diff --git a/swh/indexer/ctags.py b/swh/indexer/ctags.py index bae65f2..df78e59 100644 --- a/swh/indexer/ctags.py +++ b/swh/indexer/ctags.py @@ -1,147 +1,151 @@ # 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 json import subprocess -from typing import Any, Dict, List +from typing import Any, Dict, Iterator, List, Optional, Union from swh.core.config import merge_configs +from swh.indexer.storage.model import ContentCtagsRow from swh.model import hashutil +from swh.model.model import Revision from .indexer import ContentIndexer, write_to_temp # Options used to compute tags __FLAGS = [ "--fields=+lnz", # +l: language # +n: line number of tag definition # +z: include the symbol's kind (function, variable, ...) "--sort=no", # sort output on tag name "--links=no", # do not follow symlinks "--output-format=json", # outputs in json ] def compute_language(content, log=None): raise NotImplementedError( "Language detection was unreliable, so it is currently disabled. " "See https://forge.softwareheritage.org/D1455" ) -def run_ctags(path, lang=None, ctags_command="ctags"): +def run_ctags(path, lang=None, ctags_command="ctags") -> Iterator[Dict[str, Any]]: """Run ctags on file path with optional language. Args: path: path to the file lang: language for that path (optional) Yields: dict: ctags' output """ optional = [] if lang: optional = ["--language-force=%s" % lang] cmd = [ctags_command] + __FLAGS + optional + [path] output = subprocess.check_output(cmd, universal_newlines=True) for symbol in output.split("\n"): if not symbol: continue js_symbol = json.loads(symbol) yield { "name": js_symbol["name"], "kind": js_symbol["kind"], "line": js_symbol["line"], "lang": js_symbol["language"], } DEFAULT_CONFIG: Dict[str, Any] = { "workdir": "/tmp/swh/indexer.ctags", "tools": { "name": "universal-ctags", "version": "~git7859817b", "configuration": { "command_line": """ctags --fields=+lnz --sort=no --links=no """ """--output-format=json """ }, }, "languages": {}, } -class CtagsIndexer(ContentIndexer[Dict]): +class CtagsIndexer(ContentIndexer[ContentCtagsRow]): def __init__(self, *args, **kwargs): super().__init__(*args, **kwargs) self.config = merge_configs(DEFAULT_CONFIG, self.config) self.working_directory = self.config["workdir"] self.language_map = self.config["languages"] def filter(self, ids): """Filter out known sha1s and return only missing ones. """ yield from self.idx_storage.content_ctags_missing( ({"id": sha1, "indexer_configuration_id": self.tool["id"],} for sha1 in ids) ) - def index(self, id, data): + def index( + self, id: Union[bytes, Dict, Revision], data: Optional[bytes] = None, **kwargs + ) -> List[ContentCtagsRow]: """Index sha1s' content and store result. Args: id (bytes): content's identifier data (bytes): raw content in bytes Returns: dict: a dict representing a content_mimetype with keys: - **id** (bytes): content's identifier (sha1) - **ctags** ([dict]): ctags list of symbols """ + assert isinstance(id, bytes) + assert data is not None + lang = compute_language(data, log=self.log)["lang"] if not lang: - return None + return [] ctags_lang = self.language_map.get(lang) if not ctags_lang: - return None + return [] - ctags = { - "id": id, - } + ctags = [] filename = hashutil.hash_to_hex(id) with write_to_temp( filename=filename, data=data, working_directory=self.working_directory ) as content_path: - result = run_ctags(content_path, lang=ctags_lang) - ctags.update( - {"ctags": list(result), "indexer_configuration_id": self.tool["id"],} - ) + for ctag_kwargs in run_ctags(content_path, lang=ctags_lang): + ctags.append( + ContentCtagsRow( + id=id, indexer_configuration_id=self.tool["id"], **ctag_kwargs, + ) + ) - return [ctags] + return ctags def persist_index_computations( - self, results: List[Dict], policy_update: str + self, results: List[ContentCtagsRow], policy_update: str ) -> Dict[str, int]: """Persist the results in storage. Args: - results: list of content_mimetype, dict with the - following keys: - - id (bytes): content's identifier (sha1) - - ctags ([dict]): ctags list of symbols + results: list of ctags returned by index() policy_update: either 'update-dups' or 'ignore-dups' to respectively update duplicates or ignore them """ return self.idx_storage.content_ctags_add( results, conflict_update=(policy_update == "update-dups") ) diff --git a/swh/indexer/storage/__init__.py b/swh/indexer/storage/__init__.py index f474c3e..d1aaa16 100644 --- a/swh/indexer/storage/__init__.py +++ b/swh/indexer/storage/__init__.py @@ -1,669 +1,684 @@ # 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 collections import Counter -import itertools import json from typing import Dict, Iterable, List, Optional, Tuple import psycopg2 import psycopg2.pool from swh.core.db.common import db_transaction from swh.model.hashutil import hash_to_bytes, hash_to_hex from swh.model.model import SHA1_SIZE from swh.storage.exc import StorageDBError from swh.storage.utils import get_partition_bounds_bytes from . import converters from .db import Db from .exc import DuplicateId, IndexerStorageArgumentException from .interface import PagedResult, Sha1 from .metrics import process_metrics, send_metric, timed from .model import ( ContentCtagsRow, ContentLanguageRow, ContentLicenseRow, ContentMetadataRow, ContentMimetypeRow, OriginIntrinsicMetadataRow, RevisionIntrinsicMetadataRow, ) INDEXER_CFG_KEY = "indexer_storage" MAPPING_NAMES = ["codemeta", "gemspec", "maven", "npm", "pkg-info"] def get_indexer_storage(cls, args): """Get an indexer storage object of class `storage_class` with arguments `storage_args`. Args: cls (str): storage's class, either 'local' or 'remote' args (dict): dictionary of arguments passed to the storage class constructor Returns: an instance of swh.indexer's storage (either local or remote) Raises: ValueError if passed an unknown storage class. """ if cls == "remote": from .api.client import RemoteStorage as IndexerStorage elif cls == "local": from . import IndexerStorage elif cls == "memory": from .in_memory import IndexerStorage else: raise ValueError("Unknown indexer storage class `%s`" % cls) return IndexerStorage(**args) def check_id_duplicates(data): """ If any two row models in `data` have the same unique key, raises a `ValueError`. Values associated to the key must be hashable. Args: data (List[dict]): List of dictionaries to be inserted >>> check_id_duplicates([ ... ContentLanguageRow(id=b'foo', indexer_configuration_id=42, lang="python"), ... ContentLanguageRow(id=b'foo', indexer_configuration_id=32, lang="python"), ... ]) >>> check_id_duplicates([ ... ContentLanguageRow(id=b'foo', indexer_configuration_id=42, lang="python"), ... ContentLanguageRow(id=b'foo', indexer_configuration_id=42, lang="python"), ... ]) Traceback (most recent call last): ... swh.indexer.storage.exc.DuplicateId: [{'id': b'foo', 'indexer_configuration_id': 42}] """ # noqa counter = Counter(tuple(sorted(item.unique_key().items())) for item in data) duplicates = [id_ for (id_, count) in counter.items() if count >= 2] if duplicates: raise DuplicateId(list(map(dict, duplicates))) class IndexerStorage: """SWH Indexer Storage """ def __init__(self, db, min_pool_conns=1, max_pool_conns=10): """ Args: db_conn: either a libpq connection string, or a psycopg2 connection """ 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) def get_db(self): if self._db: return self._db return Db.from_pool(self._pool) def put_db(self, db): if db is not self._db: db.put_conn() @timed @db_transaction() def check_config(self, *, check_write, db=None, cur=None): # Check permissions on one of the tables if check_write: check = "INSERT" else: check = "SELECT" cur.execute( "select has_table_privilege(current_user, 'content_mimetype', %s)", # noqa (check,), ) return cur.fetchone()[0] @timed @db_transaction() def content_mimetype_missing( self, mimetypes: Iterable[Dict], db=None, cur=None ) -> List[Tuple[Sha1, int]]: return [obj[0] for obj in db.content_mimetype_missing_from_list(mimetypes, cur)] @timed @db_transaction() def get_partition( self, indexer_type: str, indexer_configuration_id: int, partition_id: int, nb_partitions: int, page_token: Optional[str] = None, limit: int = 1000, with_textual_data=False, db=None, cur=None, ) -> PagedResult[Sha1]: """Retrieve ids of content with `indexer_type` within within partition partition_id bound by limit. Args: **indexer_type**: Type of data content to index (mimetype, language, etc...) **indexer_configuration_id**: The tool used to index data **partition_id**: index of the partition to fetch **nb_partitions**: total number of partitions to split into **page_token**: opaque token used for pagination **limit**: Limit result (default to 1000) **with_textual_data** (bool): Deal with only textual content (True) or all content (all contents by defaults, False) Raises: IndexerStorageArgumentException for; - limit to None - wrong indexer_type provided Returns: PagedResult of Sha1. If next_page_token is None, there is no more data to fetch """ if limit is None: raise IndexerStorageArgumentException("limit should not be None") if indexer_type not in db.content_indexer_names: err = f"Wrong type. Should be one of [{','.join(db.content_indexer_names)}]" raise IndexerStorageArgumentException(err) start, end = get_partition_bounds_bytes(partition_id, nb_partitions, SHA1_SIZE) if page_token is not None: start = hash_to_bytes(page_token) if end is None: end = b"\xff" * SHA1_SIZE next_page_token: Optional[str] = None ids = [ row[0] for row in db.content_get_range( indexer_type, start, end, indexer_configuration_id, limit=limit + 1, with_textual_data=with_textual_data, cur=cur, ) ] if len(ids) >= limit: next_page_token = hash_to_hex(ids[-1]) ids = ids[:limit] assert len(ids) <= limit return PagedResult(results=ids, next_page_token=next_page_token) @timed @db_transaction() def content_mimetype_get_partition( self, indexer_configuration_id: int, partition_id: int, nb_partitions: int, page_token: Optional[str] = None, limit: int = 1000, db=None, cur=None, ) -> PagedResult[Sha1]: return self.get_partition( "mimetype", indexer_configuration_id, partition_id, nb_partitions, page_token=page_token, limit=limit, db=db, cur=cur, ) @timed @process_metrics @db_transaction() def content_mimetype_add( self, mimetypes: List[ContentMimetypeRow], conflict_update: bool = False, db=None, cur=None, ) -> Dict[str, int]: check_id_duplicates(mimetypes) mimetypes.sort(key=lambda m: m.id) db.mktemp_content_mimetype(cur) db.copy_to( [m.to_dict() for m in mimetypes], "tmp_content_mimetype", ["id", "mimetype", "encoding", "indexer_configuration_id"], cur, ) count = db.content_mimetype_add_from_temp(conflict_update, cur) return {"content_mimetype:add": count} @timed @db_transaction() def content_mimetype_get( self, ids: Iterable[Sha1], db=None, cur=None ) -> List[ContentMimetypeRow]: return [ ContentMimetypeRow.from_dict( converters.db_to_mimetype(dict(zip(db.content_mimetype_cols, c))) ) for c in db.content_mimetype_get_from_list(ids, cur) ] @timed @db_transaction() def content_language_missing( self, languages: Iterable[Dict], db=None, cur=None ) -> List[Tuple[Sha1, int]]: return [obj[0] for obj in db.content_language_missing_from_list(languages, cur)] @timed @db_transaction() def content_language_get( self, ids: Iterable[Sha1], db=None, cur=None ) -> List[ContentLanguageRow]: return [ ContentLanguageRow.from_dict( converters.db_to_language(dict(zip(db.content_language_cols, c))) ) for c in db.content_language_get_from_list(ids, cur) ] @timed @process_metrics @db_transaction() def content_language_add( self, languages: List[ContentLanguageRow], conflict_update: bool = False, db=None, cur=None, ) -> Dict[str, int]: check_id_duplicates(languages) languages.sort(key=lambda m: m.id) db.mktemp_content_language(cur) # empty language is mapped to 'unknown' db.copy_to( ( { "id": lang.id, "lang": lang.lang