents[0]]) assert s == { 'content:add': 1, 'content:add:bytes': contents[0]['length'], - 'skipped_content:add': 0 } missing_contents = storage.content_missing([contents[0]]) assert list(missing_contents) == [] s = storage.flush() assert s == {} def test_buffering_proxy_storage_content_threshold_bytes_hit(sample_data): contents = sample_data['content'] content_bytes_min_batch_size = 2 storage = BufferingProxyStorage( storage={'cls': 'memory'}, min_batch_size={ 'content': 10, 'content_bytes': content_bytes_min_batch_size, } ) assert contents[0]['length'] > content_bytes_min_batch_size s = storage.content_add([contents[0]]) assert s == { 'content:add': 1, 'content:add:bytes': contents[0]['length'], - 'skipped_content:add': 0 } missing_contents = storage.content_missing([contents[0]]) assert list(missing_contents) == [] s = storage.flush() assert s == {} +def test_buffering_proxy_storage_skipped_content_threshold_not_hit( + sample_data): + contents = sample_data['skipped_content'] + storage = BufferingProxyStorage( + storage={'cls': 'memory'}, + min_batch_size={ + 'skipped_content': 10, + } + ) + s = storage.skipped_content_add([contents[0], contents[1]]) + assert s == {} + + # contents have not been written to storage + missing_contents = storage.skipped_content_missing( + [contents[0], contents[1]]) + assert {c['sha1'] for c in missing_contents} \ + == {c['sha1'] for c in contents} + + s = storage.flush() + assert s == { + 'skipped_content:add': 1 + 1 + } + + missing_contents = storage.skipped_content_missing( + [contents[0], contents[1]]) + assert list(missing_contents) == [] + + +def test_buffering_proxy_storage_skipped_content_threshold_nb_hit(sample_data): + contents = sample_data['skipped_content'] + storage = BufferingProxyStorage( + storage={'cls': 'memory'}, + min_batch_size={ + 'skipped_content': 1, + } + ) + + s = storage.skipped_content_add([contents[0]]) + assert s == { + 'skipped_content:add': 1 + } + + missing_contents = storage.skipped_content_missing([contents[0]]) + assert list(missing_contents) == [] + + s = storage.flush() + assert s == {} + + def test_buffering_proxy_storage_directory_threshold_not_hit(sample_data): directories = sample_data['directory'] storage = BufferingProxyStorage( storage={'cls': 'memory'}, min_batch_size={ 'directory': 10, } ) s = storage.directory_add([directories[0]]) assert s == {} directory_id = directories[0]['id'] missing_directories = storage.directory_missing( [directory_id]) assert list(missing_directories) == [directory_id] s = storage.flush() assert s == { 'directory:add': 1, } missing_directories = storage.directory_missing( [directory_id]) assert list(missing_directories) == [] def test_buffering_proxy_storage_directory_threshold_hit(sample_data): directories = sample_data['directory'] storage = BufferingProxyStorage( storage={'cls': 'memory'}, min_batch_size={ 'directory': 1, } ) s = storage.directory_add([directories[0]]) assert s == { 'directory:add': 1, } missing_directories = storage.directory_missing( [directories[0]['id']]) assert list(missing_directories) == [] s = storage.flush() assert s == {} def test_buffering_proxy_storage_revision_threshold_not_hit(sample_data): revisions = sample_data['revision'] storage = BufferingProxyStorage( storage={'cls': 'memory'}, min_batch_size={ 'revision': 10, } ) s = storage.revision_add([revisions[0]]) assert s == {} revision_id = revisions[0]['id'] missing_revisions = storage.revision_missing( [revision_id]) assert list(missing_revisions) == [revision_id] s = storage.flush() assert s == { 'revision:add': 1, } missing_revisions = storage.revision_missing( [revision_id]) assert list(missing_revisions) == [] def test_buffering_proxy_storage_revision_threshold_hit(sample_data): revisions = sample_data['revision'] storage = BufferingProxyStorage( storage={'cls': 'memory'}, min_batch_size={ 'revision': 1, } ) s = storage.revision_add([revisions[0]]) assert s == { 'revision:add': 1, } missing_revisions = storage.revision_missing( [revisions[0]['id']]) assert list(missing_revisions) == [] s = storage.flush() assert s == {} def test_buffering_proxy_storage_release_threshold_not_hit(sample_data): releases = sample_data['release'] threshold = 10 assert len(releases) < threshold storage = BufferingProxyStorage( storage={'cls': 'memory'}, min_batch_size={ 'release': threshold, # configuration set } ) s = storage.release_add(releases) assert s == {} release_ids = [r['id'] for r in releases] missing_releases = storage.release_missing(release_ids) assert list(missing_releases) == release_ids s = storage.flush() assert s == { 'release:add': len(releases), } missing_releases = storage.release_missing(release_ids) assert list(missing_releases) == [] def test_buffering_proxy_storage_release_threshold_hit(sample_data): releases = sample_data['release'] threshold = 2 assert len(releases) > threshold storage = BufferingProxyStorage( storage={'cls': 'memory'}, min_batch_size={ 'release': threshold, # configuration set } ) s = storage.release_add(releases) assert s == { 'release:add': len(releases), } release_ids = [r['id'] for r in releases] missing_releases = storage.release_missing(release_ids) assert list(missing_releases) == [] s = storage.flush() assert s == {} diff --git a/swh/storage/tests/test_cassandra.py b/swh/storage/tests/test_cassandra.py index ad0ba7d8..7ca4be11 100644 --- a/swh/storage/tests/test_cassandra.py +++ b/swh/storage/tests/test_cassandra.py @@ -1,274 +1,269 @@ # Copyright (C) 2018-2019 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import os import signal import socket import subprocess import time import pytest from swh.storage import get_storage from swh.storage.cassandra import create_keyspace from swh.storage.cassandra.schema import TABLES from swh.storage.tests.test_storage import TestStorage as _TestStorage from swh.storage.tests.test_storage import TestStorageGeneratedData \ as _TestStorageGeneratedData CONFIG_TEMPLATE = ''' data_file_directories: - {data_dir}/data commitlog_directory: {data_dir}/commitlog hints_directory: {data_dir}/hints saved_caches_directory: {data_dir}/saved_caches commitlog_sync: periodic commitlog_sync_period_in_ms: 1000000 partitioner: org.apache.cassandra.dht.Murmur3Partitioner endpoint_snitch: SimpleSnitch seed_provider: - class_name: org.apache.cassandra.locator.SimpleSeedProvider parameters: - seeds: "127.0.0.1" storage_port: {storage_port} native_transport_port: {native_transport_port} start_native_transport: true listen_address: 127.0.0.1 enable_user_defined_functions: true # speed-up by disabling period saving to disk key_cache_save_period: 0 row_cache_save_period: 0 trickle_fsync: false commitlog_sync_period_in_ms: 100000 ''' def free_port(): sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM) sock.bind(('127.0.0.1', 0)) port = sock.getsockname()[1] sock.close() return port def wait_for_peer(addr, port): wait_until = time.time() + 20 while time.time() < wait_until: try: sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM) sock.connect((addr, port)) except ConnectionRefusedError: time.sleep(0.1) else: sock.close() return True return False @pytest.fixture(scope='session') def cassandra_cluster(tmpdir_factory): cassandra_conf = tmpdir_factory.mktemp('cassandra_conf') cassandra_data = tmpdir_factory.mktemp('cassandra_data') cassandra_log = tmpdir_factory.mktemp('cassandra_log') native_transport_port = free_port() storage_port = free_port() jmx_port = free_port() with open(str(cassandra_conf.join('cassandra.yaml')), 'w') as fd: fd.write(CONFIG_TEMPLATE.format( data_dir=str(cassandra_data), storage_port=storage_port, native_transport_port=native_transport_port, )) if os.environ.get('LOG_CASSANDRA'): stdout = stderr = None else: stdout = stderr = subprocess.DEVNULL proc = subprocess.Popen( [ '/usr/sbin/cassandra', '-Dcassandra.config=file://%s/cassandra.yaml' % cassandra_conf, '-Dcassandra.logdir=%s' % cassandra_log, '-Dcassandra.jmx.local.port=%d' % jmx_port, '-Dcassandra-foreground=yes', ], start_new_session=True, env={ 'MAX_HEAP_SIZE': '300M', 'HEAP_NEWSIZE': '50M', 'JVM_OPTS': '-Xlog:gc=error:file=%s/gc.log' % cassandra_log, }, stdout=stdout, stderr=stderr, ) running = wait_for_peer('127.0.0.1', native_transport_port) if running: yield (['127.0.0.1'], native_transport_port) if not running or os.environ.get('LOG_CASSANDRA'): with open(str(cassandra_log.join('debug.log'))) as fd: print(fd.read()) if not running: raise Exception('cassandra process stopped unexpectedly.') pgrp = os.getpgid(proc.pid) os.killpg(pgrp, signal.SIGKILL) class RequestHandler: def on_request(self, rf): if hasattr(rf.message, 'query'): print() print(rf.message.query) @pytest.fixture(scope='session') def keyspace(cassandra_cluster): (hosts, port) = cassandra_cluster keyspace = os.urandom(10).hex() create_keyspace(hosts, keyspace, port) return keyspace # tests are executed using imported classes (TestStorage and # TestStorageGeneratedData) using overloaded swh_storage fixture # below @pytest.fixture def swh_storage(cassandra_cluster, keyspace): (hosts, port) = cassandra_cluster storage = get_storage( 'cassandra', hosts=hosts, port=port, keyspace=keyspace, journal_writer={ 'cls': 'memory', }, objstorage={ 'cls': 'memory', 'args': {}, }, ) yield storage for table in TABLES: storage._cql_runner._session.execute('TRUNCATE TABLE "%s"' % table) @pytest.mark.cassandra class TestCassandraStorage(_TestStorage): @pytest.mark.skip('content_update is not yet implemented for Cassandra') def test_content_update(self): pass - @pytest.mark.skip( - 'not implemented, see https://forge.softwareheritage.org/T1633') - def test_skipped_content_add(self): - pass - @pytest.mark.skip( 'The "person" table of the pgsql is a legacy thing, and not ' 'supported by the cassandra backend.') def test_person_fullname_unicity(self): pass @pytest.mark.skip( 'The "person" table of the pgsql is a legacy thing, and not ' 'supported by the cassandra backend.') def test_person_get(self): pass @pytest.mark.skip('Not yet implemented') def test_metadata_provider_add(self): pass @pytest.mark.skip('Not yet implemented') def test_metadata_provider_get(self): pass @pytest.mark.skip('Not yet implemented') def test_metadata_provider_get_by(self): pass @pytest.mark.skip('Not yet implemented') def test_origin_metadata_add(self): pass @pytest.mark.skip('Not yet implemented') def test_origin_metadata_get(self): pass @pytest.mark.skip('Not yet implemented') def test_origin_metadata_get_by_provider_type(self): pass @pytest.mark.skip('Not supported by Cassandra') def test_origin_count(self): pass @pytest.mark.cassandra class TestCassandraStorageGeneratedData(_TestStorageGeneratedData): @pytest.mark.skip('Not supported by Cassandra') def test_origin_count(self): pass @pytest.mark.skip('Not supported by Cassandra') def test_origin_get_range(self): pass @pytest.mark.skip('Not supported by Cassandra') def test_origin_get_range_from_zero(self): pass @pytest.mark.skip('Not supported by Cassandra') def test_generate_content_get_range_limit(self): pass @pytest.mark.skip('Not supported by Cassandra') def test_generate_content_get_range_no_limit(self): pass @pytest.mark.skip('Not supported by Cassandra') def test_generate_content_get_range(self): pass @pytest.mark.skip('Not supported by Cassandra') def test_generate_content_get_range_empty(self): pass @pytest.mark.skip('Not supported by Cassandra') def test_generate_content_get_range_limit_none(self): pass @pytest.mark.skip('Not supported by Cassandra') def test_generate_content_get_range_full(self): pass @pytest.mark.skip('Not supported by Cassandra') def test_origin_count_with_visit_no_visits(self): pass @pytest.mark.skip('Not supported by Cassandra') def test_origin_count_with_visit_with_visits_and_snapshot(self): pass @pytest.mark.skip('Not supported by Cassandra') def test_origin_count_with_visit_with_visits_no_snapshot(self): pass diff --git a/swh/storage/tests/test_filter.py b/swh/storage/tests/test_filter.py index 8ec68e71..951d5535 100644 --- a/swh/storage/tests/test_filter.py +++ b/swh/storage/tests/test_filter.py @@ -1,74 +1,93 @@ # Copyright (C) 2019 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information from swh.storage.filter import FilteringProxyStorage def test_filtering_proxy_storage_content(sample_data): sample_content = sample_data['content'][0] storage = FilteringProxyStorage(storage={'cls': 'memory'}) content = next(storage.content_get([sample_content['sha1']])) assert not content s = storage.content_add([sample_content]) assert s == { 'content:add': 1, 'content:add:bytes': sample_content['length'], - 'skipped_content:add': 0 } content = next(storage.content_get([sample_content['sha1']])) assert content is not None s = storage.content_add([sample_content]) assert s == { 'content:add': 0, 'content:add:bytes': 0, - 'skipped_content:add': 0 + } + + +def test_filtering_proxy_storage_skipped_content(sample_data): + sample_content = sample_data['skipped_content'][0] + storage = FilteringProxyStorage(storage={'cls': 'memory'}) + + content = next(storage.skipped_content_missing([sample_content])) + assert content['sha1'] == sample_content['sha1'] + + s = storage.skipped_content_add([sample_content]) + assert s == { + 'skipped_content:add': 1, + } + + content = list(storage.skipped_content_missing([sample_content])) + assert content == [] + + s = storage.skipped_content_add([sample_content]) + assert s == { + 'skipped_content:add': 0, } def test_filtering_proxy_storage_revision(sample_data): sample_revision = sample_data['revision'][0] storage = FilteringProxyStorage(storage={'cls': 'memory'}) revision = next(storage.revision_get([sample_revision['id']])) assert not revision s = storage.revision_add([sample_revision]) assert s == { 'revision:add': 1, } revision = next(storage.revision_get([sample_revision['id']])) assert revision is not None s = storage.revision_add([sample_revision]) assert s == { 'revision:add': 0, } def test_filtering_proxy_storage_directory(sample_data): sample_directory = sample_data['directory'][0] storage = FilteringProxyStorage(storage={'cls': 'memory'}) directory = next(storage.directory_missing([sample_directory['id']])) assert directory s = storage.directory_add([sample_directory]) assert s == { 'directory:add': 1, } directory = list(storage.directory_missing([sample_directory['id']])) assert not directory s = storage.directory_add([sample_directory]) assert s == { 'directory:add': 0, } diff --git a/swh/storage/tests/test_retry.py b/swh/storage/tests/test_retry.py index a2d7684a..06100c31 100644 --- a/swh/storage/tests/test_retry.py +++ b/swh/storage/tests/test_retry.py @@ -1,899 +1,897 @@ # Copyright (C) 2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import psycopg2 import pytest from typing import Dict from unittest.mock import call from swh.storage import HashCollision from swh.storage.retry import RetryingProxyStorage @pytest.fixture def swh_storage(): return RetryingProxyStorage(storage={'cls': 'memory'}) def test_retrying_proxy_storage_content_add(swh_storage, sample_data): """Standard content_add works as before """ sample_content = sample_data['content'][0] content = next(swh_storage.content_get([sample_content['sha1']])) assert not content s = swh_storage.content_add([sample_content]) assert s == { 'content:add': 1, 'content:add:bytes': sample_content['length'], - 'skipped_content:add': 0 } content = next(swh_storage.content_get([sample_content['sha1']])) assert content['sha1'] == sample_content['sha1'] def test_retrying_proxy_storage_content_add_with_retry( swh_storage, sample_data, mocker): """Multiple retries for hash collision and psycopg2 error but finally ok """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.content_add') mock_memory.side_effect = [ # first try goes ko HashCollision('content hash collision'), # second try goes ko psycopg2.IntegrityError('content already inserted'), # ok then! {'content:add': 1} ] sample_content = sample_data['content'][0] content = next(swh_storage.content_get([sample_content['sha1']])) assert not content s = swh_storage.content_add([sample_content]) assert s == {'content:add': 1} assert mock_memory.has_calls([ call([sample_content]), call([sample_content]), call([sample_content]), ]) def test_retrying_proxy_swh_storage_content_add_failure( swh_storage, sample_data, mocker): """Unfiltered errors are raising without retry """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.content_add') mock_memory.side_effect = ValueError('Refuse to add content always!') sample_content = sample_data['content'][0] content = next(swh_storage.content_get([sample_content['sha1']])) assert not content with pytest.raises(ValueError, match='Refuse to add'): swh_storage.content_add([sample_content]) assert mock_memory.call_count == 1 def test_retrying_proxy_storage_content_add_metadata(swh_storage, sample_data): """Standard content_add_metadata works as before """ sample_content = sample_data['content_metadata'][0] pk = sample_content['sha1'] content_metadata = swh_storage.content_get_metadata([pk]) assert not content_metadata[pk] s = swh_storage.content_add_metadata([sample_content]) assert s == { 'content:add': 1, - 'skipped_content:add': 0 } content_metadata = swh_storage.content_get_metadata([pk]) assert len(content_metadata[pk]) == 1 assert content_metadata[pk][0]['sha1'] == pk def test_retrying_proxy_storage_content_add_metadata_with_retry( swh_storage, sample_data, mocker): """Multiple retries for hash collision and psycopg2 error but finally ok """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.content_add_metadata') mock_memory.side_effect = [ # first try goes ko HashCollision('content_metadata hash collision'), # second try goes ko psycopg2.IntegrityError('content_metadata already inserted'), # ok then! {'content:add': 1} ] sample_content = sample_data['content_metadata'][0] s = swh_storage.content_add_metadata([sample_content]) assert s == {'content:add': 1} assert mock_memory.has_calls([ call([sample_content]), call([sample_content]), call([sample_content]), ]) def