or "unknown", "indexer_configuration_id": lang.indexer_configuration_id, } for lang in languages ), "tmp_content_language", ["id", "lang", "indexer_configuration_id"], cur, ) count = db.content_language_add_from_temp(conflict_update, cur) return {"content_language:add": count} @timed @db_transaction() - def content_ctags_missing(self, ctags, db=None, cur=None): + def content_ctags_missing( + self, ctags: Iterable[Dict], db=None, cur=None + ) -> List[Tuple[Sha1, int]]: return [obj[0] for obj in db.content_ctags_missing_from_list(ctags, cur)] @timed @db_transaction() - def content_ctags_get(self, ids, db=None, cur=None): + def content_ctags_get( + self, ids: Iterable[Sha1], db=None, cur=None + ) -> List[ContentCtagsRow]: return [ - converters.db_to_ctags(dict(zip(db.content_ctags_cols, c))) + ContentCtagsRow.from_dict( + converters.db_to_ctags(dict(zip(db.content_ctags_cols, c))) + ) for c in db.content_ctags_get_from_list(ids, cur) ] @timed @process_metrics @db_transaction() def content_ctags_add( - self, ctags: List[Dict], conflict_update: bool = False, db=None, cur=None + self, + ctags: List[ContentCtagsRow], + conflict_update: bool = False, + db=None, + cur=None, ) -> Dict[str, int]: - rows = list(itertools.chain.from_iterable(map(converters.ctags_to_db, ctags))) - check_id_duplicates(map(ContentCtagsRow.from_dict, rows)) - ctags.sort(key=lambda m: m["id"]) + check_id_duplicates(ctags) + ctags.sort(key=lambda m: m.id) db.mktemp_content_ctags(cur) db.copy_to( - rows, + [ctag.to_dict() for ctag in ctags], tblname="tmp_content_ctags", columns=["id", "name", "kind", "line", "lang", "indexer_configuration_id"], cur=cur, ) count = db.content_ctags_add_from_temp(conflict_update, cur) return {"content_ctags:add": count} @timed @db_transaction() def content_ctags_search( - self, expression, limit=10, last_sha1=None, db=None, cur=None - ): + self, + expression: str, + limit: int = 10, + last_sha1: Optional[Sha1] = None, + db=None, + cur=None, + ) -> List[ContentCtagsRow]: return [ - converters.db_to_ctags(dict(zip(db.content_ctags_cols, obj))) + ContentCtagsRow.from_dict( + converters.db_to_ctags(dict(zip(db.content_ctags_cols, obj))) + ) for obj in db.content_ctags_search(expression, last_sha1, limit, cur=cur) ] @timed @db_transaction() def content_fossology_license_get( self, ids: Iterable[Sha1], db=None, cur=None ) -> List[ContentLicenseRow]: return [ ContentLicenseRow.from_dict( converters.db_to_fossology_license( dict(zip(db.content_fossology_license_cols, c)) ) ) for c in db.content_fossology_license_get_from_list(ids, cur) ] @timed @process_metrics @db_transaction() def content_fossology_license_add( self, licenses: List[ContentLicenseRow], conflict_update: bool = False, db=None, cur=None, ) -> Dict[str, int]: check_id_duplicates(licenses) licenses.sort(key=lambda m: m.id) db.mktemp_content_fossology_license(cur) db.copy_to( [license.to_dict() for license in licenses], tblname="tmp_content_fossology_license", columns=["id", "license", "indexer_configuration_id"], cur=cur, ) count = db.content_fossology_license_add_from_temp(conflict_update, cur) return {"content_fossology_license:add": count} @timed @db_transaction() def content_fossology_license_get_partition( self, indexer_configuration_id: int, partition_id: int, nb_partitions: int, page_token: Optional[str] = None, limit: int = 1000, db=None, cur=None, ) -> PagedResult[Sha1]: return self.get_partition( "fossology_license", indexer_configuration_id, partition_id, nb_partitions, page_token=page_token, limit=limit, with_textual_data=True, db=db, cur=cur, ) @timed @db_transaction() def content_metadata_missing(self, metadata, db=None, cur=None): return [obj[0] for obj in db.content_metadata_missing_from_list(metadata, cur)] @timed @db_transaction() def content_metadata_get(self, ids, db=None, cur=None): return [ converters.db_to_metadata(dict(zip(db.content_metadata_cols, c))) for c in db.content_metadata_get_from_list(ids, cur) ] @timed @process_metrics @db_transaction() def content_metadata_add( self, metadata: List[Dict], conflict_update: bool = False, db=None, cur=None ) -> Dict[str, int]: check_id_duplicates(map(ContentMetadataRow.from_dict, metadata)) metadata.sort(key=lambda m: m["id"]) db.mktemp_content_metadata(cur) db.copy_to( metadata, "tmp_content_metadata", ["id", "metadata", "indexer_configuration_id"], cur, ) count = db.content_metadata_add_from_temp(conflict_update, cur) return { "content_metadata:add": count, } @timed @db_transaction() def revision_intrinsic_metadata_missing(self, metadata, db=None, cur=None): return [ obj[0] for obj in db.revision_intrinsic_metadata_missing_from_list(metadata, cur) ] @timed @db_transaction() def revision_intrinsic_metadata_get(self, ids, db=None, cur=None): return [ converters.db_to_metadata(dict(zip(db.revision_intrinsic_metadata_cols, c))) for c in db.revision_intrinsic_metadata_get_from_list(ids, cur) ] @timed @process_metrics @db_transaction() def revision_intrinsic_metadata_add( self, metadata: List[Dict], conflict_update: bool = False, db=None, cur=None ) -> Dict[str, int]: check_id_duplicates(map(RevisionIntrinsicMetadataRow.from_dict, metadata)) metadata.sort(key=lambda m: m["id"]) db.mktemp_revision_intrinsic_metadata(cur) db.copy_to( metadata, "tmp_revision_intrinsic_metadata", ["id", "metadata", "mappings", "indexer_configuration_id"], cur, ) count = db.revision_intrinsic_metadata_add_from_temp(conflict_update, cur) return { "revision_intrinsic_metadata:add": count, } @timed @process_metrics @db_transaction() def revision_intrinsic_metadata_delete( self, entries: List[Dict], db=None, cur=None ) -> Dict: count = db.revision_intrinsic_metadata_delete(entries, cur) return {"revision_intrinsic_metadata:del": count} @timed @db_transaction() def origin_intrinsic_metadata_get(self, ids, db=None, cur=None): return [ converters.db_to_metadata(dict(zip(db.origin_intrinsic_metadata_cols, c))) for c in db.origin_intrinsic_metadata_get_from_list(ids, cur) ] @timed @process_metrics @db_transaction() def origin_intrinsic_metadata_add( self, metadata: List[Dict], conflict_update: bool = False, db=None, cur=None ) -> Dict[str, int]: check_id_duplicates(map(OriginIntrinsicMetadataRow.from_dict, metadata)) metadata.sort(key=lambda m: m["id"]) db.mktemp_origin_intrinsic_metadata(cur) db.copy_to( metadata, "tmp_origin_intrinsic_metadata", ["id", "metadata", "indexer_configuration_id", "from_revision", "mappings"], cur, ) count = db.origin_intrinsic_metadata_add_from_temp(conflict_update, cur) return { "origin_intrinsic_metadata:add": count, } @timed @process_metrics @db_transaction() def origin_intrinsic_metadata_delete( self, entries: List[Dict], db=None, cur=None ) -> Dict: count = db.origin_intrinsic_metadata_delete(entries, cur) return { "origin_intrinsic_metadata:del": count, } @timed @db_transaction() def origin_intrinsic_metadata_search_fulltext( self, conjunction, limit=100, db=None, cur=None ): return [ converters.db_to_metadata(dict(zip(db.origin_intrinsic_metadata_cols, c))) for c in db.origin_intrinsic_metadata_search_fulltext( conjunction, limit=limit, cur=cur ) ] @timed @db_transaction() def origin_intrinsic_metadata_search_by_producer( self, page_token="", limit=100, ids_only=False, mappings=None, tool_ids=None, db=None, cur=None, ): assert isinstance(page_token, str) # we go to limit+1 to check whether we should add next_page_token in # the response res = db.origin_intrinsic_metadata_search_by_producer( page_token, limit + 1, ids_only, mappings, tool_ids, cur ) result = {} if ids_only: result["origins"] = [origin for (origin,) in res] if len(result["origins"]) > limit: result["origins"][limit:] = [] result["next_page_token"] = result["origins"][-1] else: result["origins"] = [ converters.db_to_metadata( dict(zip(db.origin_intrinsic_metadata_cols, c)) ) for c in res ] if len(result["origins"]) > limit: result["origins"][limit:] = [] result["next_page_token"] = result["origins"][-1]["id"] return result @timed @db_transaction() def origin_intrinsic_metadata_stats(self, db=None, cur=None): mapping_names = [m for m in MAPPING_NAMES] select_parts = [] # Count rows for each mapping for mapping_name in mapping_names: select_parts.append( ( "sum(case when (mappings @> ARRAY['%s']) " " then 1 else 0 end)" ) % mapping_name ) # Total select_parts.append("sum(1)") # Rows whose metadata has at least one key that is not '@context' select_parts.append( "sum(case when ('{}'::jsonb @> (metadata - '@context')) " " then 0 else 1 end)" ) cur.execute( "select " + ", ".join(select_parts) + " from origin_intrinsic_metadata" ) results = dict(zip(mapping_names + ["total", "non_empty"], cur.fetchone())) return { "total": results.pop("total"), "non_empty": results.pop("non_empty"), "per_mapping": results, } @timed @db_transaction() def indexer_configuration_add(self, tools, db=None, cur=None): db.mktemp_indexer_configuration(cur) db.copy_to( tools, "tmp_indexer_configuration", ["tool_name", "tool_version", "tool_configuration"], cur, ) tools = db.indexer_configuration_add_from_temp(cur) results = [dict(zip(db.indexer_configuration_cols, line)) for line in tools] send_metric( "indexer_configuration:add", len(results), method_name="indexer_configuration_add", ) return results @timed @db_transaction() def indexer_configuration_get(self, tool, db=None, cur=None): tool_conf = tool["tool_configuration"] if isinstance(tool_conf, dict): tool_conf = json.dumps(tool_conf) idx = db.indexer_configuration_get( tool["tool_name"], tool["tool_version"], tool_conf ) if not idx: return None return dict(zip(db.indexer_configuration_cols, idx)) diff --git a/swh/indexer/storage/in_memory.py b/swh/indexer/storage/in_memory.py index 8932c77..9675a90 100644 --- a/swh/indexer/storage/in_memory.py +++ b/swh/indexer/storage/in_memory.py @@ -1,525 +1,510 @@ # Copyright (C) 2018-2020 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 collections import Counter, defaultdict import itertools import json import math import operator import re from typing import ( Any, Dict, Generic, Iterable, List, Optional, Set, Tuple, Type, TypeVar, ) from swh.core.collections import SortedList from swh.model.hashutil import hash_to_bytes, hash_to_hex from swh.model.model import SHA1_SIZE, Sha1Git from swh.storage.utils import get_partition_bounds_bytes -from . import MAPPING_NAMES, check_id_duplicates, converters +from . import MAPPING_NAMES, check_id_duplicates from .exc import IndexerStorageArgumentException from .interface import PagedResult, Sha1 from .model import ( BaseRow, ContentCtagsRow, ContentLanguageRow, ContentLicenseRow, ContentMetadataRow, ContentMimetypeRow, OriginIntrinsicMetadataRow, RevisionIntrinsicMetadataRow, ) SHA1_DIGEST_SIZE = 160 def _transform_tool(tool): return { "id": tool["id"], "name": tool["tool_name"], "version": tool["tool_version"], "configuration": tool["tool_configuration"], } def check_id_types(data: List[Dict[str, Any]]): """Checks all elements of the list have an 'id' whose type is 'bytes'.""" if not all(isinstance(item.get("id"), bytes) for item in data): raise IndexerStorageArgumentException("identifiers must be bytes.") def _key_from_dict(d): return tuple(sorted(d.items())) ToolId = int TValue = TypeVar("TValue", bound=BaseRow) class SubStorage(Generic[TValue]): """Implements common missing/get/add logic for each indexer type.""" _data: Dict[Sha1, Dict[Tuple, Dict[str, Any]]] _tools_per_id: Dict[Sha1, Set[ToolId]] def __init__(self, row_class: Type[TValue], tools): self.row_class = row_class self._tools = tools self._sorted_ids = SortedList[bytes, Sha1]() self._data = defaultdict(dict) self._tools_per_id = defaultdict(set) def _key_from_dict(self, d) -> Tuple: """Like the global _key_from_dict, but filters out dict keys that don't belong in the unique key.""" return _key_from_dict({k: d[k] for k in self.row_class.UNIQUE_KEY_FIELDS}) def