test_retrying_proxy_swh_storage_content_add_metadata_failure( swh_storage, sample_data, mocker): """Unfiltered errors are raising without retry """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.content_add_metadata') mock_memory.side_effect = ValueError('Refuse to add content_metadata!') sample_content = sample_data['content_metadata'][0] pk = sample_content['sha1'] content_metadata = swh_storage.content_get_metadata([pk]) assert not content_metadata[pk] with pytest.raises(ValueError, match='Refuse to add'): swh_storage.content_add_metadata([sample_content]) assert mock_memory.call_count == 1 def test_retrying_proxy_swh_storage_origin_add_one(swh_storage, sample_data): """Standard origin_add_one works as before """ sample_origin = sample_data['origin'][0] origin = swh_storage.origin_get(sample_origin) assert not origin r = swh_storage.origin_add_one(sample_origin) assert r == sample_origin['url'] origin = swh_storage.origin_get(sample_origin) assert origin['url'] == sample_origin['url'] def test_retrying_proxy_swh_storage_origin_add_one_retry( swh_storage, sample_data, mocker): """Multiple retries for hash collision and psycopg2 error but finally ok """ sample_origin = sample_data['origin'][1] mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.origin_add_one') mock_memory.side_effect = [ # first try goes ko HashCollision('origin hash collision'), # second try goes ko psycopg2.IntegrityError('origin already inserted'), # ok then! sample_origin['url'] ] origin = swh_storage.origin_get(sample_origin) assert not origin r = swh_storage.origin_add_one(sample_origin) assert r == sample_origin['url'] assert mock_memory.has_calls([ call([sample_origin]), call([sample_origin]), call([sample_origin]), ]) def test_retrying_proxy_swh_storage_origin_add_one_failure( swh_storage, sample_data, mocker): """Unfiltered errors are raising without retry """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.origin_add_one') mock_memory.side_effect = ValueError('Refuse to add origin always!') sample_origin = sample_data['origin'][0] origin = swh_storage.origin_get(sample_origin) assert not origin with pytest.raises(ValueError, match='Refuse to add'): swh_storage.origin_add_one([sample_origin]) assert mock_memory.call_count == 1 def test_retrying_proxy_swh_storage_origin_visit_add(swh_storage, sample_data): """Standard origin_visit_add works as before """ sample_origin = sample_data['origin'][0] swh_storage.origin_add_one(sample_origin) origin_url = sample_origin['url'] origin = list(swh_storage.origin_visit_get(origin_url)) assert not origin origin_visit = swh_storage.origin_visit_add( origin_url, '2020-01-01', 'hg') assert origin_visit['origin'] == origin_url assert isinstance(origin_visit['visit'], int) origin_visit = next(swh_storage.origin_visit_get(origin_url)) assert origin_visit['origin'] == origin_url assert isinstance(origin_visit['visit'], int) def test_retrying_proxy_swh_storage_origin_visit_add_retry( swh_storage, sample_data, mocker): """Multiple retries for hash collision and psycopg2 error but finally ok """ sample_origin = sample_data['origin'][1] swh_storage.origin_add_one(sample_origin) origin_url = sample_origin['url'] mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.origin_visit_add') mock_memory.side_effect = [ # first try goes ko HashCollision('origin hash collision'), # second try goes ko psycopg2.IntegrityError('origin already inserted'), # ok then! {'origin': origin_url, 'visit': 1} ] origin = list(swh_storage.origin_visit_get(origin_url)) assert not origin r = swh_storage.origin_visit_add(sample_origin, '2020-01-01', 'git') assert r == {'origin': origin_url, 'visit': 1} assert mock_memory.has_calls([ call(sample_origin, '2020-01-01', 'git'), call(sample_origin, '2020-01-01', 'git'), call(sample_origin, '2020-01-01', 'git') ]) def test_retrying_proxy_swh_storage_origin_visit_add_failure( swh_storage, sample_data, mocker): """Unfiltered errors are raising without retry """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.origin_visit_add') mock_memory.side_effect = ValueError('Refuse to add origin always!') origin_url = sample_data['origin'][0]['url'] origin = list(swh_storage.origin_visit_get(origin_url)) assert not origin with pytest.raises(ValueError, match='Refuse to add'): swh_storage.origin_visit_add(origin_url, '2020-01-31', 'svn') assert mock_memory.has_calls([ call(origin_url, '2020-01-31', 'svn'), ]) def test_retrying_proxy_storage_tool_add(swh_storage, sample_data): """Standard tool_add works as before """ sample_tool = sample_data['tool'][0] tool = swh_storage.tool_get(sample_tool) assert not tool tools = swh_storage.tool_add([sample_tool]) assert tools tool = tools[0] tool.pop('id') assert tool == sample_tool tool = swh_storage.tool_get(sample_tool) tool.pop('id') assert tool == sample_tool def test_retrying_proxy_storage_tool_add_with_retry( swh_storage, sample_data, mocker): """Multiple retries for hash collision and psycopg2 error but finally ok """ sample_tool = sample_data['tool'][0] mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.tool_add') mock_memory.side_effect = [ # first try goes ko HashCollision('tool hash collision'), # second try goes ko psycopg2.IntegrityError('tool already inserted'), # ok then! [sample_tool] ] tool = swh_storage.tool_get(sample_tool) assert not tool tools = swh_storage.tool_add([sample_tool]) assert tools == [sample_tool] assert mock_memory.has_calls([ call([sample_tool]), call([sample_tool]), call([sample_tool]), ]) def test_retrying_proxy_swh_storage_tool_add_failure( swh_storage, sample_data, mocker): """Unfiltered errors are raising without retry """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.tool_add') mock_memory.side_effect = ValueError('Refuse to add tool always!') sample_tool = sample_data['tool'][0] tool = swh_storage.tool_get(sample_tool) assert not tool with pytest.raises(ValueError, match='Refuse to add'): swh_storage.tool_add([sample_tool]) assert mock_memory.call_count == 1 def to_provider(provider: Dict) -> Dict: return { 'provider_name': provider['name'], 'provider_url': provider['url'], 'provider_type': provider['type'], 'metadata': provider['metadata'], } def test_retrying_proxy_storage_metadata_provider_add( swh_storage, sample_data): """Standard metadata_provider_add works as before """ provider = sample_data['provider'][0] provider_get = to_provider(provider) provider = swh_storage.metadata_provider_get_by(provider_get) assert not provider provider_id = swh_storage.metadata_provider_add(**provider_get) assert provider_id actual_provider = swh_storage.metadata_provider_get(provider_id) assert actual_provider actual_provider_id = actual_provider.pop('id') assert actual_provider_id == provider_id assert actual_provider == provider_get def test_retrying_proxy_storage_metadata_provider_add_with_retry( swh_storage, sample_data, mocker): """Multiple retries for hash collision and psycopg2 error but finally ok """ provider = sample_data['provider'][0] provider_get = to_provider(provider) mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.metadata_provider_add') mock_memory.side_effect = [ # first try goes ko HashCollision('provider_id hash collision'), # second try goes ko psycopg2.IntegrityError('provider_id already inserted'), # ok then! 'provider_id', ] provider = swh_storage.metadata_provider_get_by(provider_get) assert not provider provider_id = swh_storage.metadata_provider_add(**provider_get) assert provider_id == 'provider_id' assert mock_memory.has_calls([ call(**provider_get), call(**provider_get), call(**provider_get), ]) def test_retrying_proxy_swh_storage_metadata_provider_add_failure( swh_storage, sample_data, mocker): """Unfiltered errors are raising without retry """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.metadata_provider_add') mock_memory.side_effect = ValueError('Refuse to add provider_id always!') provider = sample_data['provider'][0] provider_get = to_provider(provider) provider_id = swh_storage.metadata_provider_get_by(provider_get) assert not provider_id with pytest.raises(ValueError, match='Refuse to add'): swh_storage.metadata_provider_add(**provider_get) assert mock_memory.call_count == 1 def test_retrying_proxy_storage_origin_metadata_add( swh_storage, sample_data): """Standard origin_metadata_add works as before """ ori_meta = sample_data['origin_metadata'][0] origin = ori_meta['origin'] swh_storage.origin_add_one(origin) provider_get = to_provider(ori_meta['provider']) provider_id = swh_storage.metadata_provider_add(**provider_get) origin_metadata = swh_storage.origin_metadata_get_by(origin['url']) assert not origin_metadata swh_storage.origin_metadata_add( origin['url'], ori_meta['discovery_date'], provider_id, ori_meta['tool'], ori_meta['metadata']) origin_metadata = swh_storage.origin_metadata_get_by( origin['url']) assert origin_metadata def test_retrying_proxy_storage_origin_metadata_add_with_retry( swh_storage, sample_data, mocker): """Multiple retries for hash collision and psycopg2 error but finally ok """ ori_meta = sample_data['origin_metadata'][0] origin = ori_meta['origin'] swh_storage.origin_add_one(origin) provider_get = to_provider(ori_meta['provider']) provider_id = swh_storage.metadata_provider_add(**provider_get) mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.origin_metadata_add') mock_memory.side_effect = [ # first try goes ko HashCollision('provider_id hash collision'), # second try goes ko psycopg2.IntegrityError('provider_id already inserted'), # ok then! None ] url = origin['url'] ts = ori_meta['discovery_date'] tool_id = ori_meta['tool'] metadata = ori_meta['metadata'] # No exception raised as insertion finally came through swh_storage.origin_metadata_add(url, ts, provider_id, tool_id, metadata) mock_memory.assert_has_calls([ # 3 calls, as long as error raised call(url, ts, provider_id, tool_id, metadata), call(url, ts, provider_id, tool_id, metadata), call(url, ts, provider_id, tool_id, metadata) ]) def test_retrying_proxy_swh_storage_origin_metadata_add_failure( swh_storage, sample_data, mocker): """Unfiltered errors are raising without retry """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.origin_metadata_add') mock_memory.side_effect = ValueError('Refuse to add always!') ori_meta = sample_data['origin_metadata'][0] origin = ori_meta['origin'] swh_storage.origin_add_one(origin) url = origin['url'] ts = ori_meta['discovery_date'] provider_id = 'provider_id' tool_id = ori_meta['tool'] metadata = ori_meta['metadata'] with pytest.raises(ValueError, match='Refuse to add'): swh_storage.origin_metadata_add(url, ts, provider_id, tool_id, metadata) assert mock_memory.call_count == 1 def test_retrying_proxy_swh_storage_origin_visit_update( swh_storage, sample_data): """Standard origin_visit_update works as before """ sample_origin = sample_data['origin'][0] swh_storage.origin_add_one(sample_origin) origin_url = sample_origin['url'] origin_visit = swh_storage.origin_visit_add( origin_url, '2020-01-01', 'hg') ov = next(swh_storage.origin_visit_get(origin_url)) assert ov['origin'] == origin_url assert ov['visit'] == origin_visit['visit'] assert ov['status'] == 'ongoing' assert ov['snapshot'] is None assert ov['metadata'] is None swh_storage.origin_visit_update(origin_url, ov['visit'], status='full') ov = next(swh_storage.origin_visit_get(origin_url)) assert ov['origin'] == origin_url assert ov['visit'] == origin_visit['visit'] assert ov['status'] == 'full' assert ov['snapshot'] is None assert ov['metadata'] is None def test_retrying_proxy_swh_storage_origin_visit_update_retry( swh_storage, sample_data, mocker): """Multiple retries for hash collision and psycopg2 error but finally ok """ sample_origin = sample_data['origin'][1] origin_url = sample_origin['url'] mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.origin_visit_update') mock_memory.side_effect = [ # first try goes ko HashCollision('origin hash collision'), # second try goes ko psycopg2.IntegrityError('origin already inserted'), # ok then! {'origin': origin_url, 'visit': 1} ] visit_id = 1 swh_storage.origin_visit_update(origin_url, visit_id, status='full') assert mock_memory.has_calls([ call(origin_url, visit_id, status='full'), call(origin_url, visit_id, status='full'), call(origin_url, visit_id, status='full'), ]) def test_retrying_proxy_swh_storage_origin_visit_update_failure( swh_storage, sample_data, mocker): """Unfiltered errors are raising without retry """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.origin_visit_update') mock_memory.side_effect = ValueError('Refuse to add origin always!') origin_url = sample_data['origin'][0]['url'] visit_id = 9 with pytest.raises(ValueError, match='Refuse to add'): swh_storage.origin_visit_update(origin_url, visit_id, 'partial') assert mock_memory.call_count == 1 def test_retrying_proxy_storage_directory_add(swh_storage, sample_data): """Standard directory_add works as before """ sample_dir = sample_data['directory'][0] directory = swh_storage.directory_get_random() # no directory assert not directory s = swh_storage.directory_add([sample_dir]) assert s == { 'directory:add': 1, } directory_id = swh_storage.directory_get_random() # only 1 assert directory_id == sample_dir['id'] def test_retrying_proxy_storage_directory_add_with_retry( swh_storage, sample_data, mocker): """Multiple retries for hash collision and psycopg2 error but finally ok """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.directory_add') mock_memory.side_effect = [ # first try goes ko HashCollision('directory hash collision'), # second try goes ko psycopg2.IntegrityError('directory already inserted'), # ok then! {'directory:add': 1} ] sample_dir = sample_data['directory'][1] directory_id = swh_storage.directory_get_random() # no directory assert not directory_id s = swh_storage.directory_add([sample_dir]) assert s == { 'directory:add': 1, } assert mock_memory.has_calls([ call([sample_dir]), call([sample_dir]), call([sample_dir]), ]) def test_retrying_proxy_swh_storage_directory_add_failure( swh_storage, sample_data, mocker): """Unfiltered errors are raising without retry """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.directory_add') mock_memory.side_effect = ValueError('Refuse to add directory always!') sample_dir = sample_data['directory'][0] directory_id = swh_storage.directory_get_random() # no directory assert not directory_id with pytest.raises(ValueError, match='Refuse to add'): swh_storage.directory_add([sample_dir]) assert mock_memory.call_count == 1 def test_retrying_proxy_storage_revision_add(swh_storage, sample_data): """Standard revision_add works as before """ sample_rev = sample_data['revision'][0] revision = next(swh_storage.revision_get([sample_rev['id']])) assert not revision s = swh_storage.revision_add([sample_rev]) assert s == { 'revision:add': 1, } revision = next(swh_storage.revision_get([sample_rev['id']])) assert revision['id'] == sample_rev['id'] def test_retrying_proxy_storage_revision_add_with_retry( swh_storage, sample_data, mocker): """Multiple retries for hash collision and psycopg2 error but finally ok """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.revision_add') mock_memory.side_effect = [ # first try goes ko HashCollision('revision hash collision'), # second try goes ko psycopg2.IntegrityError('revision already inserted'), # ok then! {'revision:add': 1} ] sample_rev = sample_data['revision'][0] revision = next(swh_storage.revision_get([sample_rev['id']])) assert not revision s = swh_storage.revision_add([sample_rev]) assert s == { 'revision:add': 1, } assert mock_memory.has_calls([ call([sample_rev]), call([sample_rev]), call([sample_rev]), ]) def test_retrying_proxy_swh_storage_revision_add_failure( swh_storage, sample_data, mocker): """Unfiltered errors are raising without retry """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.revision_add') mock_memory.side_effect = ValueError('Refuse to add revision always!') sample_rev = sample_data['revision'][0] revision = next(swh_storage.revision_get([sample_rev['id']])) assert not revision with pytest.raises(ValueError, match='Refuse to add'): swh_storage.revision_add([sample_rev]) assert mock_memory.call_count == 1 def test_retrying_proxy_storage_release_add(swh_storage, sample_data): """Standard release_add works as before """ sample_rel = sample_data['release'][0] release = next(swh_storage.release_get([sample_rel['id']])) assert not release s = swh_storage.release_add([sample_rel]) assert s == { 'release:add': 1, } release = next(swh_storage.release_get([sample_rel['id']])) assert release['id'] == sample_rel['id'] def test_retrying_proxy_storage_release_add_with_retry( swh_storage, sample_data, mocker): """Multiple retries for hash collision and psycopg2 error but finally ok """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.release_add') mock_memory.side_effect = [ # first try goes ko HashCollision('release hash collision'), # second try goes ko psycopg2.IntegrityError('release already inserted'), # ok then! {'release:add': 1} ] sample_rel = sample_data['release'][0] release = next(swh_storage.release_get([sample_rel['id']])) assert not release s = swh_storage.release_add([sample_rel]) assert s == { 'release:add': 1, } assert mock_memory.has_calls([ call([sample_rel]), call([sample_rel]), call([sample_rel]), ]) def test_retrying_proxy_swh_storage_release_add_failure( swh_storage, sample_data, mocker): """Unfiltered errors are raising without retry """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.release_add') mock_memory.side_effect = ValueError('Refuse to add release always!') sample_rel = sample_data['release'][0] release = next(swh_storage.release_get([sample_rel['id']])) assert not release with pytest.raises(ValueError, match='Refuse to add'): swh_storage.release_add([sample_rel]) assert mock_memory.call_count == 1 def test_retrying_proxy_storage_snapshot_add(swh_storage, sample_data): """Standard snapshot_add works as before """ sample_snap = sample_data['snapshot'][0] snapshot = swh_storage.snapshot_get(sample_snap['id']) assert not snapshot s = swh_storage.snapshot_add([sample_snap]) assert