missing(self, keys: Iterable[Dict]) -> List[Sha1]: """List data missing from storage. Args: data (iterable): dictionaries with keys: - **id** (bytes): sha1 identifier - **indexer_configuration_id** (int): tool used to compute the results Yields: missing sha1s """ results = [] for key in keys: tool_id = key["indexer_configuration_id"] id_ = key["id"] if tool_id not in self._tools_per_id.get(id_, set()): results.append(id_) return results def get(self, ids: Iterable[Sha1]) -> List[TValue]: """Retrieve data per id. Args: ids (iterable): sha1 checksums Yields: dict: dictionaries with the following keys: - **id** (bytes) - **tool** (dict): tool used to compute metadata - arbitrary data (as provided to `add`) """ results = [] for id_ in ids: for entry in self._data[id_].values(): entry = entry.copy() tool_id = entry.pop("indexer_configuration_id") results.append( self.row_class( id=id_, tool=_transform_tool(self._tools[tool_id]), **entry, ) ) return results def get_all(self) -> List[TValue]: return self.get(self._sorted_ids) def get_partition( self, indexer_configuration_id: int, partition_id: int, nb_partitions: int, page_token: Optional[str] = None, limit: int = 1000, ) -> PagedResult[Sha1]: """Retrieve ids of content with `indexer_type` within partition partition_id bound by limit. Args: **indexer_type**: Type of data content to index (mimetype, language, etc...) **indexer_configuration_id**: The tool used to index data **partition_id**: index of the partition to fetch **nb_partitions**: total number of partitions to split into **page_token**: opaque token used for pagination **limit**: Limit result (default to 1000) **with_textual_data** (bool): Deal with only textual content (True) or all content (all contents by defaults, False) Raises: IndexerStorageArgumentException for; - limit to None - wrong indexer_type provided Returns: PagedResult of Sha1. If next_page_token is None, there is no more data to fetch """ if limit is None: raise IndexerStorageArgumentException("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 next_page_token: Optional[str] = None ids: List[Sha1] = [] sha1s = (sha1 for sha1 in self._sorted_ids.iter_from(start)) for counter, sha1 in enumerate(sha1s): if sha1 > end: break if counter >= limit: next_page_token = hash_to_hex(sha1) break ids.append(sha1) assert len(ids) <= limit return PagedResult(results=ids, next_page_token=next_page_token) def add(self, data: Iterable[TValue], conflict_update: bool) -> int: """Add data not present in storage. Args: data (iterable): dictionaries with keys: - **id**: sha1 - **indexer_configuration_id**: tool used to compute the results - arbitrary data conflict_update (bool): Flag to determine if we want to overwrite (true) or skip duplicates (false) """ data = list(data) check_id_duplicates(data) count = 0 for obj in data: item = obj.to_dict() id_ = item.pop("id") tool_id = item["indexer_configuration_id"] key = _key_from_dict(obj.unique_key()) if not conflict_update and key in self._data[id_]: # Duplicate, should not be updated continue self._data[id_][key] = item self._tools_per_id[id_].add(tool_id) count += 1 if id_ not in self._sorted_ids: self._sorted_ids.add(id_) return count def delete(self, entries: List[Dict]) -> int: """Delete entries and return the number of entries deleted. """ deleted = 0 for entry in entries: (id_, tool_id) = (entry["id"], entry["indexer_configuration_id"]) if tool_id in self._tools_per_id[id_]: self._tools_per_id[id_].remove(tool_id) if id_ in self._data: key = self._key_from_dict(entry) if key in self._data[id_]: deleted += 1 del self._data[id_][key] return deleted class IndexerStorage: """In-memory SWH indexer storage.""" def __init__(self): self._tools = {} self._mimetypes = SubStorage(ContentMimetypeRow, self._tools) self._languages = SubStorage(ContentLanguageRow, self._tools) self._content_ctags = SubStorage(ContentCtagsRow, self._tools) self._licenses = SubStorage(ContentLicenseRow, self._tools) self._content_metadata = SubStorage(ContentMetadataRow, self._tools) self._revision_intrinsic_metadata = SubStorage( RevisionIntrinsicMetadataRow, self._tools ) self._origin_intrinsic_metadata = SubStorage( OriginIntrinsicMetadataRow, self._tools ) def check_config(self, *, check_write): return True def content_mimetype_missing( self, mimetypes: Iterable[Dict] ) -> List[Tuple[Sha1, int]]: return self._mimetypes.missing(mimetypes) def content_mimetype_get_partition( self, indexer_configuration_id: int, partition_id: int, nb_partitions: int, page_token: Optional[str] = None, limit: int = 1000, ) -> PagedResult[Sha1]: return self._mimetypes.get_partition( indexer_configuration_id, partition_id, nb_partitions, page_token, limit ) def content_mimetype_add( self, mimetypes: List[ContentMimetypeRow], conflict_update: bool = False ) -> Dict[str, int]: added = self._mimetypes.add(mimetypes, conflict_update) return {"content_mimetype:add": added} def content_mimetype_get(self, ids: Iterable[Sha1]) -> List[ContentMimetypeRow]: return self._mimetypes.get(ids) def content_language_missing( self, languages: Iterable[Dict] ) -> List[Tuple[Sha1, int]]: return self._languages.missing(languages) def content_language_get(self, ids: Iterable[Sha1]) -> List[ContentLanguageRow]: return self._languages.get(ids) def content_language_add( self, languages: List[ContentLanguageRow], conflict_update: bool = False ) -> Dict[str, int]: added = self._languages.add(languages, conflict_update) return {"content_language:add": added} - def content_ctags_missing(self, ctags): + def content_ctags_missing(self, ctags: Iterable[Dict]) -> List[Tuple[Sha1, int]]: return self._content_ctags.missing(ctags) - def content_ctags_get(self, ids): - results = [] - for item in self._content_ctags.get(ids): - results.append({"id": item.id, "tool": item.tool, **item.to_dict()}) - return results + def content_ctags_get(self, ids: Iterable[Sha1]) -> List[ContentCtagsRow]: + return self._content_ctags.get(ids) def content_ctags_add( - self, ctags: List[Dict], conflict_update: bool = False + self, ctags: List[ContentCtagsRow], conflict_update: bool = False ) -> Dict[str, int]: - check_id_types(ctags) - added = self._content_ctags.add( - map( - ContentCtagsRow.from_dict, - itertools.chain.from_iterable(map(converters.ctags_to_db, ctags)), - ), - conflict_update, - ) + added = self._content_ctags.add(ctags, conflict_update,) return {"content_ctags:add": added} - def content_ctags_search(self, expression, limit=10, last_sha1=None): + def content_ctags_search( + self, expression: str, limit: int = 10, last_sha1: Optional[Sha1] = None + ) -> List[ContentCtagsRow]: nb_matches = 0 items_per_id: Dict[Tuple[Sha1Git, ToolId], List[ContentCtagsRow]] = {} for item in sorted(self._content_ctags.get_all()): if item.id <= (last_sha1 or bytes(0 for _ in range(SHA1_DIGEST_SIZE))): continue items_per_id.setdefault( (item.id, item.indexer_configuration_id), [] ).append(item) results = [] for items in items_per_id.values(): - ctags = [] for item in items: if item.name != expression: continue nb_matches += 1 if nb_matches > limit: break - item_dict = item.to_dict() - id_ = item_dict.pop("id") - tool = item_dict.pop("tool") - ctags.append(item_dict) - - if ctags: - for ctag in ctags: - results.append({"id": id_, "tool": tool, **ctag}) + results.append(item) + return results def content_fossology_license_get( self, ids: Iterable[Sha1] ) -> List[ContentLicenseRow]: return self._licenses.get(ids) def content_fossology_license_add( self, licenses: List[ContentLicenseRow], conflict_update: bool = False ) -> Dict[str, int]: added = self._licenses.add(licenses, conflict_update) return {"content_fossology_license:add": added} def content_fossology_license_get_partition( self, indexer_configuration_id: int, partition_id: int, nb_partitions: int, page_token: Optional[str] = None, limit: int = 1000, ) -> PagedResult[Sha1]: return self._licenses.get_partition( indexer_configuration_id, partition_id, nb_partitions, page_token, limit ) def content_metadata_missing(self, metadata): return self._content_metadata.missing(metadata) def content_metadata_get(self, ids): return [obj.to_dict() for obj in self._content_metadata.get(ids)] def content_metadata_add( self, metadata: List[Dict], conflict_update: bool = False ) -> Dict[str, int]: check_id_types(metadata) added = self._content_metadata.add( map(ContentMetadataRow.from_dict, metadata), conflict_update ) return {"content_metadata:add": added} def revision_intrinsic_metadata_missing(self, metadata): return self._revision_intrinsic_metadata.missing(metadata) def revision_intrinsic_metadata_get(self, ids): return [obj.to_dict() for obj in self._revision_intrinsic_metadata.get(ids)] def revision_intrinsic_metadata_add( self, metadata: List[Dict], conflict_update: bool = False ) -> Dict[str, int]: check_id_types(metadata) added = self._revision_intrinsic_metadata.add( map(RevisionIntrinsicMetadataRow.from_dict, metadata), conflict_update ) return {"revision_intrinsic_metadata:add": added} def revision_intrinsic_metadata_delete(self, entries: List[Dict]) -> Dict: deleted = self._revision_intrinsic_metadata.delete(entries) return {"revision_intrinsic_metadata:del": deleted} def origin_intrinsic_metadata_get(self, ids): return [obj.to_dict() for obj in self._origin_intrinsic_metadata.get(ids)] def origin_intrinsic_metadata_add( self, metadata: List[Dict], conflict_update: bool = False ) -> Dict[str, int]: added = self._origin_intrinsic_metadata.add( map(OriginIntrinsicMetadataRow.from_dict, metadata), conflict_update ) return {"origin_intrinsic_metadata:add": added} def origin_intrinsic_metadata_delete(self, entries: List[Dict]) -> Dict: deleted = self._origin_intrinsic_metadata.delete(entries) return {"origin_intrinsic_metadata:del": deleted} def origin_intrinsic_metadata_search_fulltext(self, conjunction, limit=100): # A very crude fulltext search implementation, but that's enough # to work on English metadata tokens_re = re.compile("[a-zA-Z0-9]+") search_tokens = list(itertools.chain(*map(tokens_re.findall, conjunction))) def rank(data): # Tokenize the metadata text = json.dumps(data.metadata) text_tokens = tokens_re.findall(text) text_token_occurences = Counter(text_tokens) # Count the number of occurrences of search tokens in the text score = 0 for search_token in search_tokens: if text_token_occurences[search_token] == 0: # Search token is not in the text. return 0 score += text_token_occurences[search_token] # Normalize according to the text's length return score / math.log(len(text_tokens)) results = [ (rank(data), data) for data in self._origin_intrinsic_metadata.get_all() ] results = [(rank_, data) for (rank_, data) in results if rank_ > 0] results.sort( key=operator.itemgetter(0), reverse=True # Don't try to order 'data' ) return [result.to_dict() for (rank_, result) in results[:limit]] def origin_intrinsic_metadata_search_by_producer( self, page_token="", limit=100, ids_only=False, mappings=None, tool_ids=None ): assert isinstance(page_token, str) nb_results = 0 if mappings is not None: mappings = frozenset(mappings) if tool_ids is not None: tool_ids = frozenset(tool_ids) origins = [] # we go to limit+1 to check whether we should add next_page_token in # the response for entry in self._origin_intrinsic_metadata.get_all(): if entry.id <= page_token: continue if nb_results >= (limit + 1): break if mappings is not None and mappings.isdisjoint(entry.mappings): continue if tool_ids is not None and entry.tool["id"] not in tool_ids: continue origins.append(entry.to_dict()) nb_results += 1 result = {} if len(origins) > limit: origins = origins[:limit] result["next_page_token"] = origins[-1]["id"] if ids_only: origins = [origin["id"] for origin in origins] result["origins"] = origins return result def origin_intrinsic_metadata_stats(self): mapping_count = {m: 0 for m in MAPPING_NAMES} total = non_empty = 0 for data in self._origin_intrinsic_metadata.get_all(): total += 1 if set(data.metadata) - {"@context"}: non_empty += 1 for mapping in data.mappings: mapping_count[mapping] += 1 return {"per_mapping": mapping_count, "total": total, "non_empty": non_empty} def indexer_configuration_add(self, tools): inserted = [] for tool in tools: tool = tool.copy() id_ = self._tool_key(tool) tool["id"] = id_ self._tools[id_] = tool inserted.append(tool) return inserted def indexer_configuration_get(self, tool): return self._tools.get(self._tool_key(tool)) def _tool_key(self, tool): return hash( ( tool["tool_name"], tool["tool_version"], json.dumps(tool["tool_configuration"], sort_keys=True), ) ) diff --git a/swh/indexer/storage/interface.py b/swh/indexer/storage/interface.py index f087c9e..8645990 100644 --- a/swh/indexer/storage/interface.py +++ b/swh/indexer/storage/interface.py @@ -1,603 +1,600 @@ # 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 Dict, Iterable, List, Optional, Tuple, TypeVar from swh.core.api import remote_api_endpoint from swh.core.api.classes import PagedResult as CorePagedResult from swh.indexer.storage.model import ( + ContentCtagsRow, ContentLanguageRow, ContentLicenseRow, ContentMimetypeRow, ) TResult = TypeVar("TResult") PagedResult = CorePagedResult[TResult, str] Sha1 = bytes class IndexerStorageInterface: @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_mimetype/missing") def content_mimetype_missing( self, mimetypes: Iterable[Dict] ) -> List[Tuple[Sha1, int]]: """Generate mimetypes missing from storage. Args: mimetypes (iterable): iterable of dict with keys: - **id** (bytes): sha1 identifier - **indexer_configuration_id** (int): tool used to compute the results Returns: list of tuple (id, indexer_configuration_id) missing """ ... @remote_api_endpoint("content_mimetype/range") def content_mimetype_get_partition( self, indexer_configuration_id: int, partition_id: int, nb_partitions: int, page_token: Optional[str] = None, limit: int = 1000, ) -> PagedResult[Sha1]: """Retrieve mimetypes within partition partition_id bound by limit. Args: **indexer_configuration_id**: The tool used to index data **partition_id**: index of the partition to fetch **nb_partitions**: total number of partitions to split into **page_token**: opaque token used for pagination **limit**: Limit result (default to 1000) Raises: IndexerStorageArgumentException for; - limit to None - wrong indexer_type provided Returns: PagedResult of Sha1. If next_page_token is None, there is no more data to fetch """ ... @remote_api_endpoint("content_mimetype/add") def content_mimetype_add( self, mimetypes: List[ContentMimetypeRow], conflict_update: bool = False ) -> Dict[str, int]: """Add mimetypes not present in storage. Args: mimetypes: mimetype rows to be added, with their `tool` attribute set to None. conflict_update: Flag to determine if we want to overwrite (``True``) or skip duplicates (``False``, the default) Returns: Dict summary of number of rows added """ ... @remote_api_endpoint("content_mimetype") def content_mimetype_get(self, ids: Iterable[Sha1]) -> List[ContentMimetypeRow]: """Retrieve full content mimetype per ids. Args: ids: sha1 identifiers Returns: mimetype row objects """ ... @remote_api_endpoint("content_language/missing") def content_language_missing( self, languages: Iterable[Dict] ) -> List[Tuple[Sha1, int]]: """List languages missing from storage. Args: languages (iterable): dictionaries with keys: - **id** (bytes): sha1 identifier - **indexer_configuration_id** (int): tool used to compute the results Returns: list of tuple (id, indexer_configuration_id) missing """ ... @remote_api_endpoint("content_language") def content_language_get(self, ids: Iterable[Sha1]) -> List[ContentLanguageRow]: """Retrieve full content language per ids. Args: ids (iterable): sha1 identifier Returns: language row objects """ ... @remote_api_endpoint("content_language/add") def content_language_add( self, languages: List[ContentLanguageRow], conflict_update: bool = False ) -> Dict[str, int]: """Add languages not present in storage. Args: languages: language row objects conflict_update (bool): Flag to determine if we want to overwrite (true) or skip duplicates (false, the default) Returns: Dict summary of number of rows added """ ... @remote_api_endpoint("content/ctags/missing") - def content_ctags_missing(self, ctags): + def content_ctags_missing(self, ctags: Iterable[Dict]) -> List[Tuple[Sha1, int]]: """List ctags missing from storage. Args: ctags (iterable): dicts with keys: - **id** (bytes): sha1 identifier - **indexer_configuration_id** (int): tool used to compute the results - Yields: - an iterable of missing id for the tuple (id, + Returns: + list of missing id for the tuple (id, indexer_configuration_id) """ ... @remote_api_endpoint("content/ctags") - def content_ctags_get(self, ids): + def content_ctags_get(self, ids: Iterable[Sha1]) -> List[ContentCtagsRow]: """Retrieve ctags per id. Args: ids (iterable): sha1 checksums - Yields: - Dictionaries with keys: - - - **id** (bytes): content's identifier - - **name** (str): symbol's name - - **kind** (str): symbol's kind - - **lang** (str): language for that content - - **tool** (dict): tool used to compute the ctags' info + Returns: + list of language rows """ ... @remote_api_endpoint("content/ctags/add") def content_ctags_add( - self, ctags: List[Dict], conflict_update: bool = False + self, ctags: List[ContentCtagsRow], conflict_update: bool = False ) -> Dict[str, int]: """Add ctags not present in storage Args: ctags (iterable): dictionaries with keys: - **id** (bytes): sha1 - **ctags** ([list): List of dictionary with keys: name, kind, line, lang Returns: Dict summary of number of rows added """ ... @remote_api_endpoint("content/ctags/search") - def content_ctags_search(self, expression, limit=10, last_sha1=None): + def content_ctags_search( + self, expression: str, limit: int = 10, last_sha1: Optional[Sha1] = None + ) -> List[ContentCtagsRow]: """Search through content's raw ctags symbols. Args: expression (str): Expression to search for limit (int): Number of rows to return (default to 10). last_sha1 (str): Offset from which retrieving data (default to ''). - Yields: + Returns: rows of ctags including id, name, lang, kind, line, etc... """ ... @remote_api_endpoint("content/fossology_license") def content_fossology_license_get( self, ids: Iterable[Sha1] ) -> List[ContentLicenseRow]: """Retrieve licenses per id. Args: ids: sha1 identifiers Yields: license rows; possibly more than one per (sha1, tool_id) if there are multiple licenses. """ ... @remote_api_endpoint("content/fossology_license/add") def content_fossology_license_add( self, licenses: List[ContentLicenseRow], conflict_update: bool = False ) -> Dict[str, int]: """Add licenses not present in storage. Args: license: license rows to be added, with their `tool` attribute set to None. conflict_update: Flag to determine if we want to overwrite (true) or skip duplicates (false, the default) Returns: Dict summary of number of rows added """ ... @remote_api_endpoint("content/fossology_license/range") def content_fossology_license_get_partition( self, indexer_configuration_id: int, partition_id: int, nb_partitions: int, page_token: Optional[str] = None, limit: int = 1000, ) -> PagedResult[Sha1]: """Retrieve licenses within the partition partition_id bound by limit. Args: **indexer_configuration_id**: The tool used to index data **partition_id**: index of the partition to fetch **nb_partitions**: total number of partitions to split into **page_token**: opaque token used for pagination **limit**: Limit result (default to 1000) Raises: IndexerStorageArgumentException for; - limit to None - wrong indexer_type provided Returns: PagedResult of Sha1. If next_page_token is None, there is no more data to fetch """ ... @remote_api_endpoint("content_metadata/missing") def content_metadata_missing(self, metadata): """List metadata missing from storage. Args: metadata (iterable): dictionaries with keys: - **id** (bytes): sha1 identifier - **indexer_configuration_id** (int): tool used to compute the results Yields: missing sha1s """ ... @remote_api_endpoint("content_metadata") def content_metadata_get(self, ids): """Retrieve metadata per id. Args: ids (iterable): sha1 checksums Yields: dictionaries with the following keys: id (bytes) metadata (str): associated metadata tool (dict): tool used to compute metadata """ ... @remote_api_endpoint("content_metadata/add") def content_metadata_add( self, metadata: List[Dict], conflict_update: bool = False ) -> Dict[str, int]: """Add metadata not present in storage. Args: metadata (iterable): dictionaries with keys: - **id**: sha1 - **metadata**: arbitrary dict conflict_update: Flag to determine if we want to overwrite (true) or skip duplicates (false, the default) Returns: Dict summary of number of rows added """ ... @remote_api_endpoint("revision_intrinsic_metadata/missing") def revision_intrinsic_metadata_missing(self, metadata): """List metadata missing from storage. Args: metadata (iterable): dictionaries with keys: - **id** (bytes): sha1_git revision identifier - **indexer_configuration_id** (int): tool used to compute the results Yields: missing ids """ ... @remote_api_endpoint("revision_intrinsic_metadata") def revision_intrinsic_metadata_get(self, ids): """Retrieve revision metadata per id. Args: ids (iterable): sha1 checksums Yields: : dictionaries with the following keys: - **id** (bytes) - **metadata** (str): associated metadata - **tool** (dict): tool used to compute metadata - **mappings** (List[str]): list of mappings used to translate these metadata """ ... @remote_api_endpoint("revision_intrinsic_metadata/add") def revision_intrinsic_metadata_add( self, metadata: List[Dict], conflict_update: bool = False ) -> Dict[str, int]: """Add metadata not present in storage. Args: metadata (iterable): dictionaries with keys: - **id**: sha1_git of revision - **metadata**: arbitrary dict - **indexer_configuration_id**: tool used to compute metadata - **mappings** (List[str]): list of mappings used to translate these metadata conflict_update: Flag to determine if we want to overwrite (true) or skip duplicates (false, the default) Returns: Dict summary of number of rows added """ ... @remote_api_endpoint("revision_intrinsic_metadata/delete") def revision_intrinsic_metadata_delete(self, entries: List[Dict]) -> Dict: """Remove revision metadata from the storage. Args: entries (dict): dictionaries with the following keys: - **id** (bytes): revision identifier - **indexer_configuration_id** (int): tool used to compute metadata Returns: Summary of number of rows deleted """ ... @remote_api_endpoint("origin_intrinsic_metadata") def origin_intrinsic_metadata_get(self, ids): """Retrieve origin metadata per id. Args: ids (iterable): origin identifiers Yields: list: dictionaries with the following keys: - **id** (str): origin url - **from_revision** (bytes): which revision this metadata was extracted from - **metadata** (str): associated metadata - **tool** (dict): tool used to compute metadata - **mappings** (List[str]): list of mappings used to translate these metadata """ ... @remote_api_endpoint("origin_intrinsic_metadata/add") def origin_intrinsic_metadata_add( self, metadata: List[Dict], conflict_update: bool = False ) -> Dict[str, int]: """Add origin metadata not present in storage. Args: metadata (iterable): dictionaries with keys: - **id**: origin urls - **from_revision**: sha1 id of the revision used to generate these metadata. - **metadata**: arbitrary dict - **indexer_configuration_id**: tool used to compute metadata - **mappings** (List[str]): list of mappings used to translate these metadata conflict_update: Flag to determine if we want to overwrite (true) or skip duplicates (false, the default) Returns: Dict summary of number of rows added """ ... @remote_api_endpoint("origin_intrinsic_metadata/delete") def origin_intrinsic_metadata_delete(self, entries: List[Dict]) -> Dict: """Remove origin metadata from the storage. Args: entries (dict): dictionaries with the following keys: - **id** (str): origin urls - **indexer_configuration_id** (int): tool used to compute metadata Returns: Summary of number of rows deleted """ ... @remote_api_endpoint("origin_intrinsic_metadata/search/fulltext") def origin_intrinsic_metadata_search_fulltext(self, conjunction, limit=100): """Returns the list of origins whose metadata contain all the terms. Args: conjunction (List[str]): List of terms to be searched for. limit (int): The maximum number of results to return Yields: list: dictionaries with the following keys: - **id** (str): origin urls - **from_revision**: sha1 id of the revision used to generate these metadata. - **metadata** (str): associated metadata - **tool** (dict): tool used to compute metadata - **mappings** (List[str]): list of mappings used to translate these metadata """ ... @remote_api_endpoint("origin_intrinsic_metadata/search/by_producer") def origin_intrinsic_metadata_search_by_producer( self, page_token="", limit=100, ids_only=False, mappings=None, tool_ids=None ): """Returns the list of origins whose metadata contain all the terms. Args: page_token (str): Opaque token used for pagination. limit (int): The maximum number of results to return ids_only (bool): Determines whether only origin urls are returned or the content as well mappings (List[str]): Returns origins whose intrinsic metadata were generated using at least one of these mappings. 