s == { 'snapshot:add': 1, } snapshot = swh_storage.snapshot_get(sample_snap['id']) assert snapshot['id'] == sample_snap['id'] def test_retrying_proxy_storage_snapshot_add_with_retry( swh_storage, sample_data, mocker): """Multiple retries for hash collision and psycopg2 error but finally ok """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.snapshot_add') mock_memory.side_effect = [ # first try goes ko HashCollision('snapshot hash collision'), # second try goes ko psycopg2.IntegrityError('snapshot already inserted'), # ok then! {'snapshot:add': 1} ] sample_snap = sample_data['snapshot'][0] snapshot = swh_storage.snapshot_get(sample_snap['id']) assert not snapshot s = swh_storage.snapshot_add([sample_snap]) assert s == { 'snapshot:add': 1, } assert mock_memory.has_calls([ call([sample_snap]), call([sample_snap]), call([sample_snap]), ]) def test_retrying_proxy_swh_storage_snapshot_add_failure( swh_storage, sample_data, mocker): """Unfiltered errors are raising without retry """ mock_memory = mocker.patch( 'swh.storage.in_memory.InMemoryStorage.snapshot_add') mock_memory.side_effect = ValueError('Refuse to add snapshot always!') sample_snap = sample_data['snapshot'][0] snapshot = swh_storage.snapshot_get(sample_snap['id']) assert not snapshot with pytest.raises(ValueError, match='Refuse to add'): swh_storage.snapshot_add([sample_snap]) assert mock_memory.call_count == 1 diff --git a/swh/storage/tests/test_storage.py b/swh/storage/tests/test_storage.py index 94e70961..ab68ec65 100644 --- a/swh/storage/tests/test_storage.py +++ b/swh/storage/tests/test_storage.py @@ -1,3803 +1,3800 @@ # Copyright (C) 2015-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import copy from contextlib import contextmanager import datetime import inspect import itertools import math import queue import random import threading from collections import defaultdict from datetime import timedelta from unittest.mock import Mock import psycopg2 import pytest from hypothesis import given, strategies, settings, HealthCheck from typing import ClassVar, Optional from swh.model import from_disk, identifiers from swh.model.hashutil import hash_to_bytes from swh.model.hypothesis_strategies import objects from swh.storage import HashCollision from swh.storage.converters import origin_url_to_sha1 as sha1 from swh.storage.interface import StorageInterface from .storage_data import data @contextmanager def db_transaction(storage): with storage.db() as db: with db.transaction() as cur: yield db, cur def normalize_entity(entity): entity = copy.deepcopy(entity) for key in ('date', 'committer_date'): if key in entity: entity[key] = identifiers.normalize_timestamp(entity[key]) return entity def transform_entries(dir_, *, prefix=b''): for ent in dir_['entries']: yield { 'dir_id': dir_['id'], 'type': ent['type'], 'target': ent['target'], 'name': prefix + ent['name'], 'perms': ent['perms'], 'status': None, 'sha1': None, 'sha1_git': None, 'sha256': None, 'length': None, } def cmpdir(directory): return (directory['type'], directory['dir_id']) def short_revision(revision): return [revision['id'], revision['parents']] def assert_contents_ok(expected_contents, actual_contents, keys_to_check={'sha1', 'data'}): """Assert that a given list of contents matches on a given set of keys. """ for k in keys_to_check: expected_list = set([c.get(k) for c in expected_contents]) actual_list = set([c.get(k) for c in actual_contents]) assert actual_list == expected_list, k class TestStorage: """Main class for Storage testing. This class is used as-is to test local storage (see TestLocalStorage below) and remote storage (see TestRemoteStorage in test_remote_storage.py. We need to have the two classes inherit from this base class separately to avoid nosetests running the tests from the base class twice. """ maxDiff = None # type: ClassVar[Optional[int]] def test_types(self, swh_storage): """Checks all methods of StorageInterface are implemented by this backend, and that they have the same signature.""" # Create an instance of the protocol (which cannot be instantiated # directly, so this creates a subclass, then instantiates it) interface = type('_', (StorageInterface,), {})() assert 'content_add' in dir(interface) missing_methods = [] for meth_name in dir(interface): if meth_name.startswith('_'): continue interface_meth = getattr(interface, meth_name) try: concrete_meth = getattr(swh_storage, meth_name) except AttributeError: if not getattr(interface_meth, 'deprecated_endpoint', False): # The backend is missing a (non-deprecated) endpoint missing_methods.append(meth_name) continue expected_signature = inspect.signature(interface_meth) actual_signature = inspect.signature(concrete_meth) assert expected_signature == actual_signature, meth_name assert missing_methods == [] def test_check_config(self, swh_storage): assert swh_storage.check_config(check_write=True) assert swh_storage.check_config(check_write=False) def test_content_add(self, swh_storage): cont = data.cont insertion_start_time = datetime.datetime.now(tz=datetime.timezone.utc) actual_result = swh_storage.content_add([cont]) insertion_end_time = datetime.datetime.now(tz=datetime.timezone.utc) assert actual_result == { 'content:add': 1, 'content:add:bytes': cont['length'], - 'skipped_content:add': 0 } assert list(swh_storage.content_get([cont['sha1']])) == \ [{'sha1': cont['sha1'], 'data': cont['data']}] expected_cont = data.cont del expected_cont['data'] journal_objects = list(swh_storage.journal_writer.objects) for (obj_type, obj) in journal_objects: assert insertion_start_time <= obj['ctime'] assert obj['ctime'] <= insertion_end_time del obj['ctime'] assert journal_objects == [('content', expected_cont)] swh_storage.refresh_stat_counters() assert swh_storage.stat_counters()['content'] == 1 def test_content_add_from_generator(self, swh_storage): def _cnt_gen(): yield data.cont actual_result = swh_storage.content_add(_cnt_gen()) assert actual_result == { 'content:add': 1, 'content:add:bytes': data.cont['length'], - 'skipped_content:add': 0 } swh_storage.refresh_stat_counters() assert swh_storage.stat_counters()['content'] == 1 def test_content_add_validation(self, swh_storage): cont = data.cont + with pytest.raises(ValueError, match='status'): + swh_storage.content_add([{**cont, 'status': 'absent'}]) + with pytest.raises(ValueError, match='status'): swh_storage.content_add([{**cont, 'status': 'foobar'}]) with pytest.raises(ValueError, match="(?i)length"): swh_storage.content_add([{**cont, 'length': -2}]) + with pytest.raises( + (ValueError, TypeError), + match="reason"): + swh_storage.content_add([{**cont, 'reason': 'foobar'}]) + + def test_skipped_content_add_validation(self, swh_storage): + cont = data.cont.copy() + del cont['data'] + + with pytest.raises(ValueError, match='status'): + swh_storage.skipped_content_add([{**cont, 'status': 'visible'}]) + with pytest.raises((ValueError, psycopg2.IntegrityError), match='reason') as cm: - swh_storage.content_add([{**cont, 'status': 'absent'}]) + swh_storage.skipped_content_add([{**cont, 'status': 'absent'}]) if type(cm.value) == psycopg2.IntegrityError: assert cm.exception.pgcode == \ psycopg2.errorcodes.NOT_NULL_VIOLATION - with pytest.raises( - ValueError, - match="^Must not provide a reason if content is not absent.$"): - swh_storage.content_add([{**cont, 'reason': 'foobar'}]) - def test_content_get_missing(self, swh_storage): cont = data.cont swh_storage.content_add([cont]) # Query a single missing content results = list(swh_storage.content_get( [data.cont2['sha1']])) assert results == [None] # Check content_get does not abort after finding a missing content results = list(swh_storage.content_get( [data.cont['sha1'], data.cont2['sha1']])) assert results == [{'sha1': cont['sha1'], 'data': cont['data']}, None] # Check content_get does not discard found countent when it finds # a missing content. results = list(swh_storage.content_get( [data.cont2['sha1'], data.cont['sha1']])) assert results == [None, {'sha1': cont['sha1'], 'data': cont['data']}] def test_content_add_different_input(self, swh_storage): cont = data.cont cont2 = data.cont2 actual_result = swh_storage.content_add([cont, cont2]) assert actual_result == { 'content:add': 2, 'content:add:bytes': cont['length'] + cont2['length'], - 'skipped_content:add': 0 } def test_content_add_twice(self, swh_storage): actual_result = swh_storage.content_add([data.cont]) assert actual_result == { 'content:add': 1, 'content:add:bytes': data.cont['length'], - 'skipped_content:add': 0 } assert len(swh_storage.journal_writer.objects) == 1 actual_result = swh_storage.content_add([data.cont, data.cont2]) assert actual_result == { 'content:add': 1, 'content:add:bytes': data.cont2['length'], - 'skipped_content:add': 0 } - assert len(swh_storage.journal_writer.objects) == 2 + assert 2 <= len(swh_storage.journal_writer.objects) <= 3 assert len(swh_storage.content_find(data.cont)) == 1 assert len(swh_storage.content_find(data.cont2)) == 1 def test_content_add_collision(self, swh_storage): cont1 = data.cont # create (corrupted) content with same sha1{,_git} but != sha256 cont1b = cont1.copy() sha256_array = bytearray(cont1b['sha256']) sha256_array[0] += 1 cont1b['sha256'] = bytes(sha256_array) with pytest.raises(HashCollision) as cm: swh_storage.content_add([cont1, cont1b]) assert cm.value.args[0] in ['sha1', 'sha1_git', 'blake2s256'] def test_content_update(self, swh_storage): swh_storage.journal_writer = None # TODO, not supported cont = copy.deepcopy(data.cont) swh_storage.content_add([cont]) # alter the sha1_git for example cont['sha1_git'] = hash_to_bytes( '3a60a5275d0333bf13468e8b3dcab90f4046e654') swh_storage.content_update([cont], keys=['sha1_git']) results = swh_storage.content_get_metadata([cont['sha1']]) del cont['data'] assert results == {cont['sha1']: [cont]} def test_content_add_metadata(self, swh_storage): cont = data.cont del cont['data'] cont['ctime'] = datetime.datetime.now() actual_result = swh_storage.content_add_metadata([cont]) assert actual_result == { 'content:add': 1, - 'skipped_content:add': 0 } expected_cont = cont.copy() del expected_cont['ctime'] assert swh_storage.content_get_metadata([cont['sha1']]) == { cont['sha1']: [expected_cont] } assert list(swh_storage.journal_writer.objects) == [('content', cont)] def test_content_add_metadata_different_input(self, swh_storage): cont = data.cont del cont['data'] cont['ctime'] = datetime.datetime.now() cont2 = data.cont2 del cont2['data'] cont2['ctime'] = datetime.datetime.now() actual_result = swh_storage.content_add_metadata([cont, cont2]) assert actual_result == { 'content:add': 2, - 'skipped_content:add': 0 } def test_content_add_metadata_collision(self, swh_storage): cont1 = data.cont del cont1['data'] cont1['ctime'] = datetime.datetime.now() # create (corrupted) content with same sha1{,_git} but != sha256 cont1b = cont1.copy() sha256_array = bytearray(cont1b['sha256']) sha256_array[0] += 1 cont1b['sha256'] = bytes(sha256_array) with pytest.raises(HashCollision) as cm: swh_storage.content_add_metadata([cont1, cont1b]) assert cm.value.args[0] in ['sha1', 'sha1_git', 'blake2s256'] def test_skipped_content_add(self, swh_storage): cont = data.skipped_cont cont2 = data.skipped_cont2 cont2['blake2s256'] = None missing = list(swh_storage.skipped_content_missing([cont, cont2])) assert len(missing) == 2 - actual_result = swh_storage.content_add([cont, cont, cont2]) + actual_result = swh_storage.skipped_content_add([cont, cont, cont2]) - assert actual_result == { - 'content:add': 0, - 'content:add:bytes': 0, - 'skipped_content:add': 2, - } + assert 2 <= actual_result.pop('skipped_content:add') <= 3 + assert actual_result == {} missing = list(swh_storage.skipped_content_missing([cont, cont2])) assert missing == [] @pytest.mark.property_based @settings(deadline=None) # this test is very slow @given(strategies.sets( elements=strategies.sampled_from( ['sha256', 'sha1_git', 'blake2s256']), min_size=0)) def test_content_missing(self, swh_storage, algos): algos |= {'sha1'} cont2 = data.cont2 missing_cont = data.missing_cont swh_storage.content_add([cont2]) test_contents = [cont2] missing_per_hash = defaultdict(list) for i in range(256): test_content = missing_cont.copy() for hash in algos: test_content[hash] = bytes([i]) + test_content[hash][1:] missing_per_hash[hash].append(test_content[hash]) test_contents.append(test_content) assert set(swh_storage.content_missing(test_contents)) == \ set(missing_per_hash['sha1']) for hash in algos: assert set(swh_storage.content_missing( test_contents, key_hash=hash)) == set(missing_per_hash[hash]) @pytest.mark.property_based @given(strategies.sets( elements=strategies.sampled_from( ['sha256', 'sha1_git', 'blake2s256']), min_size=0)) def test_content_missing_unknown_algo(self, swh_storage, algos): algos |= {'sha1'} cont2 = data.cont2 missing_cont = data.missing_cont swh_storage.content_add([cont2]) test_contents = [cont2] missing_per_hash = defaultdict(list) for i in range(16): test_content = missing_cont.copy() for hash in algos: test_content[hash] = bytes([i]) + test_content[hash][1:] missing_per_hash[hash].append(test_content[hash]) test_content['nonexisting_algo'] = b'\x00' test_contents.append(test_content) assert set( swh_storage.content_missing(test_contents)) == set( missing_per_hash['sha1']) for hash in algos: assert set(swh_storage.content_missing( test_contents, key_hash=hash)) == set( missing_per_hash[hash]) def test_content_missing_per_sha1(self, swh_storage): # given cont2 = data.cont2 missing_cont = data.missing_cont swh_storage.content_add([cont2]) # when gen = swh_storage.content_missing_per_sha1([cont2['sha1'], missing_cont['sha1']]) # then assert list(gen) == [missing_cont['sha1']] def test_content_missing_per_sha1_git(self, swh_storage): cont = data.cont cont2 = data.cont2 missing_cont = data.missing_cont swh_storage.content_add([cont, cont2]) contents = [cont['sha1_git'], cont2['sha1_git'], missing_cont['sha1_git']] missing_contents = swh_storage.content_missing_per_sha1_git(contents) assert list(missing_contents) == [missing_cont['sha1_git']] def test_content_get_partition(self, swh_storage, swh_contents): """content_get_partition paginates results if limit exceeded""" expected_contents = [c for c in swh_contents if c['status'] != 'absent'] actual_contents = [] for i in range(16): actual_result = swh_storage.content_get_partition(i, 16) assert actual_result['next_page_token'] is None actual_contents.extend(actual_result['contents']) assert_contents_ok( expected_contents, actual_contents, ['sha1']) def test_content_get_partition_full(self, swh_storage, swh_contents): """content_get_partition for a single partition returns all available contents""" expected_contents = [c for c in swh_contents if c['status'] != 'absent'] actual_result = swh_storage.content_get_partition(0, 1) assert actual_result['next_page_token'] is None actual_contents = actual_result['contents'] assert_contents_ok( expected_contents, actual_contents, ['sha1']) def test_content_get_partition_empty(self, swh_storage, swh_contents): """content_get_partition when at least one of the partitions is empty""" expected_contents = {cont['sha1'] for cont in swh_contents if cont['status'] != 'absent'} # nb_partitions = smallest power of 2 such that at least one of # the partitions is empty nb_partitions = 1 << math.floor(math.log2(len(swh_contents)) + 1) seen_sha1s = [] for i in range(nb_partitions): actual_result = swh_storage.content_get_partition( i, nb_partitions, limit=len(swh_contents)+1) for cont in actual_result['contents']: seen_sha1s.append(cont['sha1']) # Limit is higher than the max number of results assert actual_result['next_page_token'] is None assert set(seen_sha1s) == expected_contents def test_content_get_partition_limit_none(self, swh_storage): """content_get_partition call with wrong limit input should fail""" with pytest.raises(ValueError) as e: swh_storage.content_get_partition(1, 16, limit=None) assert e.value.args == ('Development error: limit should not be None',) def test_generate_content_get_partition_pagination( self, swh_storage, swh_contents): """content_get_partition returns contents within range provided""" expected_contents = [c for c in swh_contents if c['status'] != 'absent'] # retrieve contents actual_contents = [] for i in range(4): page_token = None while True: actual_result = swh_storage.content_get_partition( i, 4, limit=3, page_token=page_token) actual_contents.extend(actual_result['contents']) page_token = actual_result['next_page_token'] if page_token is None: break assert_contents_ok( expected_contents, actual_contents, ['sha1']) def test_content_get_metadata(self, swh_storage): cont1 = data.cont cont2 = data.cont2 swh_storage.content_add([cont1, cont2]) actual_md = swh_storage.content_get_metadata( [cont1['sha1'], cont2['sha1']]) # we only retrieve the metadata cont1.pop('data') cont2.pop('data') assert actual_md[cont1['sha1']] == [cont1] assert actual_md[cont2['sha1']] == [cont2] assert len(actual_md.keys()) == 2 def test_content_get_metadata_missing_sha1(self, swh_storage): cont1 = data.cont cont2 = data.cont2 missing_cont = data.missing_cont swh_storage.content_add([cont1, cont2]) actual_contents = swh_storage.content_get_metadata( [missing_cont['sha1']]) assert actual_contents == {missing_cont['sha1']: []} def test_content_get_random(self, swh_storage): swh_storage.content_add([data.cont, data.cont2, data.cont3]) assert swh_storage.content_get_random() in { data.cont['sha1_git'], data.cont2['sha1_git'], data.cont3['sha1_git']} def test_directory_add(self, swh_storage): init_missing = list(swh_storage.directory_missing([data.dir['id']])) assert [data.dir['id']] == init_missing actual_result = swh_storage.directory_add([data.dir]) assert actual_result == {'directory:add': 1} assert list(swh_storage.journal_writer.objects) == \ [('directory', data.dir)] actual_data = list(swh_storage.directory_ls(data.dir['id'])) expected_data = list(transform_entries(data.dir)) assert sorted(expected_data, key=cmpdir) \ == sorted(actual_data, key=cmpdir) after_missing = list(swh_storage.directory_missing([data.dir['id']])) assert after_missing == [] swh_storage.refresh_stat_counters() assert swh_storage.stat_counters()['directory'] == 1 def test_directory_add_from_generator(self, swh_storage): def _dir_gen(): yield data.dir actual_result = swh_storage.directory_add(directories=_dir_gen()) assert actual_result == {'directory:add': 1} assert list(swh_storage.journal_writer.objects) == \ [('directory', data.dir)] swh_storage.refresh_stat_counters() assert swh_storage.stat_counters()['directory'] == 1 def test_directory_add_validation(self, swh_storage): dir_ = copy.deepcopy(data.dir) dir_['entries'][0]['type'] = 'foobar' with pytest.raises(ValueError, match='type.*foobar'): swh_storage.directory_add([dir_]) dir_ = copy.deepcopy(data.dir) del dir_['entries'][0]['target'] with pytest.raises((TypeError, psycopg2.IntegrityError), match='target') as cm: swh_storage.directory_add([dir_]) if type(cm.value) == psycopg2.IntegrityError: assert cm.value.pgcode == psycopg2.errorcodes.NOT_NULL_VIOLATION def test_directory_add_twice(self, swh_storage): actual_result = swh_storage.directory_add([data.dir]) assert actual_result == {'directory:add': 1} assert list(swh_storage.journal_writer.objects) \ == [('directory', data.dir)] actual_result = swh_storage.directory_add([data.dir]) assert actual_result == {'directory:add': 0} assert list(swh_storage.journal_writer.objects) \ == [('directory', data.dir)] def test_directory_get_recursive(self, swh_storage): init_missing = list(swh_storage.directory_missing([data.dir['id']])) assert init_missing == [data.dir['id']] actual_result = swh_storage.directory_add( [data.dir, data.dir2, data.dir3]) assert actual_result == {'directory:add': 3} assert list(swh_storage.journal_writer.objects) == [ ('directory', data.dir), ('directory', data.dir2), ('directory', data.dir3)] # List directory containing a file and an unknown subdirectory actual_data = list(swh_storage.directory_ls( data.dir['id'], recursive=True)) expected_data = list(transform_entries(data.dir)) assert sorted(expected_data, key=cmpdir) \ == sorted(actual_data, key=cmpdir) # List directory containing a file and an unknown subdirectory actual_data = list(swh_storage.directory_ls( data.dir2['id'], recursive=True)) expected_data = list(transform_entries(data.dir2)) assert sorted(expected_data, key=cmpdir) \ == sorted(actual_data, key=cmpdir) # List directory containing a known subdirectory, entries should # be both those of the directory and of the subdir actual_data = list(swh_storage.directory_ls( data.dir3['id'], recursive=True)) expected_data = list(itertools.chain( transform_entries(data.dir3), transform_entries(data.dir, prefix=b'subdir/'))) assert sorted(expected_data, key=cmpdir) \ == sorted(actual_data, key=cmpdir) def test_directory_get_non_recursive(self, swh_storage): init_missing = list(swh_storage.directory_missing([data.dir['id']])) assert init_missing == [data.dir['id']] actual_result = swh_storage.directory_add( [data.dir, data.dir2, data.dir3]) assert actual_result == {'directory:add': 3} assert list(swh_storage.journal_writer.objects) == [ ('directory', data.dir), ('directory', data.dir2), ('directory', data.dir3)] # List directory containing a file and an unknown subdirectory actual_data = list(swh_storage.directory_ls(data.dir['id'])) expected_data = list(transform_entries(data.dir)) assert sorted(expected_data, key=cmpdir) \ == sorted(actual_data, key=cmpdir) # List directory contaiining a single file actual_data = list(swh_storage.directory_ls(data.dir2['id'])) expected_data = list(transform_entries(data.dir2)) assert sorted(expected_data, key=cmpdir) \ == sorted(actual_data, key=cmpdir) # List directory containing a known subdirectory, entries should # only be those of the parent directory, not of the subdir actual_data = list(swh_storage.directory_ls(data.dir3['id'])) expected_data = list(transform_entries(data.dir3)) assert sorted(expected_data, key=cmpdir) \ == sorted(actual_data, key=cmpdir) def test_directory_entry_get_by_path(self, swh_storage): # given init_missing = list(swh_storage.directory_missing([data.dir3['id']])) assert [data.dir3['id']] == init_missing actual_result = swh_storage.directory_add([data.dir3, data.dir4]) assert actual_result == {'directory:add': 2} expected_entries = [ { 'dir_id': data.dir3['id'], 'name': b'foo', 'type': 'file', 'target': data.cont['sha1_git'], 'sha1': None, 'sha1_git': None, 'sha256': None, 'status': None, 'perms': from_disk.DentryPerms.content, 'length': None, }, { 'dir_id': data.dir3['id'], 'name': b'subdir', 'type': 'dir', 'target': data.dir['id'], 'sha1': None, 'sha1_git': None, 'sha256': None, 'status': None, 'perms': from_disk.DentryPerms.directory, 'length': None, }, { 'dir_id': data.dir3['id'], 'name': b'hello', 'type': 'file', 'target': b'12345678901234567890', 'sha1': None, 'sha1_git': None, 'sha256': None, 'status': None, 'perms': from_disk.DentryPerms.content, 'length': None, }, ] # when (all must be found here) for entry, expected_entry in zip( data.dir3['entries'], expected_entries): actual_entry = swh_storage.directory_entry_get_by_path( data.dir3['id'], [entry['name']]) assert actual_entry == expected_entry # same, but deeper for entry, expected_entry in zip( data.dir3['entries'], expected_entries): actual_entry = swh_storage.directory_entry_get_by_path( data.dir4['id'], [b'subdir1', entry['name']]) expected_entry = expected_entry.copy() expected_entry['name'] = b'subdir1/' + expected_entry['name'] assert actual_entry == expected_entry # when (nothing should be found here since data.dir is not persisted.) for entry in data.dir['entries']: actual_entry = swh_storage.directory_entry_get_by_path( data.dir['id'], [entry['name']]) assert actual_entry is None def test_directory_get_random(self, swh_storage): swh_storage.directory_add([data.dir, data.dir2, data.dir3]) assert swh_storage.directory_get_random() in \ {data.dir['id'], data.dir2['id'], data.dir3['id']} def test_revision_add(self, swh_storage): init_missing = swh_storage.revision_missing([data.revision['id']]) assert list(init_missing) == [data.revision['id']] actual_result = swh_storage.revision_add([data.revision]) assert actual_result == {'revision:add': 1} end_missing = swh_storage.revision_missing([data.revision['id']]) assert list(end_missing) == [] assert list(swh_storage.journal_writer.objects) \ == [('revision', data.revision)] # already there so nothing added actual_result = swh_storage.revision_add([data.revision]) assert actual_result == {'revision:add': 0} swh_storage.refresh_stat_counters() assert swh_storage.stat_counters()['revision'] == 1 def test_revision_add_from_generator(self, swh_storage): def _rev_gen(): yield data.revision actual_result = swh_storage.revision_add(_rev_gen()) assert actual_result == {'revision:add': 1} swh_storage.refresh_stat_counters() assert swh_storage.stat_counters()['revision'] == 1 def test_revision_add_validation(self, swh_storage): rev = copy.deepcopy(data.revision) rev['date']['offset'] = 2**16 with pytest.raises((ValueError, psycopg2.DataError), match='offset') as cm: swh_storage.revision_add([rev]) if type(cm.value) == psycopg2.DataError: assert cm.value.pgcode \ == psycopg2.errorcodes.NUMERIC_VALUE_OUT_OF_RANGE rev = copy.deepcopy(data.revision) rev['committer_date']['offset'] = 2**16 with pytest.raises((ValueError, psycopg2.DataError), match='offset') as cm: swh_storage.revision_add([rev]) if type(cm.value) == psycopg2.DataError: assert cm.value.pgcode \ == psycopg2.errorcodes.NUMERIC_VALUE_OUT_OF_RANGE rev = copy.deepcopy(data.revision) rev['type'] = 'foobar' with pytest.raises((ValueError, psycopg2.DataError), match='(?i)type') as cm: swh_storage.revision_add([rev]) if type(cm.value) == psycopg2.DataError: assert cm.value.pgcode == \ psycopg2.errorcodes.INVALID_TEXT_REPRESENTATION def test_revision_add_twice(self, swh_storage): actual_result = swh_storage.revision_add([data.revision]) assert actual_result == {'revision:add': 1} assert list(swh_storage.journal_writer.objects) \ == [('revision', data.revision)] actual_result = swh_storage.revision_add( [data.revision, data.revision2]) assert actual_result == {'revision:add': 1} assert list(swh_storage.journal_writer.objects) \ == [('revision', data.revision), ('revision', data.revision2)] def test_revision_add_name_clash(self, swh_storage): revision1 = data.revision revision2 = data.revision2 revision1['author'] = { 'fullname': b'John Doe ', 'name': b'John Doe', 'email': b'john.doe@example.com' } revision2['author'] = { 'fullname': b'John Doe ', 'name': b'John Doe ', 'email': b'john.doe@example.com ' } actual_result = swh_storage.revision_add([revision1, revision2]) assert actual_result == {'revision:add': 2} def test_revision_log(self, swh_storage): # given # data.revision4 -is-child-of-> data.revision3 swh_storage.revision_add([data.revision3, data.revision4]) # when actual_results = list(swh_storage.revision_log( [data.revision4['id']])) # hack: ids generated for actual_result in actual_results: if 'id' in actual_result['author']: del actual_result['author']['id'] if 'id' in actual_result['committer']: del actual_result['committer']['id'] assert len(actual_results) == 2 # rev4 -child-> rev3 assert actual_results[0] == normalize_entity(data.revision4) assert actual_results[1] == normalize_entity(data.revision3) assert list(swh_storage.journal_writer.objects) == [ ('revision', data.revision3), ('revision', data.revision4)] def test_revision_log_with_limit(self, swh_storage): # given # data.revision4 -is-child-of-> data.revision3 swh_storage.revision_add([data.revision3, data.revision4]) actual_results = list(swh_storage.revision_log( [data.revision4['id']], 1)) # hack: ids generated for actual_result in actual_results: if 'id' in actual_result['author']: del actual_result['author']['id'] if 'id' in actual_result['committer']: del actual_result['committer']['id'] assert len(actual_results) == 1 assert actual_results[0] == data.revision4 def test_revision_log_unknown_revision(self, swh_storage): rev_log = list(swh_storage.revision_log([data.revision['id']])) assert rev_log == [] def test_revision_shortlog(self, swh_storage): # given # data.revision4 -is-child-of-> data.revision3 swh_storage.revision_add([data.revision3, data.revision4]) # when actual_results = list(swh_storage.revision_shortlog( [data.revision4['id']])) assert len(actual_results) == 2 # rev4 -child-> rev3 assert list(actual_results[0]) == short_revision(data.revision4) assert list(actual_results[1]) == short_revision(data.revision3) def test_revision_shortlog_with_limit(self, swh_storage): # given # data.revision4 -is-child-of-> data.revision3 swh_storage.revision_add([data.revision3, data.revision4]) actual_results = list(swh_storage.revision_shortlog( [data.revision4['id']], 1)) assert len(actual_results) == 1 assert list(actual_results[0]) == short_revision(data.revision4) def test_revision_get(self, swh_storage): swh_storage.revision_add([data.revision]) actual_revisions = list(swh_storage.revision_get( [data.revision['id'], data.revision2['id']])) # when if 'id' in actual_revisions[0]['author']: del actual_revisions[0]['author']['id'] # hack: ids are generated if 'id' in actual_revisions[0]['committer']: del actual_revisions[0]['committer']['id'] assert len(actual_revisions) == 2 assert actual_revisions[0] == normalize_entity(data.revision) assert actual_revisions[1] is None def test_revision_get_no_parents(self, swh_storage): swh_storage.revision_add([data.revision3]) get = list(swh_storage.revision_get([data.revision3['id']])) assert len(get) == 1 assert get[0]['parents'] == [] # no parents on this one def test_revision_get_random(self, swh_storage): swh_storage.revision_add( [data.revision, data.revision2, data.revision3]) assert swh_storage.revision_get_random() in \ {data.revision['id'], data.revision2['id'], data.revision3['id']} def test_release_add(self, swh_storage): init_missing = swh_storage.release_missing([data.release['id'], data.release2['id']]) assert [data.release['id'], data.release2['id']] == list(init_missing) actual_result = swh_storage.release_add([data.release, data.release2]) assert actual_result == {'release:add': 2} end_missing = swh_storage.release_missing([data.release['id'], data.release2['id']]) assert list(end_missing) == [] assert list(swh_storage.journal_writer.objects) == [ ('release', data.release), ('release', data.release2)] # already present so nothing added actual_result = swh_storage.release_add([data.release, data.release2]) assert actual_result == {'release:add': 0} swh_storage.refresh_stat_counters() assert swh_storage.stat_counters()['release'] == 2 def test_release_add_from_generator(self, swh_storage): def _rel_gen(): yield data.release yield data.release2 actual_result = swh_storage.release_add(_rel_gen()) assert actual_result == {'release:add': 2} assert list(swh_storage.journal_writer.objects) == [ ('release', data.release), ('release', data.release2)] swh_storage.refresh_stat_counters() assert swh_storage.stat_counters()['release'] == 2 def test_release_add_no_author_date(self, swh_storage): release = data.release release['author'] = None release['date'] = None actual_result = swh_storage.release_add([release]) assert actual_result == {'release:add': 1} end_missing = swh_storage.release_missing([data.release['id']]) assert list(end_missing) == [] assert list(swh_storage.journal_writer.objects) \ == [('release', release)] def test_release_add_validation(self, swh_storage): rel = copy.deepcopy(data.release) rel['date']['offset'] = 2**16 with pytest.raises((ValueError, psycopg2.DataError), match='offset') as cm: swh_storage.release_add([rel]) if type(cm.value) == psycopg2.DataError: assert cm.value.pgcode \ == psycopg2.errorcodes.NUMERIC_VALUE_OUT_OF_RANGE rel = copy.deepcopy(data.release) rel['author'] = None with pytest.raises((ValueError, psycopg2.IntegrityError), match='date') as cm: swh_storage.release_add([rel]) if type(cm.value) == psycopg2.IntegrityError: assert cm.value.pgcode == psycopg2.errorcodes.CHECK_VIOLATION def test_release_add_twice(self, swh_storage): actual_result = swh_storage.release_add([data.release]) assert actual_result == {'release:add': 1} assert list(swh_storage.journal_writer.objects) \ == [('release', data.release)] actual_result = swh_storage.release_add([data.release, data.release2]) assert actual_result == {'release:add': 1} assert list(swh_storage.journal_writer.objects) \ == [('release', data.release), ('release', data.release2)] def test_release_add_name_clash(self, swh_storage): release1 = data.release.copy() release2 = data.release2.copy() release1['author'] = { 'fullname': b'John Doe ', 'name': b'John Doe', 'email': b'john.doe@example.com' } release2['author'] = { 'fullname': b'John Doe ', 'name': b'John Doe ', 'email': b'john.doe@example.com ' } actual_result = swh_storage.release_add([release1, release2]) assert actual_result == {'release:add': 2} def test_release_get(self, swh_storage): # given swh_storage.release_add([data.release, data.release2]) # when actual_releases = list(swh_storage.release_get([data.release['id'], data.release2['id']])) # then for actual_release in actual_releases: if 'id' in actual_release['author']: del actual_release['author']['id'] # hack: ids are generated assert [ normalize_entity(data.release), normalize_entity(data.release2)] \ == [actual_releases[0], actual_releases[1]] unknown_releases = \ list(swh_storage.release_get([data.release3['id']])) assert unknown_releases[0] is None def test_release_get_random(self, swh_storage): swh_storage.release_add([data.release, data.release2, data.release3]) assert swh_storage.release_get_random() in \ {data.release['id'], data.release2['id'], data.release3['id']} def test_origin_add_one(self, swh_storage): origin0 = swh_storage.origin_get(data.origin) assert origin0 is None id = swh_storage.origin_add_one(data.origin) actual_origin = swh_storage.origin_get({'url': data.origin['url']}) assert actual_origin['url'] == data.origin['url'] id2 = swh_storage.origin_add_one(data.origin) assert id == id2 def test_origin_add(self, swh_storage): origin0 = swh_storage.origin_get([data.origin])[0] assert origin0 is None origin1, origin2 = swh_storage.origin_add([data.origin, data.origin2]) actual_origin = swh_storage.origin_get([{ 'url': data.origin['url'], }])[0] assert actual_origin['url'] == origin1['url'] actual_origin2 = swh_storage.origin_get([{ 'url': data.origin2['url'], }])[0] assert actual_origin2['url'] == origin2['url'] if 'id' in actual_origin: del actual_origin['id'] del actual_origin2['id'] assert list(swh_storage.journal_writer.objects) \ == [('origin', actual_origin), ('origin', actual_origin2)] swh_storage.refresh_stat_counters() assert swh_storage.stat_counters()['origin'] == 2 def test_origin_add_from_generator(self, swh_storage): def _ori_gen(): yield data.origin yield data.origin2 origin1, origin2 = swh_storage.origin_add(_ori_gen()) actual_origin = swh_storage.origin_get([{ 'url': data.origin['url'], }])[0] assert actual_origin['url'] == origin1['url'] actual_origin2 = swh_storage.origin_get([{ 'url': data.origin2['url'], }])[0] assert actual_origin2['url'] == origin2['url'] if 'id' in actual_origin: del actual_origin['id'] del actual_origin2['id'] assert list(swh_storage.journal_writer.objects) \ == [('origin', actual_origin), ('origin', actual_origin2)] swh_storage.refresh_stat_counters() assert swh_storage.stat_counters()['origin'] == 2 def test_origin_add_twice(self, swh_storage): add1 = swh_storage.origin_add([data.origin, data.origin2]) assert list(swh_storage.journal_writer.objects) \ == [('origin', data.origin), ('origin', data.origin2)] add2 = swh_storage.origin_add([data.origin, data.origin2]) assert list(swh_storage.journal_writer.objects) \ == [('origin', data.origin), ('origin', data.origin2)] assert add1 == add2 def test_origin_add_validation(self, swh_storage): with pytest.raises((TypeError, KeyError), match='url'): swh_storage.origin_add([{'type': 'git'}]) def test_origin_get_legacy(self, swh_storage): assert