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): list of origin url (str) if `ids_only=True` else dictionaries with the following keys: - **id** (str): origin urls - **from_revision**: sha1 id of the revision used to generate these metadata. - **metadata** (str): associated metadata - **tool** (dict): tool used to compute metadata - **mappings** (List[str]): list of mappings used to translate these metadata """ ... @remote_api_endpoint("origin_intrinsic_metadata/stats") def origin_intrinsic_metadata_stats(self): """Returns counts of indexed metadata per origins, broken down into metadata types. Returns: dict: dictionary with keys: - total (int): total number of origins that were indexed (possibly yielding an empty metadata dictionary) - non_empty (int): total number of origins that we extracted a non-empty metadata dictionary from - per_mapping (dict): a dictionary with mapping names as keys and number of origins whose indexing used this mapping. Note that indexing a given origin may use 0, 1, or many mappings. """ ... @remote_api_endpoint("indexer_configuration/add") def indexer_configuration_add(self, tools): """Add new tools to the storage. Args: tools ([dict]): List of dictionary representing tool to insert in the db. Dictionary with the following keys: - **tool_name** (str): tool's name - **tool_version** (str): tool's version - **tool_configuration** (dict): tool's configuration (free form dict) Returns: List of dict inserted in the db (holding the id key as well). The order of the list is not guaranteed to match the order of the initial list. """ ... @remote_api_endpoint("indexer_configuration/data") def indexer_configuration_get(self, tool): """Retrieve tool information. Args: tool (dict): Dictionary representing a tool with the following keys: - **tool_name** (str): tool's name - **tool_version** (str): tool's version - **tool_configuration** (dict): tool's configuration (free form dict) Returns: The same dictionary with an `id` key, None otherwise. """ ... diff --git a/swh/indexer/storage/model.py b/swh/indexer/storage/model.py index eec1c43..c91f1b7 100644 --- a/swh/indexer/storage/model.py +++ b/swh/indexer/storage/model.py @@ -1,114 +1,115 @@ # 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 """Classes used internally by the in-memory idx-storage, and will be used for the interface of the idx-storage in the near future.""" from __future__ import annotations from typing import Any, Dict, List, Optional, Tuple, Type, TypeVar import attr from swh.model.model import Sha1Git, dictify TSelf = TypeVar("TSelf") @attr.s class BaseRow: UNIQUE_KEY_FIELDS: Tuple = ("id", "indexer_configuration_id") id = attr.ib(type=Sha1Git) indexer_configuration_id = attr.ib(type=Optional[int], default=None, kw_only=True) tool = attr.ib(type=Optional[Dict], default=None, kw_only=True) def __attrs_post_init__(self): if self.indexer_configuration_id is None and self.tool is None: raise TypeError("Either indexer_configuration_id or tool must be not None.") if self.indexer_configuration_id is not None and self.tool is not None: raise TypeError( "indexer_configuration_id and tool are mutually exclusive; " "only one may be not None." ) def anonymize(self: TSelf) -> Optional[TSelf]: # Needed to implement swh.journal.writer.ValueProtocol return None def to_dict(self) -> Dict[str, Any]: """Wrapper of `attr.asdict` that can be overridden by subclasses that have special handling of some of the fields.""" d = dictify(attr.asdict(self, recurse=False)) if d["indexer_configuration_id"] is None: del d["indexer_configuration_id"] if d["tool"] is None: del d["tool"] return d @classmethod def from_dict(cls: Type[TSelf], d) -> TSelf: return cls(**d) # type: ignore def unique_key(self) -> Dict: if self.indexer_configuration_id is None: raise ValueError( - "Can only call unique_key() on objects with indexer_configuration_id." + "Can only call unique_key() on objects without " + "indexer_configuration_id." ) return {key: getattr(self, key) for key in self.UNIQUE_KEY_FIELDS} @attr.s class ContentMimetypeRow(BaseRow): mimetype = attr.ib(type=str) encoding = attr.ib(type=str) @attr.s class ContentLanguageRow(BaseRow): lang = attr.ib(type=str) @attr.s class ContentCtagsRow(BaseRow): UNIQUE_KEY_FIELDS = ( "id", "indexer_configuration_id", "name", "kind", "line", "lang", ) name = attr.ib(type=str) kind = attr.ib(type=str) line = attr.ib(type=int) lang = attr.ib(type=str) @attr.s class ContentLicenseRow(BaseRow): UNIQUE_KEY_FIELDS = ("id", "indexer_configuration_id", "license") license = attr.ib(type=str) @attr.s class ContentMetadataRow(BaseRow): metadata = attr.ib(type=Dict[str, Any]) @attr.s class RevisionIntrinsicMetadataRow(BaseRow): metadata = attr.ib(type=Dict[str, Any]) mappings = attr.ib(type=List[str]) @attr.s class OriginIntrinsicMetadataRow(BaseRow): metadata = attr.ib(type=Dict[str, Any]) from_revision = attr.ib(type=Sha1Git) mappings = attr.ib(type=List[str]) diff --git a/swh/indexer/tests/storage/test_storage.py b/swh/indexer/tests/storage/test_storage.py index a64cef6..1cd5e16 100644 --- a/swh/indexer/tests/storage/test_storage.py +++ b/swh/indexer/tests/storage/test_storage.py @@ -1,2049 +1,1935 @@ # 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 inspect import math import threading from typing import Any, Dict, List, Tuple, Union +import attr import pytest from swh.indexer.storage.exc import DuplicateId, IndexerStorageArgumentException from swh.indexer.storage.interface import IndexerStorageInterface from swh.indexer.storage.model import ( BaseRow, + ContentCtagsRow, ContentLanguageRow, ContentLicenseRow, ContentMimetypeRow, ) from swh.model.hashutil import hash_to_bytes def prepare_mimetypes_from_licenses( fossology_licenses: List[ContentLicenseRow], ) -> List[ContentMimetypeRow]: """Fossology license needs some consistent data in db to run. """ mimetypes = [] for c in fossology_licenses: mimetypes.append( ContentMimetypeRow( id=c.id, mimetype="text/plain", # for filtering on textual data to work encoding="utf-8", indexer_configuration_id=c.indexer_configuration_id, ) ) return mimetypes def endpoint_name(etype: str, ename: str) -> str: """Compute the storage's endpoint's name >>> endpoint_name('content_mimetype', 'add') 'content_mimetype_add' >>> endpoint_name('content_fosso_license', 'delete') 'content_fosso_license_delete' """ return f"{etype}_{ename}" def endpoint(storage, etype: str, ename: str): return getattr(storage, endpoint_name(etype, ename)) def expected_summary(count: int, etype: str, ename: str = "add") -> Dict[str, int]: """Compute the expected summary The key is determine according to etype and ename >>> expected_summary(10, 'content_mimetype', 'add') {'content_mimetype:add': 10} >>> expected_summary(9, 'origin_intrinsic_metadata', 'delete') {'origin_intrinsic_metadata:del': 9} """ pattern = ename[0:3] key = endpoint_name(etype, ename).replace(f"_{ename}", f":{pattern}") return {key: count} def test_check_config(swh_indexer_storage) -> None: assert swh_indexer_storage.check_config(check_write=True) assert swh_indexer_storage.check_config(check_write=False) def test_types(swh_indexer_storage) -> None: """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("_", (IndexerStorageInterface,), {})() assert "content_mimetype_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_indexer_storage, meth_name) except AttributeError: 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 == [] class StorageETypeTester: """Base class for testing a series of common behaviour between a bunch of endpoint types supported by an IndexerStorage. This is supposed to be inherited with the following class attributes: - endpoint_type - tool_name - example_data See below for example usage. """ endpoint_type: str tool_name: str example_data: List[Dict] # For old endpoints (which use dicts), this is just the identity function. # For new endpoints, it returns a BaseRow instance. @staticmethod def row_from_dict(d: Dict) -> Union[Dict, BaseRow]: return d # Inverse function of row_from_dict # TODO: remove this once all endpoints are migrated to rows @staticmethod def dict_from_row(r: Union[Dict, BaseRow]) -> Dict: assert isinstance(r, Dict) return r def test_missing( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool_id = data.tools[self.tool_name]["id"] # given 2 (hopefully) unknown objects query = [ {"id": data.sha1_1, "indexer_configuration_id": tool_id,}, {"id": data.sha1_2, "indexer_configuration_id": tool_id,}, ] # we expect these are both returned by the xxx_missing endpoint actual_missing = endpoint(storage, etype, "missing")(query) assert list(actual_missing) == [ data.sha1_1, data.sha1_2, ] # now, when we add one of them summary = endpoint(storage, etype, "add")( [ self.row_from_dict( { "id": data.sha1_2, **self.example_data[0], "indexer_configuration_id": tool_id, } ) ] ) assert summary == expected_summary(1, etype) # we expect only the other one returned actual_missing = endpoint(storage, etype, "missing")(query) assert list(actual_missing) == [data.sha1_1] def test_add__drop_duplicate( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool_id = data.tools[self.tool_name]["id"] # add the first object data_v1 = { "id": data.sha1_2, **self.example_data[0], "indexer_configuration_id": tool_id, } summary = endpoint(storage, etype, "add")([self.row_from_dict(data_v1)]) assert summary == expected_summary(1, etype) # should be able to retrieve it actual_data = list(endpoint(storage, etype, "get")([data.sha1_2])) expected_data_v1 = [ self.row_from_dict( { "id": data.sha1_2, **self.example_data[0], "tool": data.tools[self.tool_name], } ) ] assert actual_data == expected_data_v1 # now if we add a modified version of the same object (same id) data_v2 = data_v1.copy() data_v2.update(self.example_data[1]) summary2 = endpoint(storage, etype, "add")([self.row_from_dict(data_v2)]) assert summary2 == expected_summary(0, etype) # not added # we expect to retrieve the original data, not the modified one actual_data = list(endpoint(storage, etype, "get")([data.sha1_2])) assert actual_data == expected_data_v1 def test_add__update_in_place_duplicate( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool = data.tools[self.tool_name] data_v1 = { "id": data.sha1_2, **self.example_data[0], "indexer_configuration_id": tool["id"], } # given summary = endpoint(storage, etype, "add")([self.row_from_dict(data_v1)]) assert summary == expected_summary(1, etype) # not added # when actual_data = list(endpoint(storage, etype, "get")([data.sha1_2])) expected_data_v1 = [ self.row_from_dict( {"id": data.sha1_2, **self.example_data[0], "tool": tool} ) ] # then assert actual_data == expected_data_v1 # given data_v2 = data_v1.copy() data_v2.update(self.example_data[1]) endpoint(storage, etype, "add")( [self.row_from_dict(data_v2)], conflict_update=True ) assert summary == expected_summary(1, etype) # modified so counted actual_data = list(endpoint(storage, etype, "get")([data.sha1_2])) expected_data_v2 = [ self.row_from_dict( {"id": data.sha1_2, **self.example_data[1], "tool": tool,} ) ] # data did change as the v2 was used to overwrite v1 assert actual_data == expected_data_v2 def test_add__update_in_place_deadlock( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool = data.tools[self.tool_name] hashes = [ hash_to_bytes("34973274ccef6ab4dfaaf86599792fa9c3fe4{:03d}".format(i)) for i in range(1000) ] data_v1 = [ self.row_from_dict( { "id": hash_, **self.example_data[0], "indexer_configuration_id": tool["id"], } ) for hash_ in hashes ] data_v2 = [ self.row_from_dict( { "id": hash_, **self.example_data[1], "indexer_configuration_id": tool["id"], } ) for hash_ in hashes ] # Remove one item from each, so that both queries have to succeed for # all items to be in the DB. data_v2a = data_v2[1:] data_v2b = list(reversed(data_v2[0:-1])) # given endpoint(storage, etype, "add")(data_v1) # when actual_data = list(endpoint(storage, etype, "get")(hashes)) expected_data_v1 = [ self.row_from_dict({"id": hash_, **self.example_data[0], "tool": tool}) for hash_ in hashes ] # then assert actual_data == expected_data_v1 # given def f1() -> None: endpoint(storage, etype, "add")(data_v2a, conflict_update=True) def f2() -> None: endpoint(storage, etype, "add")(data_v2b, conflict_update=True) t1 = threading.Thread(target=f1) t2 = threading.Thread(target=f2) t2.start() t1.start() t1.join() t2.join() actual_data = sorted( map(self.dict_from_row, endpoint(storage, etype, "get")(hashes)), key=lambda x: x["id"], ) expected_data_v2 = [ {"id": hash_, **self.example_data[1], "tool": tool} for hash_ in hashes ] assert actual_data == expected_data_v2 def test_add__duplicate_twice( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool = data.tools[self.tool_name] data_rev1 = self.row_from_dict( { "id": data.revision_id_2, **self.example_data[0], "indexer_configuration_id": tool["id"], } ) data_rev2 = self.row_from_dict( { "id": data.revision_id_2, **self.example_data[1], "indexer_configuration_id": tool["id"], } ) # when summary = endpoint(storage, etype, "add")([data_rev1]) assert summary == expected_summary(1, etype) with pytest.raises(DuplicateId): endpoint(storage, etype, "add")( [data_rev2, data_rev2], conflict_update=True ) # then actual_data = list( endpoint(storage, etype, "get")([data.revision_id_2, data.revision_id_1]) ) expected_data = [ self.row_from_dict( {"id": data.revision_id_2, **self.example_data[0], "tool": tool} ) ] assert actual_data == expected_data def test_get( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool = data.tools[self.tool_name] query = [data.sha1_2, data.sha1_1] data1 = self.row_from_dict( { "id": data.sha1_2, **self.example_data[0], "indexer_configuration_id": tool["id"], } ) # when summary = endpoint(storage, etype, "add")([data1]) assert summary == expected_summary(1, etype) # then actual_data = list(endpoint(storage, etype, "get")(query)) # then expected_data = [ self.row_from_dict( {"id": data.sha1_2, **self.example_data[0], "tool": tool} ) ] assert actual_data == expected_data class TestIndexerStorageContentMimetypes(StorageETypeTester): """Test Indexer Storage content_mimetype related methods """ endpoint_type = "content_mimetype" tool_name = "file" example_data = [ {"mimetype": "text/plain", "encoding": "utf-8",}, {"mimetype": "text/html", "encoding": "us-ascii",}, ] row_from_dict = ContentMimetypeRow.from_dict dict_from_row = staticmethod(lambda x: x.to_dict()) # type: ignore def test_generate_content_mimetype_get_partition_failure( self, swh_indexer_storage: IndexerStorageInterface ) -> None: """get_partition call with wrong limit input should fail""" storage = swh_indexer_storage indexer_configuration_id = 42 with pytest.raises( IndexerStorageArgumentException, match="limit should not be None" ): storage.content_mimetype_get_partition( indexer_configuration_id, 0, 3, limit=None # type: ignore ) def test_generate_content_mimetype_get_partition_no_limit( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: """get_partition should return result""" storage, data = swh_indexer_storage_with_data mimetypes = data.mimetypes expected_ids = set([c.id for c in mimetypes]) indexer_configuration_id = mimetypes[0].indexer_configuration_id assert len(mimetypes) == 16 nb_partitions = 16 actual_ids = [] for partition_id in range(nb_partitions): actual_result = storage.content_mimetype_get_partition( indexer_configuration_id, partition_id, nb_partitions ) assert actual_result.next_page_token is None actual_ids.extend(actual_result.results) assert len(actual_ids) == len(expected_ids) for actual_id in actual_ids: assert actual_id in expected_ids def test_generate_content_mimetype_get_partition_full( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: """get_partition for a single partition should return available ids """ storage, data = swh_indexer_storage_with_data mimetypes = data.mimetypes expected_ids = set([c.id for c in mimetypes]) indexer_configuration_id = mimetypes[0].indexer_configuration_id actual_result = storage.content_mimetype_get_partition( indexer_configuration_id, 0, 1 ) assert actual_result.next_page_token is None actual_ids = actual_result.results assert len(actual_ids) == len(expected_ids) for actual_id in actual_ids: assert actual_id in expected_ids def test_generate_content_mimetype_get_partition_empty( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: """get_partition when at least one of the partitions is empty""" storage, data = swh_indexer_storage_with_data mimetypes = data.mimetypes expected_ids = set([c.id for c in mimetypes]) indexer_configuration_id = mimetypes[0].indexer_configuration_id # nb_partitions = smallest power of 2 such that at least one of # the partitions is empty nb_mimetypes = len(mimetypes) nb_partitions = 1 << math.floor(math.log2(nb_mimetypes) + 1) seen_ids = [] for partition_id in range(nb_partitions): actual_result = storage.content_mimetype_get_partition( indexer_configuration_id, partition_id, nb_partitions, limit=nb_mimetypes + 1, ) for actual_id in actual_result.results: seen_ids.append(actual_id) # Limit is higher than the max number of results assert actual_result.next_page_token is None assert set(seen_ids) == expected_ids def test_generate_content_mimetype_get_partition_with_pagination( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: """get_partition should return ids provided with pagination """ storage, data = swh_indexer_storage_with_data mimetypes = data.mimetypes expected_ids = set([c.id for c in mimetypes]) indexer_configuration_id = mimetypes[0].indexer_configuration_id nb_partitions = 4 actual_ids = [] for partition_id in range(nb_partitions): next_page_token = None while True: actual_result = storage.content_mimetype_get_partition( indexer_configuration_id, partition_id, nb_partitions, limit=2, page_token=next_page_token, ) actual_ids.extend(actual_result.results) next_page_token = actual_result.next_page_token if next_page_token is None: break assert len(set(actual_ids)) == len(set(expected_ids)) for actual_id in actual_ids: assert actual_id in expected_ids class TestIndexerStorageContentLanguage(StorageETypeTester): """Test Indexer Storage content_language related methods """ endpoint_type = "content_language" tool_name = "pygments" example_data = [ {"lang": "haskell",}, {"lang": "common-lisp",}, ] row_from_dict = ContentLanguageRow.from_dict dict_from_row = staticmethod(lambda x: x.to_dict()) # type: ignore class TestIndexerStorageContentCTags(StorageETypeTester): """Test Indexer Storage content_ctags related methods """ endpoint_type = "content_ctags" tool_name = "universal-ctags" example_data = [ - { - "ctags": [ - {"name": "done", "kind": "variable", "line": 119, "lang": "OCaml",} - ] - }, - { - "ctags": [ - {"name": "done", "kind": "variable", "line": 100, "lang": "Python",}, - {"name": "main", "kind": "function", "line": 119, "lang": "Python",}, - ] - }, + {"name": "done", "kind": "variable", "line": 119, "lang": "OCaml",}, + {"name": "done", "kind": "variable", "line": 100, "lang": "Python",}, + {"name": "main", "kind": "function", "line": 119, "lang": "Python",}, ] + row_from_dict = ContentCtagsRow.from_dict + dict_from_row = staticmethod(lambda x: x.to_dict()) # type: ignore # the following tests are disabled because CTAGS behaves differently @pytest.mark.skip def test_add__drop_duplicate(self): pass @pytest.mark.skip def test_add__update_in_place_duplicate(self): pass @pytest.mark.skip def test_add__update_in_place_deadlock(self): pass @pytest.mark.skip def test_add__duplicate_twice(self): pass @pytest.mark.skip def test_get(self): pass def test_content_ctags_search( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data # 1. given tool = data.tools["universal-ctags"] tool_id = tool["id"] - ctag1 = { - "id": data.sha1_1, - "indexer_configuration_id": tool_id, - "ctags": [ + ctags1 = [ + ContentCtagsRow( + id=data.sha1_1, + indexer_configuration_id=tool_id, + **kwargs, # type: ignore + ) + for kwargs in [ {"name": "hello", "kind": "function", "line": 133, "lang": "Python",}, {"name": "counter", "kind": "variable", "line": 119, "lang": "Python",}, {"name": "hello", "kind": "variable", "line": 210, "lang": "Python",}, - ], - } + ] + ] + ctags1_with_tool = [ + attr.evolve(ctag, indexer_configuration_id=None, tool=tool) + for ctag in ctags1 + ] - ctag2 = { - "id": data.sha1_2, - "indexer_configuration_id": tool_id, - "ctags": [ + ctags2 = [ + ContentCtagsRow( + id=data.sha1_2, + indexer_configuration_id=tool_id, + **kwargs, # type: ignore + ) + for kwargs in [ {"name": "hello", "kind": "variable", "line": 100, "lang": "C",}, {"name": "result", "kind": "variable", "line": 120, "lang": "C",}, - ], - } + ] + ] + ctags2_with_tool = [ + attr.evolve(ctag, indexer_configuration_id=None, tool=tool) + for ctag in ctags2 + ] - storage.content_ctags_add([ctag1, ctag2]) + storage.content_ctags_add(ctags1 + ctags2) # 1. when actual_ctags = list(storage.content_ctags_search("hello", limit=1)) # 1. then - assert actual_ctags == [ - { - "id": ctag1["id"], - "tool": tool, - "name": "hello", - "kind": "function", - "line": 133, - "lang": "Python", - } - ] + assert actual_ctags == [ctags1_with_tool[0]] # 2. when actual_ctags = list( - storage.content_ctags_search("hello", limit=1, last_sha1=ctag1["id"]) + storage.content_ctags_search("hello", limit=1, last_sha1=data.sha1_1) ) # 2. then - assert actual_ctags == [ - { - "id": ctag2["id"], - "tool": tool, - "name": "hello", - "kind": "variable", - "line": 100, - "lang": "C", - } - ] + assert actual_ctags == [ctags2_with_tool[0]] # 3. when actual_ctags = list(storage.content_ctags_search("hello")) # 3. then assert actual_ctags == [ - { - "id": ctag1["id"], - "tool": tool, - "name": "hello", - "kind": "function", - "line": 133, - "lang": "Python", - }, - { - "id": ctag1["id"], - "tool": tool, - "name": "hello", - "kind": "variable", - "line": 210, - "lang": "Python", - }, - { - "id": ctag2["id"], - "tool": tool, - "name": "hello", - "kind": "variable", - "line": 100, - "lang": "C", - }, + ctags1_with_tool[0], + ctags1_with_tool[2], + ctags2_with_tool[0], ] # 