swh_storage.origin_get(data.origin) is None swh_storage.origin_add_one(data.origin) actual_origin0 = swh_storage.origin_get( {'url': data.origin['url']}) assert actual_origin0['url'] == data.origin['url'] def test_origin_get(self, swh_storage): assert swh_storage.origin_get(data.origin) is None swh_storage.origin_add_one(data.origin) actual_origin0 = swh_storage.origin_get( [{'url': data.origin['url']}]) assert len(actual_origin0) == 1 assert actual_origin0[0]['url'] == data.origin['url'] def _generate_random_visits(self, nb_visits=100, start=0, end=7): """Generate random visits within the last 2 months (to avoid computations) """ visits = [] today = datetime.datetime.now(tz=datetime.timezone.utc) for weeks in range(nb_visits, 0, -1): hours = random.randint(0, 24) minutes = random.randint(0, 60) seconds = random.randint(0, 60) days = random.randint(0, 28) weeks = random.randint(start, end) date_visit = today - timedelta( weeks=weeks, hours=hours, minutes=minutes, seconds=seconds, days=days) visits.append(date_visit) return visits def test_origin_visit_get_random(self, swh_storage): swh_storage.origin_add(data.origins) # Add some random visits within the selection range visits = self._generate_random_visits() visit_type = 'git' # Add visits to those origins for origin in data.origins: for date_visit in visits: visit = swh_storage.origin_visit_add( origin['url'], date=date_visit, type=visit_type) swh_storage.origin_visit_update( origin['url'], visit_id=visit['visit'], status='full') swh_storage.refresh_stat_counters() stats = swh_storage.stat_counters() assert stats['origin'] == len(data.origins) assert stats['origin_visit'] == len(data.origins) * len(visits) random_origin_visit = swh_storage.origin_visit_get_random(visit_type) assert random_origin_visit assert random_origin_visit['origin'] is not None original_urls = [o['url'] for o in data.origins] assert random_origin_visit['origin'] in original_urls def test_origin_visit_get_random_nothing_found(self, swh_storage): swh_storage.origin_add(data.origins) visit_type = 'hg' # Add some visits outside of the random generation selection so nothing # will be found by the random selection visits = self._generate_random_visits(nb_visits=3, start=13, end=24) for origin in data.origins: for date_visit in visits: visit = swh_storage.origin_visit_add( origin['url'], date=date_visit, type=visit_type) swh_storage.origin_visit_update( origin['url'], visit_id=visit['visit'], status='full') random_origin_visit = swh_storage.origin_visit_get_random(visit_type) assert random_origin_visit is None def test_origin_get_by_sha1(self, swh_storage): assert swh_storage.origin_get(data.origin) is None swh_storage.origin_add_one(data.origin) origins = list(swh_storage.origin_get_by_sha1([ sha1(data.origin['url']) ])) assert len(origins) == 1 assert origins[0]['url'] == data.origin['url'] def test_origin_get_by_sha1_not_found(self, swh_storage): assert swh_storage.origin_get(data.origin) is None origins = list(swh_storage.origin_get_by_sha1([ sha1(data.origin['url']) ])) assert len(origins) == 1 assert origins[0] is None def test_origin_search_single_result(self, swh_storage): found_origins = list(swh_storage.origin_search(data.origin['url'])) assert len(found_origins) == 0 found_origins = list(swh_storage.origin_search(data.origin['url'], regexp=True)) assert len(found_origins) == 0 swh_storage.origin_add_one(data.origin) origin_data = { 'url': data.origin['url']} found_origins = list(swh_storage.origin_search(data.origin['url'])) assert len(found_origins) == 1 if 'id' in found_origins[0]: del found_origins[0]['id'] assert found_origins[0] == origin_data found_origins = list(swh_storage.origin_search( '.' + data.origin['url'][1:-1] + '.', regexp=True)) assert len(found_origins) == 1 if 'id' in found_origins[0]: del found_origins[0]['id'] assert found_origins[0] == origin_data swh_storage.origin_add_one(data.origin2) origin2_data = {'url': data.origin2['url']} found_origins = list(swh_storage.origin_search(data.origin2['url'])) assert len(found_origins) == 1 if 'id' in found_origins[0]: del found_origins[0]['id'] assert found_origins[0] == origin2_data found_origins = list(swh_storage.origin_search( '.' + data.origin2['url'][1:-1] + '.', regexp=True)) assert len(found_origins) == 1 if 'id' in found_origins[0]: del found_origins[0]['id'] assert found_origins[0] == origin2_data def test_origin_search_no_regexp(self, swh_storage): swh_storage.origin_add_one(data.origin) swh_storage.origin_add_one(data.origin2) origin = swh_storage.origin_get({'url': data.origin['url']}) origin2 = swh_storage.origin_get({'url': data.origin2['url']}) # no pagination found_origins = list(swh_storage.origin_search('/')) assert len(found_origins) == 2 # offset=0 found_origins0 = list(swh_storage.origin_search('/', offset=0, limit=1)) # noqa assert len(found_origins0) == 1 assert found_origins0[0] in [origin, origin2] # offset=1 found_origins1 = list(swh_storage.origin_search('/', offset=1, limit=1)) # noqa assert len(found_origins1) == 1 assert found_origins1[0] in [origin, origin2] # check both origins were returned assert found_origins0 != found_origins1 def test_origin_search_regexp_substring(self, swh_storage): swh_storage.origin_add_one(data.origin) swh_storage.origin_add_one(data.origin2) origin = swh_storage.origin_get({'url': data.origin['url']}) origin2 = swh_storage.origin_get({'url': data.origin2['url']}) # no pagination found_origins = list(swh_storage.origin_search('/', regexp=True)) assert len(found_origins) == 2 # offset=0 found_origins0 = list(swh_storage.origin_search('/', offset=0, limit=1, regexp=True)) # noqa assert len(found_origins0) == 1 assert found_origins0[0] in [origin, origin2] # offset=1 found_origins1 = list(swh_storage.origin_search('/', offset=1, limit=1, regexp=True)) # noqa assert len(found_origins1) == 1 assert found_origins1[0] in [origin, origin2] # check both origins were returned assert found_origins0 != found_origins1 def test_origin_search_regexp_fullstring(self, swh_storage): swh_storage.origin_add_one(data.origin) swh_storage.origin_add_one(data.origin2) origin = swh_storage.origin_get({'url': data.origin['url']}) origin2 = swh_storage.origin_get({'url': data.origin2['url']}) # no pagination found_origins = list(swh_storage.origin_search('.*/.*', regexp=True)) assert len(found_origins) == 2 # offset=0 found_origins0 = list(swh_storage.origin_search('.*/.*', offset=0, limit=1, regexp=True)) # noqa assert len(found_origins0) == 1 assert found_origins0[0] in [origin, origin2] # offset=1 found_origins1 = list(swh_storage.origin_search('.*/.*', offset=1, limit=1, regexp=True)) # noqa assert len(found_origins1) == 1 assert found_origins1[0] in [origin, origin2] # check both origins were returned assert found_origins0 != found_origins1 def test_origin_visit_add(self, swh_storage): # given swh_storage.origin_add_one(data.origin2) origin_url = data.origin2['url'] date_visit = datetime.datetime.now(datetime.timezone.utc) # Round to milliseconds before insertion, so equality doesn't fail # after a round-trip through a DB (eg. Cassandra) date_visit = date_visit.replace( microsecond=round(date_visit.microsecond, -3)) # when origin_visit1 = swh_storage.origin_visit_add( origin_url, type=data.type_visit1, date=date_visit) actual_origin_visits = list(swh_storage.origin_visit_get( origin_url)) assert { 'origin': origin_url, 'date': date_visit, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'status': 'ongoing', 'metadata': None, 'snapshot': None, } in actual_origin_visits origin_visit = { 'origin': origin_url, 'date': date_visit, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'status': 'ongoing', 'metadata': None, 'snapshot': None, } objects = list(swh_storage.journal_writer.objects) assert ('origin', data.origin2) in objects assert ('origin_visit', origin_visit) in objects def test_origin_visit_get__unknown_origin(self, swh_storage): assert [] == list(swh_storage.origin_visit_get('foo')) def test_origin_visit_add_default_type(self, swh_storage): # given swh_storage.origin_add_one(data.origin2) origin_url = data.origin2['url'] # when date_visit = datetime.datetime.now(datetime.timezone.utc) date_visit2 = date_visit + datetime.timedelta(minutes=1) # Round to milliseconds before insertion, so equality doesn't fail # after a round-trip through a DB (eg. Cassandra) date_visit = date_visit.replace( microsecond=round(date_visit.microsecond, -3)) date_visit2 = date_visit2.replace( microsecond=round(date_visit2.microsecond, -3)) origin_visit1 = swh_storage.origin_visit_add( origin_url, date=date_visit, type=data.type_visit1, ) origin_visit2 = swh_storage.origin_visit_add( origin_url, date=date_visit2, type=data.type_visit2, ) # then assert origin_visit1['origin'] == origin_url assert origin_visit1['visit'] is not None actual_origin_visits = list(swh_storage.origin_visit_get( origin_url)) expected_visits = [ { 'origin': origin_url, 'date': date_visit, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'status': 'ongoing', 'metadata': None, 'snapshot': None, }, { 'origin': origin_url, 'date': date_visit2, 'visit': origin_visit2['visit'], 'type': data.type_visit2, 'status': 'ongoing', 'metadata': None, 'snapshot': None, }, ] for visit in expected_visits: assert visit in actual_origin_visits objects = list(swh_storage.journal_writer.objects) assert ('origin', data.origin2) in objects for visit in expected_visits: assert ('origin_visit', visit) in objects def test_origin_visit_add_validation(self, swh_storage): origin_url = swh_storage.origin_add_one(data.origin2) with pytest.raises((TypeError, psycopg2.ProgrammingError)) as cm: swh_storage.origin_visit_add(origin_url, date=[b'foo']) if type(cm.value) == psycopg2.ProgrammingError: assert cm.value.pgcode \ == psycopg2.errorcodes.UNDEFINED_FUNCTION def test_origin_visit_update(self, swh_storage): # given swh_storage.origin_add_one(data.origin) origin_url = data.origin['url'] date_visit = datetime.datetime.now(datetime.timezone.utc) date_visit2 = date_visit + datetime.timedelta(minutes=1) # Round to milliseconds before insertion, so equality doesn't fail # after a round-trip through a DB (eg. Cassandra) date_visit = date_visit.replace( microsecond=round(date_visit.microsecond, -3)) date_visit2 = date_visit2.replace( microsecond=round(date_visit2.microsecond, -3)) origin_visit1 = swh_storage.origin_visit_add( origin_url, date=date_visit, type=data.type_visit1, ) origin_visit2 = swh_storage.origin_visit_add( origin_url, date=date_visit2, type=data.type_visit2 ) swh_storage.origin_add_one(data.origin2) origin_url2 = data.origin2['url'] origin_visit3 = swh_storage.origin_visit_add( origin_url2, date=date_visit2, type=data.type_visit3 ) # when visit1_metadata = { 'contents': 42, 'directories': 22, } swh_storage.origin_visit_update( origin_url, origin_visit1['visit'], status='full', metadata=visit1_metadata) swh_storage.origin_visit_update( origin_url2, origin_visit3['visit'], status='partial') # then actual_origin_visits = list(swh_storage.origin_visit_get( origin_url)) expected_visits = [{ 'origin': origin_url, 'date': date_visit, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'status': 'full', 'metadata': visit1_metadata, 'snapshot': None, }, { 'origin': origin_url, 'date': date_visit2, 'visit': origin_visit2['visit'], 'type': data.type_visit2, 'status': 'ongoing', 'metadata': None, 'snapshot': None, }] for visit in expected_visits: assert visit in actual_origin_visits actual_origin_visits_bis = list(swh_storage.origin_visit_get( origin_url, limit=1)) assert actual_origin_visits_bis == [ { 'origin': origin_url, 'date': date_visit, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'status': 'full', 'metadata': visit1_metadata, 'snapshot': None, }] actual_origin_visits_ter = list(swh_storage.origin_visit_get( origin_url, last_visit=origin_visit1['visit'])) assert actual_origin_visits_ter == [ { 'origin': origin_url, 'date': date_visit2, 'visit': origin_visit2['visit'], 'type': data.type_visit2, 'status': 'ongoing', 'metadata': None, 'snapshot': None, }] actual_origin_visits2 = list(swh_storage.origin_visit_get( origin_url2)) assert actual_origin_visits2 == [ { 'origin': origin_url2, 'date': date_visit2, 'visit': origin_visit3['visit'], 'type': data.type_visit3, 'status': 'partial', 'metadata': None, 'snapshot': None, }] data1 = { 'origin': origin_url, 'date': date_visit, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'status': 'ongoing', 'metadata': None, 'snapshot': None, } data2 = { 'origin': origin_url, 'date': date_visit2, 'visit': origin_visit2['visit'], 'type': data.type_visit2, 'status': 'ongoing', 'metadata': None, 'snapshot': None, } data3 = { 'origin': origin_url2, 'date': date_visit2, 'visit': origin_visit3['visit'], 'type': data.type_visit3, 'status': 'ongoing', 'metadata': None, 'snapshot': None, } data4 = { 'origin': origin_url, 'date': date_visit, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'metadata': visit1_metadata, 'status': 'full', 'snapshot': None, } data5 = { 'origin': origin_url2, 'date': date_visit2, 'visit': origin_visit3['visit'], 'type': data.type_visit3, 'status': 'partial', 'metadata': None, 'snapshot': None, } objects = list(swh_storage.journal_writer.objects) assert ('origin', data.origin) in objects assert ('origin', data.origin2) in objects assert ('origin_visit', data1) in objects assert ('origin_visit', data2) in objects assert ('origin_visit', data3) in objects assert ('origin_visit', data4) in objects assert ('origin_visit', data5) in objects def test_origin_visit_update_validation(self, swh_storage): origin_url = data.origin['url'] swh_storage.origin_add_one(data.origin) visit = swh_storage.origin_visit_add( origin_url, date=data.date_visit2, type=data.type_visit2, ) with pytest.raises((ValueError, psycopg2.DataError), match='status') as cm: swh_storage.origin_visit_update( origin_url, visit['visit'], status='foobar') if type(cm.value) == psycopg2.DataError: assert cm.value.pgcode == \ psycopg2.errorcodes.INVALID_TEXT_REPRESENTATION def test_origin_visit_find_by_date(self, swh_storage): # given swh_storage.origin_add_one(data.origin) swh_storage.origin_visit_add( data.origin['url'], date=data.date_visit2, type=data.type_visit1, ) origin_visit2 = swh_storage.origin_visit_add( data.origin['url'], date=data.date_visit3, type=data.type_visit2, ) origin_visit3 = swh_storage.origin_visit_add( data.origin['url'], date=data.date_visit2, type=data.type_visit3, ) # Simple case visit = swh_storage.origin_visit_find_by_date( data.origin['url'], data.date_visit3) assert visit['visit'] == origin_visit2['visit'] # There are two visits at the same date, the latest must be returned visit = swh_storage.origin_visit_find_by_date( data.origin['url'], data.date_visit2) assert visit['visit'] == origin_visit3['visit'] def test_origin_visit_find_by_date__unknown_origin(self, swh_storage): swh_storage.origin_visit_find_by_date('foo', data.date_visit2) def test_origin_visit_update_missing_snapshot(self, swh_storage): # given swh_storage.origin_add_one(data.origin) origin_url = data.origin['url'] origin_visit = swh_storage.origin_visit_add( origin_url, date=data.date_visit1, type=data.type_visit1, ) # when swh_storage.origin_visit_update( origin_url, origin_visit['visit'], snapshot=data.snapshot['id']) # then actual_origin_visit = swh_storage.origin_visit_get_by( origin_url, origin_visit['visit']) assert actual_origin_visit['snapshot'] == data.snapshot['id'] # when swh_storage.snapshot_add([data.snapshot]) assert actual_origin_visit['snapshot'] == data.snapshot['id'] def test_origin_visit_get_by(self, swh_storage): swh_storage.origin_add_one(data.origin) swh_storage.origin_add_one(data.origin2) origin_url = data.origin['url'] origin2_url = data.origin2['url'] origin_visit1 = swh_storage.origin_visit_add( origin_url, date=data.date_visit2, type=data.type_visit2, ) swh_storage.snapshot_add([data.snapshot]) swh_storage.origin_visit_update( origin_url, origin_visit1['visit'], snapshot=data.snapshot['id']) # Add some other {origin, visit} entries swh_storage.origin_visit_add( origin_url, date=data.date_visit3, type=data.type_visit3, ) swh_storage.origin_visit_add( origin2_url, date=data.date_visit3, type=data.type_visit3, ) # when visit1_metadata = { 'contents': 42, 'directories': 22, } swh_storage.origin_visit_update( origin_url, origin_visit1['visit'], status='full', metadata=visit1_metadata) expected_origin_visit = origin_visit1.copy() expected_origin_visit.update({ 'origin': origin_url, 'visit': origin_visit1['visit'], 'date': data.date_visit2, 'type': data.type_visit2, 'metadata': visit1_metadata, 'status': 'full', 'snapshot': data.snapshot['id'], }) # when actual_origin_visit1 = swh_storage.origin_visit_get_by( origin_url, origin_visit1['visit']) # then assert actual_origin_visit1 == expected_origin_visit def test_origin_visit_get_by__unknown_origin(self, swh_storage): assert swh_storage.origin_visit_get_by('foo', 10) is None def test_origin_visit_upsert_new(self, swh_storage): # given swh_storage.origin_add_one(data.origin2) origin_url = data.origin2['url'] # when swh_storage.origin_visit_upsert([ { 'origin': origin_url, 'date': data.date_visit2, 'visit': 123, 'type': data.type_visit2, 'status': 'full', 'metadata': None, 'snapshot': None, }, { 'origin': origin_url, 'date': '2018-01-01 23:00:00+00', 'visit': 1234, 'type': data.type_visit2, 'status': 'full', 'metadata': None, 'snapshot': None, }, ]) # then actual_origin_visits = list(swh_storage.origin_visit_get( origin_url)) assert actual_origin_visits == [ { 'origin': origin_url, 'date': data.date_visit2, 'visit': 123, 'type': data.type_visit2, 'status': 'full', 'metadata': None, 'snapshot': None, }, { 'origin': origin_url, 'date': data.date_visit3, 'visit': 1234, 'type': data.type_visit2, 'status': 'full', 'metadata': None, 'snapshot': None, }, ] data1 = { 'origin': origin_url, 'date': data.date_visit2, 'visit': 123, 'type': data.type_visit2, 'status': 'full', 'metadata': None, 'snapshot': None, } data2 = { 'origin': origin_url, 'date': data.date_visit3, 'visit': 1234, 'type': data.type_visit2, 'status': 'full', 'metadata': None, 'snapshot': None, } assert list(swh_storage.journal_writer.objects) == [ ('origin', data.origin2), ('origin_visit', data1), ('origin_visit', data2)] def test_origin_visit_upsert_existing(self, swh_storage): # given swh_storage.origin_add_one(data.origin2) origin_url = data.origin2['url'] # when origin_visit1 = swh_storage.origin_visit_add( origin_url, date=data.date_visit2, type=data.type_visit1, ) swh_storage.origin_visit_upsert([{ 'origin': origin_url, 'date': data.date_visit2, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'status': 'full', 'metadata': None, 'snapshot': None, }]) # then assert origin_visit1['origin'] == origin_url assert origin_visit1['visit'] is not None actual_origin_visits = list(swh_storage.origin_visit_get( origin_url)) assert actual_origin_visits == [ { 'origin': origin_url, 'date': data.date_visit2, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'status': 'full', 'metadata': None, 'snapshot': None, }] data1 = { 'origin': origin_url, 'date': data.date_visit2, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'status': 'ongoing', 'metadata': None, 'snapshot': None, } data2 = { 'origin': origin_url, 'date': data.date_visit2, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'status': 'full', 'metadata': None, 'snapshot': None, } assert list(swh_storage.journal_writer.objects) == [ ('origin', data.origin2), ('origin_visit', data1), ('origin_visit', data2)] def test_origin_visit_get_by_no_result(self, swh_storage): swh_storage.origin_add([data.origin]) actual_origin_visit = swh_storage.origin_visit_get_by( data.origin['url'], 999) assert actual_origin_visit is None def test_origin_visit_get_latest(self, swh_storage): swh_storage.origin_add_one(data.origin) origin_url = data.origin['url'] origin_visit1 = swh_storage.origin_visit_add( origin=origin_url, date=data.date_visit1, type=data.type_visit1, ) visit1_id = origin_visit1['visit'] origin_visit2 = swh_storage.origin_visit_add( origin=origin_url, date=data.date_visit2, type=data.type_visit2, ) visit2_id = origin_visit2['visit'] # Add a visit with the same date as the previous one origin_visit3 = swh_storage.origin_visit_add( origin=origin_url, date=data.date_visit2, type=data.type_visit2, ) visit3_id = origin_visit3['visit'] origin_visit1 = swh_storage.origin_visit_get_by(origin_url, visit1_id) origin_visit2 = swh_storage.origin_visit_get_by(origin_url, visit2_id) origin_visit3 = swh_storage.origin_visit_get_by(origin_url, visit3_id) # Two visits, both with no snapshot assert origin_visit3 == swh_storage.origin_visit_get_latest(origin_url) assert swh_storage.origin_visit_get_latest( origin_url, require_snapshot=True) is None # Add snapshot to visit1; require_snapshot=True makes it return # visit1 and require_snapshot=False still returns visit2 swh_storage.snapshot_add([data.complete_snapshot]) swh_storage.origin_visit_update( origin_url, visit1_id, snapshot=data.complete_snapshot['id']) assert {**origin_visit1, 'snapshot': data.complete_snapshot['id']} \ == swh_storage.origin_visit_get_latest( origin_url, require_snapshot=True) assert origin_visit3 == swh_storage.origin_visit_get_latest(origin_url) # Status filter: all three visits are status=ongoing, so no visit # returned assert swh_storage.origin_visit_get_latest( origin_url, allowed_statuses=['full']) is None # Mark the first visit as completed and check status filter again swh_storage.origin_visit_update( origin_url, visit1_id, status='full') assert { **origin_visit1, 'snapshot': data.complete_snapshot['id'], 'status': 'full'} == swh_storage.origin_visit_get_latest( origin_url, allowed_statuses=['full']) assert origin_visit3 == swh_storage.origin_visit_get_latest(origin_url) # Add snapshot to visit2 and check that the new snapshot is returned swh_storage.snapshot_add([data.empty_snapshot]) swh_storage.origin_visit_update( origin_url, visit2_id, snapshot=data.empty_snapshot['id']) assert {**origin_visit2, 'snapshot': data.empty_snapshot['id']} == \ swh_storage.origin_visit_get_latest( origin_url, require_snapshot=True) assert origin_visit3 == swh_storage.origin_visit_get_latest(origin_url) # Check that the status filter is still working assert { **origin_visit1, 'snapshot': data.complete_snapshot['id'], 'status': 'full'} == swh_storage.origin_visit_get_latest( origin_url, allowed_statuses=['full']) # Add snapshot to visit3 (same date as visit2) swh_storage.snapshot_add([data.complete_snapshot]) swh_storage.origin_visit_update( origin_url, visit3_id, snapshot=data.complete_snapshot['id']) assert { **origin_visit1, 'snapshot': data.complete_snapshot['id'], 'status': 'full'} == swh_storage.origin_visit_get_latest( origin_url, allowed_statuses=['full']) assert { **origin_visit1, 'snapshot': data.complete_snapshot['id'], 'status': 'full'} == swh_storage.origin_visit_get_latest( origin_url, allowed_statuses=['full'], require_snapshot=True) assert { **origin_visit3, 'snapshot': data.complete_snapshot['id'] } == swh_storage.origin_visit_get_latest(origin_url) assert { **origin_visit3, 'snapshot': data.complete_snapshot['id'] } == swh_storage.origin_visit_get_latest( origin_url, require_snapshot=True) def test_person_fullname_unicity(self, swh_storage): # given (person injection through revisions for example) revision = data.revision # create a revision with same committer fullname but wo name and email revision2 = copy.deepcopy(data.revision2) revision2['committer'] = dict(revision['committer']) revision2['committer']['email'] = None revision2['committer']['name'] = None swh_storage.revision_add([revision]) swh_storage.revision_add([revision2]) # when getting added revisions revisions = list( swh_storage.revision_get([revision['id'], revision2['id']])) # then # check committers are the same assert revisions[0]['committer'] == revisions[1]['committer'] def test_snapshot_add_get_empty(self, swh_storage): origin_url = data.origin['url'] swh_storage.origin_add_one(data.origin) origin_visit1 = swh_storage.origin_visit_add( origin=origin_url, date=data.date_visit1, type=data.type_visit1, ) visit_id = origin_visit1['visit'] actual_result = swh_storage.snapshot_add([data.empty_snapshot]) assert actual_result == {'snapshot:add': 1} swh_storage.origin_visit_update( origin_url, visit_id, snapshot=data.empty_snapshot['id']) by_id = swh_storage.snapshot_get(data.empty_snapshot['id']) assert by_id == {**data.empty_snapshot, 'next_branch': None} by_ov = swh_storage.snapshot_get_by_origin_visit(origin_url, visit_id) assert by_ov == {**data.empty_snapshot, 'next_branch': None} data1 = { 'origin': origin_url, 'date': data.date_visit1, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'status': 'ongoing', 'metadata': None, 'snapshot': None, } data2 = { 'origin': origin_url, 'date': data.date_visit1, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'status': 'ongoing', 'metadata': None, 'snapshot': data.empty_snapshot['id'], } assert list(swh_storage.journal_writer.objects) == \ [('origin', data.origin), ('origin_visit', data1), ('snapshot', data.empty_snapshot), ('origin_visit', data2)] def test_snapshot_add_get_complete(self, swh_storage): origin_url = data.origin['url'] swh_storage.origin_add_one(data.origin) origin_visit1 = swh_storage.origin_visit_add( origin=origin_url, date=data.date_visit1, type=data.type_visit1, ) visit_id = origin_visit1['visit'] actual_result = swh_storage.snapshot_add([data.complete_snapshot]) swh_storage.origin_visit_update( origin_url, visit_id, snapshot=data.complete_snapshot['id']) assert actual_result == {'snapshot:add': 1} by_id = swh_storage.snapshot_get(data.complete_snapshot['id']) assert by_id == {**data.complete_snapshot, 'next_branch': None} by_ov = swh_storage.snapshot_get_by_origin_visit(origin_url, visit_id) assert by_ov == {**data.complete_snapshot, 'next_branch': None} def test_snapshot_add_many(self, swh_storage): actual_result = swh_storage.snapshot_add( [data.snapshot, data.complete_snapshot]) assert actual_result == {'snapshot:add': 2} assert {**data.complete_snapshot, 'next_branch': None} \ == swh_storage.snapshot_get(data.complete_snapshot['id']) assert {**data.snapshot, 'next_branch': None} \ == swh_storage.snapshot_get(data.snapshot['id']) swh_storage.refresh_stat_counters() assert swh_storage.stat_counters()['snapshot'] == 2 def test_snapshot_add_many_from_generator(self, swh_storage): def _snp_gen(): yield data.snapshot yield data.complete_snapshot actual_result = swh_storage.snapshot_add(_snp_gen()) assert actual_result == {'snapshot:add': 2} swh_storage.refresh_stat_counters() assert swh_storage.stat_counters()['snapshot'] == 2 def test_snapshot_add_many_incremental(self, swh_storage): actual_result = swh_storage.snapshot_add([data.complete_snapshot]) assert actual_result == {'snapshot:add': 1} actual_result2 = swh_storage.snapshot_add( [data.snapshot, data.complete_snapshot]) assert actual_result2 == {'snapshot:add': 1} assert {**data.complete_snapshot, 'next_branch': None} \ == swh_storage.snapshot_get(data.complete_snapshot['id']) assert {**data.snapshot, 'next_branch': None} \ == swh_storage.snapshot_get(data.snapshot['id']) def test_snapshot_add_twice(self, swh_storage): actual_result = swh_storage.snapshot_add([data.empty_snapshot]) assert actual_result == {'snapshot:add': 1} assert list(swh_storage.journal_writer.objects) \ == [('snapshot', data.empty_snapshot)] actual_result = swh_storage.snapshot_add([data.snapshot]) assert actual_result == {'snapshot:add': 1} assert list(swh_storage.journal_writer.objects) \ == [('snapshot', data.empty_snapshot), ('snapshot', data.snapshot)] def test_snapshot_add_validation(self, swh_storage): snap = copy.deepcopy(data.snapshot) snap['branches'][b'foo'] = {'target_type': 'revision'} with pytest.raises(KeyError, match='target'): swh_storage.snapshot_add([snap]) snap = copy.deepcopy(data.snapshot) snap['branches'][b'foo'] = {'target': b'\x42'*20} with pytest.raises(KeyError, match='target_type'): swh_storage.snapshot_add([snap]) def test_snapshot_add_count_branches(self, swh_storage): actual_result = swh_storage.snapshot_add([data.complete_snapshot]) assert actual_result == {'snapshot:add': 1} snp_id = data.complete_snapshot['id'] snp_size = swh_storage.snapshot_count_branches(snp_id) expected_snp_size = { 'alias': 1, 'content': 1, 'directory': 2, 'release': 1, 'revision': 1, 'snapshot': 1, None: 1 } assert snp_size == expected_snp_size def test_snapshot_add_get_paginated(self, swh_storage): swh_storage.snapshot_add([data.complete_snapshot]) snp_id = data.complete_snapshot['id'] branches = data.complete_snapshot['branches'] branch_names = list(sorted(branches)) # Test branch_from snapshot = swh_storage.snapshot_get_branches( snp_id, branches_from=b'release') rel_idx = branch_names.index(b'release') expected_snapshot = { 'id': snp_id, 'branches': { name: branches[name] for name in branch_names[rel_idx:] }, 'next_branch': None, } assert snapshot == expected_snapshot # Test branches_count snapshot = swh_storage.snapshot_get_branches( snp_id, branches_count=1) expected_snapshot = { 'id': snp_id, 'branches': { branch_names[0]: branches[branch_names[0]], }, 'next_branch': b'content', } assert snapshot == expected_snapshot # test branch_from + branches_count snapshot = swh_storage.snapshot_get_branches( snp_id, branches_from=b'directory', branches_count=3) dir_idx = branch_names.index(b'directory') expected_snapshot = { 'id': snp_id, 'branches': { name: branches[name] for name in branch_names[dir_idx:dir_idx + 3] }, 'next_branch': branch_names[dir_idx + 3], } assert snapshot == expected_snapshot def test_snapshot_add_get_filtered(self, swh_storage): origin_url = data.origin['url'] swh_storage.origin_add_one(data.origin) origin_visit1 = swh_storage.origin_visit_add( origin=origin_url, date=data.date_visit1, type=data.type_visit1, ) visit_id = origin_visit1['visit'] swh_storage.snapshot_add([data.complete_snapshot]) swh_storage.origin_visit_update( origin_url, visit_id, snapshot=data.complete_snapshot['id']) snp_id = data.complete_snapshot['id'] branches = data.complete_snapshot['branches'] snapshot = swh_storage.snapshot_get_branches( snp_id, target_types=['release', 'revision']) expected_snapshot = { 'id': snp_id, 'branches': { name: tgt for name, tgt in branches.items() if tgt and tgt['target_type'] in ['release', 'revision'] }, 'next_branch': None, } assert snapshot == expected_snapshot snapshot = swh_storage.snapshot_get_branches( snp_id, target_types=['alias']) expected_snapshot = { 'id': snp_id, 'branches': { name: tgt for name, tgt in branches.items() if tgt and tgt['target_type'] == 'alias' }, 'next_branch': None, } assert snapshot == expected_snapshot def test_snapshot_add_get_filtered_and_paginated(self, swh_storage): swh_storage.snapshot_add([data.complete_snapshot]) snp_id = data.complete_snapshot['id'] branches = data.complete_snapshot['branches'] branch_names = list(sorted(branches)) # Test branch_from snapshot = swh_storage.snapshot_get_branches( snp_id, target_types=['directory', 'release'], branches_from=b'directory2') expected_snapshot = { 'id': snp_id, 'branches': { name: branches[name] for name in (b'directory2', b'release') }, 'next_branch': None, } assert snapshot == expected_snapshot # Test branches_count snapshot = swh_storage.snapshot_get_branches( snp_id, target_types=['directory', 'release'], branches_count=1) expected_snapshot = { 'id': snp_id, 'branches': { b'directory': branches[b'directory'] }, 'next_branch': b'directory2', } assert snapshot == expected_snapshot # Test branches_count snapshot = swh_storage.snapshot_get_branches( snp_id, target_types=['directory', 'release'], branches_count=2) expected_snapshot = { 'id': snp_id, 'branches': { name: branches[name] for name in (b'directory', b'directory2') }, 'next_branch': b'release', } assert snapshot == expected_snapshot # test branch_from + branches_count snapshot = swh_storage.snapshot_get_branches( snp_id, target_types=['directory', 'release'], branches_from=b'directory2', branches_count=1) dir_idx = branch_names.index(b'directory2') expected_snapshot = { 'id': snp_id, 'branches': { branch_names[dir_idx]: branches[branch_names[dir_idx]], }, 'next_branch': b'release', } assert snapshot == expected_snapshot def test_snapshot_add_get(self, swh_storage): origin_url = data.origin['url'] swh_storage.origin_add_one(data.origin) origin_visit1 = swh_storage.origin_visit_add( origin=origin_url, date=data.date_visit1, type=data.type_visit1, ) visit_id = origin_visit1['visit'] swh_storage.snapshot_add([data.snapshot]) swh_storage.origin_visit_update( origin_url, visit_id, snapshot=data.snapshot['id']) by_id = swh_storage.snapshot_get(data.snapshot['id']) assert by_id == {**data.snapshot, 'next_branch': None} by_ov = swh_storage.snapshot_get_by_origin_visit(origin_url, visit_id) assert by_ov == {**data.snapshot, 'next_branch': None} origin_visit_info = swh_storage.origin_visit_get_by( origin_url, visit_id) assert origin_visit_info['snapshot'] == data.snapshot['id'] def test_snapshot_add_nonexistent_visit(self, swh_storage): origin_url = data.origin['url'] swh_storage.origin_add_one(data.origin) visit_id = 54164461156 swh_storage.journal_writer.objects[:] = [] swh_storage.snapshot_add([data.snapshot]) with pytest.raises(ValueError): swh_storage.origin_visit_update( origin_url, visit_id, snapshot=data.snapshot['id']) assert list(swh_storage.journal_writer.objects) == [ ('snapshot', data.snapshot)] def test_snapshot_add_twice__by_origin_visit(self, swh_storage): origin_url = data.origin['url'] swh_storage.origin_add_one(data.origin) origin_visit1 = swh_storage.origin_visit_add( origin=origin_url, date=data.date_visit1, type=data.type_visit1, ) visit1_id = origin_visit1['visit'] swh_storage.snapshot_add([data.snapshot]) swh_storage.origin_visit_update( origin_url, visit1_id, snapshot=data.snapshot['id']) by_ov1 = swh_storage.snapshot_get_by_origin_visit( origin_url, visit1_id) assert by_ov1 == {**data.snapshot, 'next_branch': None} origin_visit2 = swh_storage.origin_visit_add( origin=origin_url, date=data.date_visit2, type=data.type_visit2, ) visit2_id = origin_visit2['visit'] swh_storage.snapshot_add([data.snapshot]) swh_storage.origin_visit_update( origin_url, visit2_id, snapshot=data.snapshot['id']) by_ov2 = swh_storage.snapshot_get_by_origin_visit( origin_url, visit2_id) assert by_ov2 == {**data.snapshot, 'next_branch': None} data1 = { 'origin': origin_url, 'date': data.date_visit1, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'status': 'ongoing', 'metadata': None, 'snapshot': None, } data2 = { 'origin': origin_url, 'date': data.date_visit1, 'visit': origin_visit1['visit'], 'type': data.type_visit1, 'status': 'ongoing', 'metadata': None, 'snapshot': data.snapshot['id'], } data3 = { 'origin': origin_url, 'date': data.date_visit2, 'visit': origin_visit2['visit'], 'type': data.type_visit2, 'status': 'ongoing', 'metadata': None, 'snapshot': None, } data4 = { 'origin': origin_url, 'date': data.date_visit2, 'visit': origin_visit2['visit'], 'type': data.type_visit2, 'status': 'ongoing', 'metadata': None, 'snapshot': data.snapshot['id'], } assert list(swh_storage.journal_writer.objects) \ == [('origin', data.origin), ('origin_visit', data1), ('snapshot', data.snapshot), ('origin_visit', data2), ('origin_visit', data3), ('origin_visit', data4)] def test_snapshot_get_latest(self, swh_storage): origin_url = data.origin['url'] swh_storage.origin_add_one(data.origin) origin_url = data.origin['url'] origin_visit1 = swh_storage.origin_visit_add( origin=origin_url, date=data.date_visit1, type=data.type_visit1, ) visit1_id = origin_visit1['visit'] origin_visit2 = swh_storage.origin_visit_add( origin=origin_url, date=data.date_visit2, type=data.type_visit2, ) visit2_id = origin_visit2['visit'] # Add a visit with the same date as the previous one origin_visit3 = swh_storage.origin_visit_add( origin=origin_url, date=data.date_visit2, type=data.type_visit3, ) visit3_id = origin_visit3['visit'] # Two visits, both with no snapshot: latest snapshot is None assert swh_storage.snapshot_get_latest(origin_url) is