4. when actual_ctags = list(storage.content_ctags_search("counter")) # then - assert actual_ctags == [ - { - "id": ctag1["id"], - "tool": tool, - "name": "counter", - "kind": "variable", - "line": 119, - "lang": "Python", - } - ] + assert actual_ctags == [ctags1_with_tool[1]] # 5. when actual_ctags = list(storage.content_ctags_search("result", limit=1)) # then - assert actual_ctags == [ - { - "id": ctag2["id"], - "tool": tool, - "name": "result", - "kind": "variable", - "line": 120, - "lang": "C", - } - ] + assert actual_ctags == [ctags2_with_tool[1]] def test_content_ctags_search_no_result( self, swh_indexer_storage: IndexerStorageInterface ) -> None: storage = swh_indexer_storage actual_ctags = list(storage.content_ctags_search("counter")) assert not actual_ctags def test_content_ctags_add__add_new_ctags_added( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data # given tool = data.tools["universal-ctags"] tool_id = tool["id"] - ctag_v1 = { - "id": data.sha1_2, - "indexer_configuration_id": tool_id, - "ctags": [ - {"name": "done", "kind": "variable", "line": 100, "lang": "Scheme",} - ], - } + ctag1 = ContentCtagsRow( + id=data.sha1_2, + indexer_configuration_id=tool_id, + name="done", + kind="variable", + line=100, + lang="Scheme", + ) + ctag1_with_tool = attr.evolve(ctag1, indexer_configuration_id=None, tool=tool) # given - storage.content_ctags_add([ctag_v1]) - storage.content_ctags_add([ctag_v1]) # conflict does nothing + storage.content_ctags_add([ctag1]) + storage.content_ctags_add([ctag1]) # conflict does nothing # when actual_ctags = list(storage.content_ctags_get([data.sha1_2])) # then - expected_ctags = [ - { - "id": data.sha1_2, - "name": "done", - "kind": "variable", - "line": 100, - "lang": "Scheme", - "tool": tool, - } - ] - - assert actual_ctags == expected_ctags + assert actual_ctags == [ctag1_with_tool] # given - ctag_v2 = ctag_v1.copy() - ctag_v2.update( - { - "ctags": [ - {"name": "defn", "kind": "function", "line": 120, "lang": "Scheme",} - ] - } + ctag2 = ContentCtagsRow( + id=data.sha1_2, + indexer_configuration_id=tool_id, + name="defn", + kind="function", + line=120, + lang="Scheme", ) + ctag2_with_tool = attr.evolve(ctag2, indexer_configuration_id=None, tool=tool) - storage.content_ctags_add([ctag_v2]) - - expected_ctags = [ - { - "id": data.sha1_2, - "name": "done", - "kind": "variable", - "line": 100, - "lang": "Scheme", - "tool": tool, - }, - { - "id": data.sha1_2, - "name": "defn", - "kind": "function", - "line": 120, - "lang": "Scheme", - "tool": tool, - }, - ] + storage.content_ctags_add([ctag2]) actual_ctags = list(storage.content_ctags_get([data.sha1_2])) - assert actual_ctags == expected_ctags + assert actual_ctags == [ctag1_with_tool, ctag2_with_tool] def test_content_ctags_add__update_in_place( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data # given tool = data.tools["universal-ctags"] tool_id = tool["id"] - ctag_v1 = { - "id": data.sha1_2, - "indexer_configuration_id": tool_id, - "ctags": [ - {"name": "done", "kind": "variable", "line": 100, "lang": "Scheme",} - ], - } + ctag1 = ContentCtagsRow( + id=data.sha1_2, + indexer_configuration_id=tool_id, + name="done", + kind="variable", + line=100, + lang="Scheme", + ) + ctag1_with_tool = attr.evolve(ctag1, indexer_configuration_id=None, tool=tool) # given - storage.content_ctags_add([ctag_v1]) + storage.content_ctags_add([ctag1]) # when actual_ctags = list(storage.content_ctags_get([data.sha1_2])) # then - expected_ctags = [ - { - "id": data.sha1_2, - "name": "done", - "kind": "variable", - "line": 100, - "lang": "Scheme", - "tool": tool, - } - ] - assert actual_ctags == expected_ctags + assert actual_ctags == [ctag1_with_tool] # given - ctag_v2 = ctag_v1.copy() - ctag_v2.update( - { - "ctags": [ - { - "name": "done", - "kind": "variable", - "line": 100, - "lang": "Scheme", - }, - { - "name": "defn", - "kind": "function", - "line": 120, - "lang": "Scheme", - }, - ] - } + ctag2 = ContentCtagsRow( + id=data.sha1_2, + indexer_configuration_id=tool_id, + name="defn", + kind="function", + line=120, + lang="Scheme", ) + ctag2_with_tool = attr.evolve(ctag2, indexer_configuration_id=None, tool=tool) - storage.content_ctags_add([ctag_v2], conflict_update=True) + storage.content_ctags_add([ctag1, ctag2], conflict_update=True) actual_ctags = list(storage.content_ctags_get([data.sha1_2])) - # ctag did change as the v2 was used to overwrite v1 - expected_ctags = [ - { - "id": data.sha1_2, - "name": "done", - "kind": "variable", - "line": 100, - "lang": "Scheme", - "tool": tool, - }, - { - "id": data.sha1_2, - "name": "defn", - "kind": "function", - "line": 120, - "lang": "Scheme", - "tool": tool, - }, - ] - assert actual_ctags == expected_ctags + assert actual_ctags == [ctag1_with_tool, ctag2_with_tool] def test_add_empty( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: (storage, data) = swh_indexer_storage_with_data etype = self.endpoint_type - tool = data.tools[self.tool_name] - summary = endpoint(storage, etype, "add")( - [{"id": data.sha1_2, "indexer_configuration_id": tool["id"], "ctags": [],}] - ) + summary = endpoint(storage, etype, "add")([]) assert summary == {"content_ctags:add": 0} actual_ctags = list(endpoint(storage, etype, "get")([data.sha1_2])) assert actual_ctags == [] def test_get_unknown( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: (storage, data) = swh_indexer_storage_with_data etype = self.endpoint_type actual_ctags = list(endpoint(storage, etype, "get")([data.sha1_2])) assert actual_ctags == [] class TestIndexerStorageContentMetadata(StorageETypeTester): """Test Indexer Storage content_metadata related methods """ tool_name = "swh-metadata-detector" endpoint_type = "content_metadata" example_data = [ { "metadata": { "other": {}, "codeRepository": { "type": "git", "url": "https://github.com/moranegg/metadata_test", }, "description": "Simple package.json test for indexer", "name": "test_metadata", "version": "0.0.1", }, }, {"metadata": {"other": {}, "name": "test_metadata", "version": "0.0.1"},}, ] class TestIndexerStorageRevisionIntrinsicMetadata(StorageETypeTester): """Test Indexer Storage revision_intrinsic_metadata related methods """ tool_name = "swh-metadata-detector" endpoint_type = "revision_intrinsic_metadata" example_data = [ { "metadata": { "other": {}, "codeRepository": { "type": "git", "url": "https://github.com/moranegg/metadata_test", }, "description": "Simple package.json test for indexer", "name": "test_metadata", "version": "0.0.1", }, "mappings": ["mapping1"], }, { "metadata": {"other": {}, "name": "test_metadata", "version": "0.0.1"}, "mappings": ["mapping2"], }, ] def test_revision_intrinsic_metadata_delete( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool = data.tools[self.tool_name] query = [data.sha1_2, data.sha1_1] data1 = { "id": data.sha1_2, **self.example_data[0], "indexer_configuration_id": tool["id"], } # when summary = endpoint(storage, etype, "add")([data1]) assert summary == expected_summary(1, etype) summary2 = endpoint(storage, etype, "delete")( [{"id": data.sha1_2, "indexer_configuration_id": tool["id"],}] ) assert summary2 == expected_summary(1, etype, "del") # then actual_data = list(endpoint(storage, etype, "get")(query)) # then assert not actual_data def test_revision_intrinsic_metadata_delete_nonexisting( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool = data.tools[self.tool_name] endpoint(storage, etype, "delete")( [{"id": data.sha1_2, "indexer_configuration_id": tool["id"],}] ) class TestIndexerStorageContentFossologyLicense: endpoint_type = "content_fossology_license" tool_name = "nomos" row_from_dict = ContentLicenseRow.from_dict dict_from_row = staticmethod(lambda x: x.to_dict()) def test_content_fossology_license_add__new_license_added( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data # given tool = data.tools["nomos"] tool_id = tool["id"] license1 = ContentLicenseRow( id=data.sha1_1, license="Apache-2.0", indexer_configuration_id=tool_id, ) # given storage.content_fossology_license_add([license1]) # conflict does nothing storage.content_fossology_license_add([license1]) # when actual_licenses = list(storage.content_fossology_license_get([data.sha1_1])) # then expected_licenses = [ ContentLicenseRow(id=data.sha1_1, license="Apache-2.0", tool=tool,) ] assert actual_licenses == expected_licenses # given license2 = ContentLicenseRow( id=data.sha1_1, license="BSD-2-Clause", indexer_configuration_id=tool_id, ) storage.content_fossology_license_add([license2]) actual_licenses = list(storage.content_fossology_license_get([data.sha1_1])) expected_licenses.append( ContentLicenseRow(id=data.sha1_1, license="BSD-2-Clause", tool=tool,) ) # first license was not removed when the second one was added assert sorted(actual_licenses) == sorted(expected_licenses) def test_generate_content_fossology_license_get_partition_failure( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: """get_partition call with wrong limit input should fail""" storage, data = swh_indexer_storage_with_data indexer_configuration_id = 42 with pytest.raises( IndexerStorageArgumentException, match="limit should not be None" ): storage.content_fossology_license_get_partition( indexer_configuration_id, 0, 3, limit=None, # type: ignore ) def test_generate_content_fossology_license_get_partition_no_limit( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: """get_partition should return results""" storage, data = swh_indexer_storage_with_data # craft some consistent mimetypes fossology_licenses = data.fossology_licenses mimetypes = prepare_mimetypes_from_licenses(fossology_licenses) indexer_configuration_id = fossology_licenses[0].indexer_configuration_id storage.content_mimetype_add(mimetypes, conflict_update=True) # add fossology_licenses to storage storage.content_fossology_license_add(fossology_licenses) # All ids from the db expected_ids = set([c.id for c in fossology_licenses]) assert len(fossology_licenses) == 10 assert len(mimetypes) == 10 nb_partitions = 4 actual_ids = [] for partition_id in range(nb_partitions): actual_result = storage.content_fossology_license_get_partition( indexer_configuration_id, partition_id, nb_partitions ) assert actual_result.next_page_token is None actual_ids.extend(actual_result.results) assert len(set(actual_ids)) == len(expected_ids) for actual_id in actual_ids: assert actual_id in expected_ids def test_generate_content_fossology_license_get_partition_full( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: """get_partition for a single partition should return available ids """ storage, data = swh_indexer_storage_with_data # craft some consistent mimetypes fossology_licenses = data.fossology_licenses mimetypes = prepare_mimetypes_from_licenses(fossology_licenses) indexer_configuration_id = fossology_licenses[0].indexer_configuration_id storage.content_mimetype_add(mimetypes, conflict_update=True) # add fossology_licenses to storage storage.content_fossology_license_add(fossology_licenses) # All ids from the db expected_ids = set([c.id for c in fossology_licenses]) actual_result = storage.content_fossology_license_get_partition( indexer_configuration_id, 0, 1 ) assert actual_result.next_page_token is None actual_ids = actual_result.results assert len(set(actual_ids)) == len(expected_ids) for actual_id in actual_ids: assert actual_id in expected_ids def test_generate_content_fossology_license_get_partition_empty( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: """get_partition when at least one of the partitions is empty""" storage, data = swh_indexer_storage_with_data # craft some consistent mimetypes fossology_licenses = data.fossology_licenses mimetypes = prepare_mimetypes_from_licenses(fossology_licenses) indexer_configuration_id = fossology_licenses[0].indexer_configuration_id