None # Add snapshot to visit1, latest snapshot = visit 1 snapshot swh_storage.snapshot_add([data.complete_snapshot]) swh_storage.origin_visit_update( origin_url, visit1_id, snapshot=data.complete_snapshot['id']) assert {**data.complete_snapshot, 'next_branch': None} \ == swh_storage.snapshot_get_latest(origin_url) # Status filter: all three visits are status=ongoing, so no snapshot # returned assert swh_storage.snapshot_get_latest( origin_url, allowed_statuses=['full']) is None # Mark the first visit as completed and check status filter again swh_storage.origin_visit_update(origin_url, visit1_id, status='full') assert {**data.complete_snapshot, 'next_branch': None} \ == swh_storage.snapshot_get_latest( origin_url, allowed_statuses=['full']) # Add snapshot to visit2 and check that the new snapshot is returned swh_storage.snapshot_add([data.empty_snapshot]) swh_storage.origin_visit_update( origin_url, visit2_id, snapshot=data.empty_snapshot['id']) assert {**data.empty_snapshot, 'next_branch': None} \ == swh_storage.snapshot_get_latest(origin_url) # Check that the status filter is still working assert {**data.complete_snapshot, 'next_branch': None} \ == swh_storage.snapshot_get_latest( origin_url, allowed_statuses=['full']) # Add snapshot to visit3 (same date as visit2) and check that # the new snapshot is returned swh_storage.snapshot_add([data.complete_snapshot]) swh_storage.origin_visit_update( origin_url, visit3_id, snapshot=data.complete_snapshot['id']) assert {**data.complete_snapshot, 'next_branch': None} \ == swh_storage.snapshot_get_latest(origin_url) def test_snapshot_get_latest__missing_snapshot(self, swh_storage): # Origin does not exist origin_url = data.origin['url'] assert swh_storage.snapshot_get_latest(origin_url) is None swh_storage.origin_add_one(data.origin) origin_visit1 = swh_storage.origin_visit_add( origin=origin_url, date=data.date_visit1, type=data.type_visit1, ) visit1_id = origin_visit1['visit'] origin_visit2 = swh_storage.origin_visit_add( origin=origin_url, date=data.date_visit2, type=data.type_visit2, ) visit2_id = origin_visit2['visit'] # Two visits, both with no snapshot: latest snapshot is None assert swh_storage.snapshot_get_latest(origin_url) is None # Add unknown snapshot to visit1, check that the inconsistency is # detected swh_storage.origin_visit_update( origin_url, visit1_id, snapshot=data.complete_snapshot['id']) with pytest.raises(ValueError): swh_storage.snapshot_get_latest( origin_url) # Status filter: both visits are status=ongoing, so no snapshot # returned assert swh_storage.snapshot_get_latest( origin_url, allowed_statuses=['full']) is None # Mark the first visit as completed and check status filter again swh_storage.origin_visit_update( origin_url, visit1_id, status='full') with pytest.raises(ValueError): swh_storage.snapshot_get_latest( origin_url, allowed_statuses=['full']), # Actually add the snapshot and check status filter again swh_storage.snapshot_add([data.complete_snapshot]) assert {**data.complete_snapshot, 'next_branch': None} \ == swh_storage.snapshot_get_latest(origin_url) # Add unknown snapshot to visit2 and check that the inconsistency # is detected swh_storage.origin_visit_update( origin_url, visit2_id, snapshot=data.snapshot['id']) with pytest.raises(ValueError): swh_storage.snapshot_get_latest( origin_url) # Actually add that snapshot and check that the new one is returned swh_storage.snapshot_add([data.snapshot]) assert{**data.snapshot, 'next_branch': None} \ == swh_storage.snapshot_get_latest(origin_url) def test_snapshot_get_random(self, swh_storage): swh_storage.snapshot_add( [data.snapshot, data.empty_snapshot, data.complete_snapshot]) assert swh_storage.snapshot_get_random() in { data.snapshot['id'], data.empty_snapshot['id'], data.complete_snapshot['id']} def test_snapshot_missing(self, swh_storage): snap = data.snapshot missing_snap = data.empty_snapshot snapshots = [snap['id'], missing_snap['id']] swh_storage.snapshot_add([snap]) missing_snapshots = swh_storage.snapshot_missing(snapshots) assert list(missing_snapshots) == [missing_snap['id']] def test_stat_counters(self, swh_storage): expected_keys = ['content', 'directory', 'origin', 'revision'] # Initially, all counters are 0 swh_storage.refresh_stat_counters() counters = swh_storage.stat_counters() assert set(expected_keys) <= set(counters) for key in expected_keys: assert counters[key] == 0 # Add a content. Only the content counter should increase. swh_storage.content_add([data.cont]) swh_storage.refresh_stat_counters() counters = swh_storage.stat_counters() assert set(expected_keys) <= set(counters) for key in expected_keys: if key != 'content': assert counters[key] == 0 assert counters['content'] == 1 # Add other objects. Check their counter increased as well. swh_storage.origin_add_one(data.origin2) origin_visit1 = swh_storage.origin_visit_add( origin=data.origin2['url'], date=data.date_visit2, type=data.type_visit2, ) swh_storage.snapshot_add([data.snapshot]) swh_storage.origin_visit_update( data.origin2['url'], origin_visit1['visit'], snapshot=data.snapshot['id']) swh_storage.directory_add([data.dir]) swh_storage.revision_add([data.revision]) swh_storage.release_add([data.release]) swh_storage.refresh_stat_counters() counters = swh_storage.stat_counters() assert counters['content'] == 1 assert counters['directory'] == 1 assert counters['snapshot'] == 1 assert counters['origin'] == 1 assert counters['origin_visit'] == 1 assert counters['revision'] == 1 assert counters['release'] == 1 assert counters['snapshot'] == 1 if 'person' in counters: assert counters['person'] == 3 def test_content_find_ctime(self, swh_storage): cont = data.cont.copy() del cont['data'] now = datetime.datetime.now(tz=datetime.timezone.utc) cont['ctime'] = now swh_storage.content_add_metadata([cont]) actually_present = swh_storage.content_find({'sha1': cont['sha1']}) # check ctime up to one second dt = actually_present[0]['ctime'] - now assert abs(dt.total_seconds()) <= 1 del actually_present[0]['ctime'] assert actually_present[0] == { 'sha1': cont['sha1'], 'sha256': cont['sha256'], 'sha1_git': cont['sha1_git'], 'blake2s256': cont['blake2s256'], 'length': cont['length'], 'status': 'visible' } def test_content_find_with_present_content(self, swh_storage): # 1. with something to find cont = data.cont swh_storage.content_add([cont, data.cont2]) actually_present = swh_storage.content_find( {'sha1': cont['sha1']} ) assert 1 == len(actually_present) actually_present[0].pop('ctime') assert actually_present[0] == { 'sha1': cont['sha1'], 'sha256': cont['sha256'], 'sha1_git': cont['sha1_git'], 'blake2s256': cont['blake2s256'], 'length': cont['length'], 'status': 'visible' } # 2. with something to find actually_present = swh_storage.content_find( {'sha1_git': cont['sha1_git']}) assert 1 == len(actually_present) actually_present[0].pop('ctime') assert actually_present[0] == { 'sha1': cont['sha1'], 'sha256': cont['sha256'], 'sha1_git': cont['sha1_git'], 'blake2s256': cont['blake2s256'], 'length': cont['length'], 'status': 'visible' } # 3. with something to find actually_present = swh_storage.content_find( {'sha256': cont['sha256']}) assert 1 == len(actually_present) actually_present[0].pop('ctime') assert actually_present[0] == { 'sha1': cont['sha1'], 'sha256': cont['sha256'], 'sha1_git': cont['sha1_git'], 'blake2s256': cont['blake2s256'], 'length': cont['length'], 'status': 'visible' } # 4. with something to find actually_present = swh_storage.content_find({ 'sha1': cont['sha1'], 'sha1_git': cont['sha1_git'], 'sha256': cont['sha256'], 'blake2s256': cont['blake2s256'], }) assert 1 == len(actually_present) actually_present[0].pop('ctime') assert actually_present[0] == { 'sha1': cont['sha1'], 'sha256': cont['sha256'], 'sha1_git': cont['sha1_git'], 'blake2s256': cont['blake2s256'], 'length': cont['length'], 'status': 'visible' } def test_content_find_with_non_present_content(self, swh_storage): # 1. with something that does not exist missing_cont = data.missing_cont actually_present = swh_storage.content_find( {'sha1': missing_cont['sha1']}) assert actually_present == [] # 2. with something that does not exist actually_present = swh_storage.content_find( {'sha1_git': missing_cont['sha1_git']}) assert actually_present == [] # 3. with something that does not exist actually_present = swh_storage.content_find( {'sha256': missing_cont['sha256']}) assert actually_present == [] def test_content_find_with_duplicate_input(self, swh_storage): cont1 = data.cont duplicate_cont = cont1.copy() # Create fake data with colliding sha256 and blake2s256 sha1_array = bytearray(duplicate_cont['sha1']) sha1_array[0] += 1 duplicate_cont['sha1'] = bytes(sha1_array) sha1git_array = bytearray(duplicate_cont['sha1_git']) sha1git_array[0] += 1 duplicate_cont['sha1_git'] = bytes(sha1git_array) # Inject the data swh_storage.content_add([cont1, duplicate_cont]) finder = {'blake2s256': duplicate_cont['blake2s256'], 'sha256': duplicate_cont['sha256']} actual_result = list(swh_storage.content_find(finder)) cont1.pop('data') duplicate_cont.pop('data') actual_result[0].pop('ctime') actual_result[1].pop('ctime') expected_result = [ cont1, duplicate_cont ] for result in expected_result: assert result in actual_result def test_content_find_with_duplicate_sha256(self, swh_storage): cont1 = data.cont duplicate_cont = cont1.copy() # Create fake data with colliding sha256 for hashalgo in ('sha1', 'sha1_git', 'blake2s256'): value = bytearray(duplicate_cont[hashalgo]) value[0] += 1 duplicate_cont[hashalgo] = bytes(value) swh_storage.content_add([cont1, duplicate_cont]) finder = { 'sha256': duplicate_cont['sha256'] } actual_result = list(swh_storage.content_find(finder)) assert len(actual_result) == 2 cont1.pop('data') duplicate_cont.pop('data') actual_result[0].pop('ctime') actual_result[1].pop('ctime') expected_result = [ cont1, duplicate_cont ] assert expected_result == sorted(actual_result, key=lambda x: x['sha1']) # Find with both sha256 and blake2s256 finder = { 'sha256': duplicate_cont['sha256'], 'blake2s256': duplicate_cont['blake2s256'] } actual_result = list(swh_storage.content_find(finder)) assert len(actual_result) == 1 actual_result[0].pop('ctime') expected_result = [duplicate_cont] assert actual_result[0] == duplicate_cont def test_content_find_with_duplicate_blake2s256(self, swh_storage): cont1 = data.cont duplicate_cont = cont1.copy() # Create fake data with colliding sha256 and blake2s256 sha1_array = bytearray(duplicate_cont['sha1']) sha1_array[0] += 1 duplicate_cont['sha1'] = bytes(sha1_array) sha1git_array = bytearray(duplicate_cont['sha1_git']) sha1git_array[0] += 1 duplicate_cont['sha1_git'] = bytes(sha1git_array) sha256_array = bytearray(duplicate_cont['sha256']) sha256_array[0] += 1 duplicate_cont['sha256'] = bytes(sha256_array) swh_storage.content_add([cont1, duplicate_cont]) finder = { 'blake2s256': duplicate_cont['blake2s256'] } actual_result = list(swh_storage.content_find(finder)) cont1.pop('data') duplicate_cont.pop('data') actual_result[0].pop('ctime') actual_result[1].pop('ctime') expected_result = [ cont1, duplicate_cont ] for result in expected_result: assert result in actual_result # Find with both sha256 and blake2s256 finder = { 'sha256': duplicate_cont['sha256'], 'blake2s256': duplicate_cont['blake2s256'] } actual_result = list(swh_storage.content_find(finder)) actual_result[0].pop('ctime') expected_result = [ duplicate_cont ] assert expected_result == actual_result def test_content_find_bad_input(self, swh_storage): # 1. with bad input with pytest.raises(ValueError): swh_storage.content_find({}) # empty is bad # 2. with bad input with pytest.raises(ValueError): swh_storage.content_find( {'unknown-sha1': 'something'}) # not the right key def test_object_find_by_sha1_git(self, swh_storage): sha1_gits = [b'00000000000000000000'] expected = { b'00000000000000000000': [], } swh_storage.content_add([data.cont]) sha1_gits.append(data.cont['sha1_git']) expected[data.cont['sha1_git']] = [{ 'sha1_git': data.cont['sha1_git'], 'type': 'content', }] swh_storage.directory_add([data.dir]) sha1_gits.append(data.dir['id']) expected[data.dir['id']] = [{ 'sha1_git': data.dir['id'], 'type': 'directory', }] swh_storage.revision_add([data.revision]) sha1_gits.append(data.revision['id']) expected[data.revision['id']] = [{ 'sha1_git': data.revision['id'], 'type': 'revision', }] swh_storage.release_add([data.release]) sha1_gits.append(data.release['id']) expected[data.release['id']] = [{ 'sha1_git': data.release['id'], 'type': 'release', }] ret = swh_storage.object_find_by_sha1_git(sha1_gits) assert expected == ret def test_tool_add(self, swh_storage): tool = { 'name': 'some-unknown-tool', 'version': 'some-version', 'configuration': {"debian-package": "some-package"}, } actual_tool = swh_storage.tool_get(tool) assert actual_tool is None # does not exist # add it actual_tools = swh_storage.tool_add([tool]) assert len(actual_tools) == 1 actual_tool = actual_tools[0] assert actual_tool is not None # now it exists new_id = actual_tool.pop('id') assert actual_tool == tool actual_tools2 = swh_storage.tool_add([tool]) actual_tool2 = actual_tools2[0] assert actual_tool2 is not None # now it exists new_id2 = actual_tool2.pop('id') assert new_id == new_id2 assert actual_tool == actual_tool2 def test_tool_add_multiple(self, swh_storage): tool = { 'name': 'some-unknown-tool', 'version': 'some-version', 'configuration': {"debian-package": "some-package"}, } actual_tools = list(swh_storage.tool_add([tool])) assert len(actual_tools) == 1 new_tools = [tool, { 'name': 'yet-another-tool', 'version': 'version', 'configuration': {}, }] actual_tools = swh_storage.tool_add(new_tools) assert len(actual_tools) == 2 # order not guaranteed, so we iterate over results to check for tool in actual_tools: _id = tool.pop('id') assert _id is not None assert tool in new_tools def test_tool_get_missing(self, swh_storage): tool = { 'name': 'unknown-tool', 'version': '3.1.0rc2-31-ga2cbb8c', 'configuration': {"command_line": "nomossa "}, } actual_tool = swh_storage.tool_get(tool) assert actual_tool is None def test_tool_metadata_get_missing_context(self, swh_storage): tool = { 'name': 'swh-metadata-translator', 'version': '0.0.1', 'configuration': {"context": "unknown-context"}, } actual_tool = swh_storage.tool_get(tool) assert actual_tool is None def test_tool_metadata_get(self, swh_storage): tool = { 'name': 'swh-metadata-translator', 'version': '0.0.1', 'configuration': {"type": "local", "context": "npm"}, } expected_tool = swh_storage.tool_add([tool])[0] # when actual_tool = swh_storage.tool_get(tool) # then assert expected_tool == actual_tool def test_metadata_provider_get(self, swh_storage): # given no_provider = swh_storage.metadata_provider_get(6459456445615) assert no_provider is None # when provider_id = swh_storage.metadata_provider_add( data.provider['name'], data.provider['type'], data.provider['url'], data.provider['metadata']) actual_provider = swh_storage.metadata_provider_get(provider_id) expected_provider = { 'provider_name': data.provider['name'], 'provider_url': data.provider['url'] } # then del actual_provider['id'] assert actual_provider, expected_provider def test_metadata_provider_get_by(self, swh_storage): # given no_provider = swh_storage.metadata_provider_get_by({ 'provider_name': data.provider['name'], 'provider_url': data.provider['url'] }) assert no_provider is None # when provider_id = swh_storage.metadata_provider_add( data.provider['name'], data.provider['type'], data.provider['url'], data.provider['metadata']) actual_provider = swh_storage.metadata_provider_get_by({ 'provider_name': data.provider['name'], 'provider_url': data.provider['url'] }) # then assert provider_id, actual_provider['id'] def test_origin_metadata_add(self, swh_storage): # given origin = data.origin swh_storage.origin_add([origin])[0] tools = swh_storage.tool_add([data.metadata_tool]) tool = tools[0] swh_storage.metadata_provider_add( data.provider['name'], data.provider['type'], data.provider['url'], data.provider['metadata']) provider = swh_storage.metadata_provider_get_by({ 'provider_name': data.provider['name'], 'provider_url': data.provider['url'] }) # when adding for the same origin 2 metadatas n_om = len(list(swh_storage.origin_metadata_get_by(origin['url']))) swh_storage.origin_metadata_add( origin['url'], data.origin_metadata['discovery_date'], provider['id'], tool['id'], data.origin_metadata['metadata']) swh_storage.origin_metadata_add( origin['url'], '2015-01-01 23:00:00+00', provider['id'], tool['id'], data.origin_metadata2['metadata']) n_actual_om = len(list( swh_storage.origin_metadata_get_by(origin['url']))) # then assert n_actual_om == n_om + 2 def test_origin_metadata_get(self, swh_storage): # given origin_url = data.origin['url'] origin_url2 = data.origin2['url'] swh_storage.origin_add([data.origin]) swh_storage.origin_add([data.origin2]) swh_storage.metadata_provider_add(data.provider['name'], data.provider['type'], data.provider['url'], data.provider['metadata']) provider = swh_storage.metadata_provider_get_by({ 'provider_name': data.provider['name'], 'provider_url': data.provider['url'] }) tool = swh_storage.tool_add([data.metadata_tool])[0] # when adding for the same origin 2 metadatas swh_storage.origin_metadata_add( origin_url, data.origin_metadata['discovery_date'], provider['id'], tool['id'], data.origin_metadata['metadata']) swh_storage.origin_metadata_add( origin_url2, data.origin_metadata2['discovery_date'], provider['id'], tool['id'], data.origin_metadata2['metadata']) swh_storage.origin_metadata_add( origin_url, data.origin_metadata2['discovery_date'], provider['id'], tool['id'], data.origin_metadata2['metadata']) all_metadatas = list(sorted(swh_storage.origin_metadata_get_by( origin_url), key=lambda x: x['discovery_date'])) metadatas_for_origin2 = list(swh_storage.origin_metadata_get_by( origin_url2)) expected_results = [{ 'origin_url': origin_url, 'discovery_date': datetime.datetime( 2015, 1, 1, 23, 0, tzinfo=datetime.timezone.utc), 'metadata': { 'name': 'test_origin_metadata', 'version': '0.0.1' }, 'provider_id': provider['id'], 'provider_name': 'hal', 'provider_type': 'deposit-client', 'provider_url': 'http:///hal/inria', 'tool_id': tool['id'] }, { 'origin_url': origin_url, 'discovery_date': datetime.datetime( 2017, 1, 1, 23, 0, tzinfo=datetime.timezone.utc), 'metadata': { 'name': 'test_origin_metadata', 'version': '0.0.1' }, 'provider_id': provider['id'], 'provider_name': 'hal', 'provider_type': 'deposit-client', 'provider_url': 'http:///hal/inria', 'tool_id': tool['id'] }] # then assert len(all_metadatas) == 2 assert len(metadatas_for_origin2) == 1 assert all_metadatas == expected_results def test_metadata_provider_add(self, swh_storage): provider = { 'provider_name': 'swMATH', 'provider_type': 'registry', 'provider_url': 'http://www.swmath.org/', 'metadata': { 'email': 'contact@swmath.org', 'license': 'All rights reserved' } } provider['id'] = provider_id = swh_storage.metadata_provider_add( **provider) assert provider == swh_storage.metadata_provider_get_by( {'provider_name': 'swMATH', 'provider_url': 'http://www.swmath.org/'}) assert provider == swh_storage.metadata_provider_get(provider_id) def test_origin_metadata_get_by_provider_type(self, swh_storage): # given origin_url = data.origin['url'] origin_url2 = data.origin2['url'] swh_storage.origin_add([data.origin]) swh_storage.origin_add([data.origin2]) provider1_id = swh_storage.metadata_provider_add( data.provider['name'], data.provider['type'], data.provider['url'], data.provider['metadata']) provider1 = swh_storage.metadata_provider_get_by({ 'provider_name': data.provider['name'], 'provider_url': data.provider['url'] }) assert provider1 == swh_storage.metadata_provider_get(provider1_id) provider2_id = swh_storage.metadata_provider_add( 'swMATH', 'registry', 'http://www.swmath.org/', {'email': 'contact@swmath.org', 'license': 'All rights reserved'}) provider2 = swh_storage.metadata_provider_get_by({ 'provider_name': 'swMATH', 'provider_url': 'http://www.swmath.org/' }) assert provider2 == swh_storage.metadata_provider_get(provider2_id) # using the only tool now inserted in the data.sql, but for this # provider should be a crawler tool (not yet implemented) tool = swh_storage.tool_add([data.metadata_tool])[0] # when adding for the same origin 2 metadatas swh_storage.origin_metadata_add( origin_url, data.origin_metadata['discovery_date'], provider1['id'], tool['id'], data.origin_metadata['metadata']) swh_storage.origin_metadata_add( origin_url2, data.origin_metadata2['discovery_date'], provider2['id'], tool['id'], data.origin_metadata2['metadata']) provider_type = 'registry' m_by_provider = list(swh_storage.origin_metadata_get_by( origin_url2, provider_type)) for item in m_by_provider: if 'id' in item: del item['id'] expected_results = [{ 'origin_url': origin_url2, 'discovery_date': datetime.datetime( 2017, 1, 1, 23, 0, tzinfo=datetime.timezone.utc), 'metadata': { 'name': 'test_origin_metadata', 'version': '0.0.1' }, 'provider_id': provider2['id'], 'provider_name': 'swMATH', 'provider_type': provider_type, 'provider_url': 'http://www.swmath.org/', 'tool_id': tool['id'] }] # then assert len(m_by_provider) == 1 assert m_by_provider == expected_results class TestStorageGeneratedData: def test_generate_content_get(self, swh_storage, swh_contents): contents_with_data = [c for c in swh_contents if c['status'] != 'absent'] # input the list of sha1s we want from storage get_sha1s = [c['sha1'] for c in contents_with_data] # retrieve contents actual_contents = list(swh_storage.content_get(get_sha1s)) assert None not in actual_contents assert_contents_ok(contents_with_data, actual_contents) def test_generate_content_get_metadata(self, swh_storage, swh_contents): # input the list of sha1s we want from storage expected_contents = [c for c in swh_contents if c['status'] != 'absent'] get_sha1s = [c['sha1'] for c in expected_contents] # retrieve contents meta_contents = swh_storage.content_get_metadata(get_sha1s) assert len(list(meta_contents)) == len(get_sha1s) actual_contents = [] for contents in meta_contents.values(): actual_contents.extend(contents) keys_to_check = {'length', 'status', 'sha1', 'sha1_git', 'sha256', 'blake2s256'} assert_contents_ok(expected_contents, actual_contents, keys_to_check=keys_to_check) def test_generate_content_get_range(self, swh_storage, swh_contents): """content_get_range returns complete range""" present_contents = [c for c in swh_contents if c['status'] != 'absent'] get_sha1s = sorted([c['sha1'] for c in swh_contents if c['status'] != 'absent']) start = get_sha1s[2] end = get_sha1s[-2] actual_result = swh_storage.content_get_range(start, end) assert actual_result['next'] is None actual_contents = actual_result['contents'] expected_contents = [c for c in present_contents if start <= c['sha1'] <= end] if expected_contents: assert_contents_ok( expected_contents, actual_contents, ['sha1']) else: assert actual_contents == [] def test_generate_content_get_range_full(self, swh_storage, swh_contents): """content_get_range for a full range returns all available contents""" present_contents = [c for c in swh_contents if c['status'] != 'absent'] start = b'0' * 40 end = b'f' * 40 actual_result = swh_storage.content_get_range(start, end) assert actual_result['next'] is None actual_contents = actual_result['contents'] expected_contents = [c for c in present_contents if start <= c['sha1'] <= end] if expected_contents: assert_contents_ok( expected_contents, actual_contents, ['sha1']) else: assert actual_contents == [] def test_generate_content_get_range_empty(self, swh_storage, swh_contents): """content_get_range for an empty range returns nothing""" start = b'0' * 40 end = b'f' * 40 actual_result = swh_storage.content_get_range(end, start) assert actual_result['next'] is None assert len(actual_result['contents']) == 0 def test_generate_content_get_range_limit_none(self, swh_storage): """content_get_range call with wrong limit input should fail""" with pytest.raises(ValueError) as e: swh_storage.content_get_range(start=None, end=None, limit=None) assert e.value.args == ('Development error: limit should not be None',) def test_generate_content_get_range_no_limit( self, swh_storage, swh_contents): """content_get_range returns contents within range provided""" # input the list of sha1s we want from storage get_sha1s = sorted([c['sha1'] for c in swh_contents if c['status'] != 'absent']) start = get_sha1s[0] end = get_sha1s[-1] # retrieve contents actual_result = swh_storage.content_get_range(start, end) actual_contents = actual_result['contents'] assert actual_result['next'] is None assert len(actual_contents) == len(get_sha1s) expected_contents = [c for c in swh_contents if c['status'] != 'absent'] assert_contents_ok( expected_contents, actual_contents, ['sha1']) def test_generate_content_get_range_limit(self, swh_storage, swh_contents): """content_get_range paginates results if limit exceeded""" contents_map = {c['sha1']: c for c in swh_contents} # input the list of sha1s we want from storage get_sha1s = sorted([c['sha1'] for c in swh_contents if c['status'] != 'absent']) start = get_sha1s[0] end = get_sha1s[-1] # retrieve contents limited to n-1 results limited_results = len(get_sha1s) - 1 actual_result = swh_storage.content_get_range( start, end, limit=limited_results) actual_contents = actual_result['contents'] assert actual_result['next'] == get_sha1s[-1] assert len(actual_contents) == limited_results expected_contents = [contents_map[sha1] for sha1 in get_sha1s[:-1]] assert_contents_ok( expected_contents, actual_contents, ['sha1']) # retrieve next part actual_results2 = swh_storage.content_get_range(start=end, end=end) assert actual_results2['next'] is None actual_contents2 = actual_results2['contents'] assert len(actual_contents2) == 1 assert_contents_ok( [contents_map[get_sha1s[-1]]], actual_contents2, ['sha1']) def test_origin_get_range_from_zero(self, swh_storage, swh_origins): actual_origins = list( swh_storage.origin_get_range(origin_from=0, origin_count=0)) assert len(actual_origins) == 0 actual_origins = list( swh_storage.origin_get_range(origin_from=0, origin_count=1)) assert len(actual_origins) == 1 assert actual_origins[0]['id'] == 1 assert actual_origins[0]['url'] == swh_origins[0]['url'] @pytest.mark.parametrize('origin_from,origin_count', [ (1, 1), (1, 10), (1, 20), (1, 101), (11, 0), (11, 10), (91, 11)]) def test_origin_get_range( self, swh_storage, swh_origins, origin_from, origin_count): actual_origins = list( swh_storage.origin_get_range(origin_from=origin_from, origin_count=origin_count)) origins_with_id = list(enumerate(swh_origins, start=1)) expected_origins = [ { 'url': origin['url'], 'id': origin_id, } for (origin_id, origin) in origins_with_id[origin_from-1:origin_from+origin_count-1] ] assert actual_origins == expected_origins @pytest.mark.parametrize('limit', [1, 7, 10, 100, 1000]) def test_origin_list(self, swh_storage, swh_origins, limit): returned_origins = [] page_token = None i = 0 while True: result = swh_storage.origin_list( page_token=page_token, limit=limit) assert len(result['origins']) <= limit returned_origins.extend( origin['url'] for origin in result['origins']) i += 1 page_token = result.get('next_page_token') if page_token is None: assert i*limit >= len(swh_origins) break else: assert len(result['origins']) == limit expected_origins = [origin['url'] for origin in swh_origins] assert sorted(returned_origins) == sorted(expected_origins) ORIGINS = [ 'https://github.com/user1/repo1', 'https://github.com/user2/repo1', 'https://github.com/user3/repo1', 'https://gitlab.com/user1/repo1', 'https://gitlab.com/user2/repo1', 'https://forge.softwareheritage.org/source/repo1', ] def test_origin_count(self, swh_storage): swh_storage.origin_add([{'url': url} for url in self.ORIGINS]) assert swh_storage.origin_count('github') == 3 assert swh_storage.origin_count('gitlab') == 2 assert swh_storage.origin_count('.*user.*', regexp=True) == 5 assert swh_storage.origin_count('.*user.*', regexp=False) == 0 assert swh_storage.origin_count('.*user1.*', regexp=True) == 2 assert swh_storage.origin_count('.*user1.*', regexp=False) == 0 def test_origin_count_with_visit_no_visits(self, swh_storage): swh_storage.origin_add([{'url': url} for url in self.ORIGINS]) # none of them have visits, so with_visit=True => 0 assert swh_storage.origin_count('github', with_visit=True) == 0 assert swh_storage.origin_count('gitlab', with_visit=True) == 0 assert swh_storage.origin_count('.*user.*', regexp=True, with_visit=True) == 0 assert swh_storage.origin_count('.*user.*', regexp=False, with_visit=True) == 0 assert swh_storage.origin_count('.*user1.*', regexp=True, with_visit=True) == 0 assert swh_storage.origin_count('.*user1.*', regexp=False, with_visit=True) == 0 def test_origin_count_with_visit_with_visits_no_snapshot( self, swh_storage): swh_storage.origin_add([{'url': url} for url in self.ORIGINS]) now = datetime.datetime.now(tz=datetime.timezone.utc) swh_storage.origin_visit_add( origin='https://github.com/user1/repo1', date=now, type='git') assert swh_storage.origin_count('github', with_visit=False) == 3 # it has a visit, but no snapshot, so with_visit=True => 0 assert swh_storage.origin_count('github', with_visit=True) == 0 assert swh_storage.origin_count('gitlab', with_visit=False) == 2 # these gitlab origins have no visit assert swh_storage.origin_count('gitlab', with_visit=True) == 0 assert swh_storage.origin_count('github.*user1', regexp=True, with_visit=False) == 1 assert swh_storage.origin_count('github.*user1', regexp=True, with_visit=True) == 0 assert swh_storage.origin_count('github', regexp=True, with_visit=True) == 0 def test_origin_count_with_visit_with_visits_and_snapshot( self, swh_storage): swh_storage.origin_add([{'url': url} for url in self.ORIGINS]) now = datetime.datetime.now(tz=datetime.timezone.utc) swh_storage.snapshot_add([data.snapshot]) visit = swh_storage.origin_visit_add( origin='https://github.com/user1/repo1', date=now, type='git') swh_storage.origin_visit_update( origin='https://github.com/user1/repo1', visit_id=visit['visit'], snapshot=data.snapshot['id']) assert swh_storage.origin_count('github', with_visit=False) == 3 # github/user1 has a visit and a snapshot, so with_visit=True => 1 assert swh_storage.origin_count('github', with_visit=True) == 1 assert swh_storage.origin_count('github.*user1', regexp=True, with_visit=False) == 1 assert swh_storage.origin_count('github.*user1', regexp=True, with_visit=True) == 1 assert swh_storage.origin_count('github', regexp=True, with_visit=True) == 1 @settings(suppress_health_check=[HealthCheck.too_slow]) @given(strategies.lists(objects(), max_size=2)) def test_add_arbitrary(self, swh_storage, objects): for (obj_type, obj) in objects: obj = obj.to_dict() if obj_type == 'origin_visit': origin = obj.pop('origin') swh_storage.origin_add_one({'url': origin}) if 'visit' in obj: del obj['visit'] swh_storage.origin_visit_add( origin, obj['date'], obj['type']) else: + if obj_type == 'content' and obj['status'] == 'absent': + obj_type = 'skipped_content' method = getattr(swh_storage, obj_type + '_add') try: method([obj]) except HashCollision: pass @pytest.mark.db class TestLocalStorage: """Test the local storage""" # This test is only relevant on the local storage, with an actual # objstorage raising an exception def test_content_add_objstorage_exception(self, swh_storage): swh_storage.objstorage.add = Mock( side_effect=Exception('mocked broken objstorage') ) with pytest.raises(Exception) as e: swh_storage.content_add([data.cont]) assert e.value.args == ('mocked broken objstorage',) missing = list(swh_storage.content_missing([data.cont])) assert missing == [data.cont['sha1']] @pytest.mark.db class TestStorageRaceConditions: @pytest.mark.xfail def test_content_add_race(self, swh_storage): results = queue.Queue() def thread(): try: with db_transaction(swh_storage) as (db, cur): ret = swh_storage.content_add([data.cont], db=db, cur=cur) results.put((threading.get_ident(), 'data', ret)) except Exception as e: results.put((threading.get_ident(), 'exc', e)) t1 = threading.Thread(target=thread) t2 = threading.Thread(target=thread) t1.start() # this avoids the race condition # import time # time.sleep(1) t2.start() t1.join() t2.join() r1 = results.get(block=False) r2 = results.get(block=False) with pytest.raises(queue.Empty): results.get(block=False) assert r1[0] != r2[0] assert r1[1] == 'data', 'Got exception %r in Thread%s' % (r1[2], r1[0]) assert r2[1] == 'data', 'Got exception %r in Thread%s' % (r2[2], r2[0]) @pytest.mark.db class TestPgStorage: """This class is dedicated for the rare case where the schema needs to be altered dynamically. Otherwise, the tests could be blocking when ran altogether. """ def test_content_update_with_new_cols(self, swh_storage): swh_storage.journal_writer = None # TODO, not supported with db_transaction(swh_storage) as (_, cur): cur.execute("""alter table content add column test text default null, add column test2 text default null""") cont = copy.deepcopy(data.cont2) swh_storage.content_add([cont]) cont['test'] = 'value-1' cont['test2'] = 'value-2' swh_storage.content_update([cont], keys=['test', 'test2']) with db_transaction(swh_storage) as (_, cur): cur.execute( '''SELECT sha1, sha1_git, sha256, length, status, test, test2 FROM content WHERE sha1 = %s''', (cont['sha1'],)) datum = cur.fetchone() assert datum == (cont['sha1'], cont['sha1_git'], cont['sha256'], cont['length'], 'visible', cont['test'], cont['test2']) with db_transaction(swh_storage) as (_, cur): cur.execute("""alter table content drop column test, drop column test2""") def test_content_add_db(self, swh_storage): cont = data.cont actual_result = swh_storage.content_add([cont]) assert actual_result == { 'content:add': 1, 'content:add:bytes': cont['length'], - 'skipped_content:add': 0 } if hasattr(swh_storage, 'objstorage'): assert cont['sha1'] in swh_storage.objstorage with db_transaction(swh_storage) as (_, cur): cur.execute('SELECT sha1, sha1_git, sha256, length, status' ' FROM content WHERE sha1 = %s', (cont['sha1'],)) datum = cur.fetchone() assert datum == (cont['sha1'], cont['sha1_git'], cont['sha256'], cont['length'], 'visible') expected_cont = cont.copy() del expected_cont['data'] journal_objects = list(swh_storage.journal_writer.objects) for (obj_type, obj) in journal_objects: del obj['ctime'] assert journal_objects == [('content', expected_cont)] def test_content_add_metadata_db(self, swh_storage): cont = data.cont del cont['data'] cont['ctime'] = datetime.datetime.now() actual_result = swh_storage.content_add_metadata([cont]) assert actual_result == { 'content:add': 1, - 'skipped_content:add': 0 } if hasattr(swh_storage, 'objstorage'): assert cont['sha1'] not in swh_storage.objstorage with db_transaction(swh_storage) as (_, cur): cur.execute('SELECT sha1, sha1_git, sha256, length, status' ' FROM content WHERE sha1 = %s', (cont['sha1'],)) datum = cur.fetchone() assert datum == (cont['sha1'], cont['sha1_git'], cont['sha256'], cont['length'], 'visible') assert list(swh_storage.journal_writer.objects) == [('content', cont)] def test_skipped_content_add_db(self, swh_storage): cont = data.skipped_cont cont2 = data.skipped_cont2 cont2['blake2s256'] = None - actual_result = swh_storage.content_add([cont, cont, cont2]) + actual_result = swh_storage.skipped_content_add([cont, cont, cont2]) - assert actual_result == { - 'content:add': 0, - 'content:add:bytes': 0, - 'skipped_content:add': 2, - } + assert 2 <= actual_result.pop('skipped_content:add') <= 3 + assert actual_result == {} with db_transaction(swh_storage) as (_, cur): cur.execute('SELECT sha1, sha1_git, sha256, blake2s256, ' 'length, status, reason ' 'FROM skipped_content ORDER BY sha1_git') dbdata = cur.fetchall() assert len(dbdata) == 2 assert dbdata[0] == (cont['sha1'], cont['sha1_git'], cont['sha256'], cont['blake2s256'], cont['length'], 'absent', 'Content too long') assert dbdata[1] == (cont2['sha1'], cont2['sha1_git'], cont2['sha256'], cont2['blake2s256'], cont2['length'], 'absent', 'Content too long')