storage.content_mimetype_add(mimetypes, conflict_update=True) # add fossology_licenses to storage storage.content_fossology_license_add(fossology_licenses) # All ids from the db expected_ids = set([c.id for c in fossology_licenses]) # nb_partitions = smallest power of 2 such that at least one of # the partitions is empty nb_licenses = len(fossology_licenses) nb_partitions = 1 << math.floor(math.log2(nb_licenses) + 1) seen_ids = [] for partition_id in range(nb_partitions): actual_result = storage.content_fossology_license_get_partition( indexer_configuration_id, partition_id, nb_partitions, limit=nb_licenses + 1, ) for actual_id in actual_result.results: seen_ids.append(actual_id) # Limit is higher than the max number of results assert actual_result.next_page_token is None assert set(seen_ids) == expected_ids def test_generate_content_fossology_license_get_partition_with_pagination( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: """get_partition should return ids provided with paginationv """ storage, data = swh_indexer_storage_with_data # craft some consistent mimetypes fossology_licenses = data.fossology_licenses mimetypes = prepare_mimetypes_from_licenses(fossology_licenses) indexer_configuration_id = fossology_licenses[0].indexer_configuration_id storage.content_mimetype_add(mimetypes, conflict_update=True) # add fossology_licenses to storage storage.content_fossology_license_add(fossology_licenses) # All ids from the db expected_ids = [c.id for c in fossology_licenses] nb_partitions = 4 actual_ids = [] for partition_id in range(nb_partitions): next_page_token = None while True: actual_result = storage.content_fossology_license_get_partition( indexer_configuration_id, partition_id, nb_partitions, limit=2, page_token=next_page_token, ) actual_ids.extend(actual_result.results) next_page_token = actual_result.next_page_token if next_page_token is None: break assert len(set(actual_ids)) == len(set(expected_ids)) for actual_id in actual_ids: assert actual_id in expected_ids def test_add_empty( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: (storage, data) = swh_indexer_storage_with_data etype = self.endpoint_type summary = endpoint(storage, etype, "add")([]) assert summary == {"content_fossology_license:add": 0} actual_license = list(endpoint(storage, etype, "get")([data.sha1_2])) assert actual_license == [] def test_get_unknown( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: (storage, data) = swh_indexer_storage_with_data etype = self.endpoint_type actual_license = list(endpoint(storage, etype, "get")([data.sha1_2])) assert actual_license == [] class TestIndexerStorageOriginIntrinsicMetadata: def test_origin_intrinsic_metadata_get( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data # given tool_id = data.tools["swh-metadata-detector"]["id"] metadata = { "version": None, "name": None, } metadata_rev = { "id": data.revision_id_2, "metadata": metadata, "mappings": ["mapping1"], "indexer_configuration_id": tool_id, } metadata_origin = { "id": data.origin_url_1, "metadata": metadata, "indexer_configuration_id": tool_id, "mappings": ["mapping1"], "from_revision": data.revision_id_2, } # when storage.revision_intrinsic_metadata_add([metadata_rev]) storage.origin_intrinsic_metadata_add([metadata_origin]) # then actual_metadata = list( storage.origin_intrinsic_metadata_get([data.origin_url_1, "no://where"]) ) expected_metadata = [ { "id": data.origin_url_1, "metadata": metadata, "tool": data.tools["swh-metadata-detector"], "from_revision": data.revision_id_2, "mappings": ["mapping1"], } ] assert actual_metadata == expected_metadata def test_origin_intrinsic_metadata_delete( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data # given tool_id = data.tools["swh-metadata-detector"]["id"] metadata = { "version": None, "name": None, } metadata_rev = { "id": data.revision_id_2, "metadata": metadata, "mappings": ["mapping1"], "indexer_configuration_id": tool_id, } metadata_origin = { "id": data.origin_url_1, "metadata": metadata, "indexer_configuration_id": tool_id, "mappings": ["mapping1"], "from_revision": data.revision_id_2, } metadata_origin2 = metadata_origin.copy() metadata_origin2["id"] = data.origin_url_2 # when storage.revision_intrinsic_metadata_add([metadata_rev]) storage.origin_intrinsic_metadata_add([metadata_origin, metadata_origin2]) storage.origin_intrinsic_metadata_delete( [{"id": data.origin_url_1, "indexer_configuration_id": tool_id}] ) # then actual_metadata = list( storage.origin_intrinsic_metadata_get( [data.origin_url_1, data.origin_url_2, "no://where"] ) ) for item in actual_metadata: item["indexer_configuration_id"] = item.pop("tool")["id"] assert actual_metadata == [metadata_origin2] def test_origin_intrinsic_metadata_delete_nonexisting( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data tool_id = data.tools["swh-metadata-detector"]["id"] storage.origin_intrinsic_metadata_delete( [{"id": data.origin_url_1, "indexer_configuration_id": tool_id}] ) def test_origin_intrinsic_metadata_add_drop_duplicate( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data # given tool_id = data.tools["swh-metadata-detector"]["id"] metadata_v1: Dict[str, Any] = { "version": None, "name": None, } metadata_rev_v1 = { "id": data.revision_id_1, "metadata": metadata_v1.copy(), "mappings": [], "indexer_configuration_id": tool_id, } metadata_origin_v1 = { "id": data.origin_url_1, "metadata": metadata_v1.copy(), "indexer_configuration_id": tool_id, "mappings": [], "from_revision": data.revision_id_1, } # given storage.revision_intrinsic_metadata_add([metadata_rev_v1]) storage.origin_intrinsic_metadata_add([metadata_origin_v1]) # when actual_metadata = list( storage.origin_intrinsic_metadata_get([data.origin_url_1, "no://where"]) ) expected_metadata_v1 = [ { "id": data.origin_url_1, "metadata": metadata_v1, "tool": data.tools["swh-metadata-detector"], "from_revision": data.revision_id_1, "mappings": [], } ] assert actual_metadata == expected_metadata_v1 # given metadata_v2 = metadata_v1.copy() metadata_v2.update( {"name": "test_metadata", "author": "MG",} ) metadata_rev_v2 = metadata_rev_v1.copy() metadata_origin_v2 = metadata_origin_v1.copy() metadata_rev_v2["metadata"] = metadata_v2 metadata_origin_v2["metadata"] = metadata_v2 storage.revision_intrinsic_metadata_add([metadata_rev_v2]) storage.origin_intrinsic_metadata_add([metadata_origin_v2]) # then actual_metadata = list( storage.origin_intrinsic_metadata_get([data.origin_url_1]) ) # metadata did not change as the v2 was dropped. assert actual_metadata == expected_metadata_v1 def test_origin_intrinsic_metadata_add_update_in_place_duplicate( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data # given tool_id = data.tools["swh-metadata-detector"]["id"] metadata_v1: Dict[str, Any] = { "version": None, "name": None, } metadata_rev_v1 = { "id": data.revision_id_2, "metadata": metadata_v1, "mappings": [], "indexer_configuration_id": tool_id, } metadata_origin_v1 = { "id": data.origin_url_1, "metadata": metadata_v1.copy(), "indexer_configuration_id": tool_id, "mappings": [], "from_revision": data.revision_id_2, } # given storage.revision_intrinsic_metadata_add([metadata_rev_v1]) storage.origin_intrinsic_metadata_add([metadata_origin_v1]) # when actual_metadata = list( storage.origin_intrinsic_metadata_get([data.origin_url_1]) ) # then expected_metadata_v1 = [ { "id": data.origin_url_1, "metadata": metadata_v1, "tool": data.tools["swh-metadata-detector"], "from_revision": data.revision_id_2, "mappings": [], } ] assert actual_metadata == expected_metadata_v1 # given metadata_v2 = metadata_v1.copy() metadata_v2.update( {"name": "test_update_duplicated_metadata", "author": "MG",} ) metadata_rev_v2 = metadata_rev_v1.copy() metadata_origin_v2 = metadata_origin_v1.copy() metadata_rev_v2["metadata"] = metadata_v2 metadata_origin_v2 = { "id": data.origin_url_1, "metadata": metadata_v2.copy(), "indexer_configuration_id": tool_id, "mappings": ["npm"], "from_revision": data.revision_id_1, } storage.revision_intrinsic_metadata_add([metadata_rev_v2], conflict_update=True) storage.origin_intrinsic_metadata_add( [metadata_origin_v2], conflict_update=True ) actual_metadata = list( storage.origin_intrinsic_metadata_get([data.origin_url_1]) ) expected_metadata_v2 = [ { "id": data.origin_url_1, "metadata": metadata_v2, "tool": data.tools["swh-metadata-detector"], "from_revision": data.revision_id_1, "mappings": ["npm"], } ] # metadata did change as the v2 was used to overwrite v1 assert actual_metadata == expected_metadata_v2 def test_origin_intrinsic_metadata_add__update_in_place_deadlock( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data # given tool_id = data.tools["swh-metadata-detector"]["id"] ids = list(range(10)) example_data1 = { "metadata": {"version": None, "name": None,}, "mappings": [], } example_data2 = { "metadata": {"version": "v1.1.1", "name": "foo",}, "mappings": [], } metadata_rev_v1 = { "id": data.revision_id_2, "metadata": {"version": None, "name": None,}, "mappings": [], "indexer_configuration_id": tool_id, } data_v1 = [ { "id": "file:///tmp/origin%d" % id_, "from_revision": data.revision_id_2, **example_data1, "indexer_configuration_id": tool_id, } for id_ in ids ] data_v2 = [ { "id": "file:///tmp/origin%d" % id_, "from_revision": data.revision_id_2, **example_data2, "indexer_configuration_id": tool_id, } for id_ in ids ] # Remove one item from each, so that both queries have to succeed for # all items to be in the DB. data_v2a = data_v2[1:] data_v2b = list(reversed(data_v2[0:-1])) # given storage.revision_intrinsic_metadata_add([metadata_rev_v1]) storage.origin_intrinsic_metadata_add(data_v1) # when origins = ["file:///tmp/origin%d" % i for i in ids] actual_data = list(storage.origin_intrinsic_metadata_get(origins)) expected_data_v1 = [ { "id": "file:///tmp/origin%d" % id_, "from_revision": data.revision_id_2, **example_data1, "tool": data.tools["swh-metadata-detector"], } for id_ in ids ] # then assert actual_data == expected_data_v1 # given def f1() -> None: storage.origin_intrinsic_metadata_add(data_v2a, conflict_update=True) def f2() -> None: storage.origin_intrinsic_metadata_add(data_v2b, conflict_update=True) t1 = threading.Thread(target=f1) t2 = threading.Thread(target=f2) t2.start() t1.start() t1.join() t2.join() actual_data = list(storage.origin_intrinsic_metadata_get(origins)) expected_data_v2 = [ { "id": "file:///tmp/origin%d" % id_, "from_revision": data.revision_id_2, **example_data2, "tool": data.tools["swh-metadata-detector"], } for id_ in ids ] assert len(actual_data) == len(expected_data_v2) assert sorted(actual_data, key=lambda x: x["id"]) == expected_data_v2 def test_origin_intrinsic_metadata_add__duplicate_twice( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data # given tool_id = data.tools["swh-metadata-detector"]["id"] metadata = { "developmentStatus": None, "name": None, } metadata_rev = { "id": data.revision_id_2, "metadata": metadata, "mappings": ["mapping1"], "indexer_configuration_id": tool_id, } metadata_origin = { "id": data.origin_url_1, "metadata": metadata, "indexer_configuration_id": tool_id, "mappings": ["mapping1"], "from_revision": data.revision_id_2, } # when storage.revision_intrinsic_metadata_add([metadata_rev]) with pytest.raises(DuplicateId): storage.origin_intrinsic_metadata_add([metadata_origin, metadata_origin]) def test_origin_intrinsic_metadata_search_fulltext( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data # given tool_id = data.tools["swh-metadata-detector"]["id"] metadata1 = { "author": "John Doe", } metadata1_rev = { "id": data.revision_id_1, "metadata": metadata1, "mappings": [], "indexer_configuration_id": tool_id, } metadata1_origin = { "id": data.origin_url_1, "metadata": metadata1, "mappings": [], "indexer_configuration_id": tool_id, "from_revision": data.revision_id_1, } metadata2 = { "author": "Jane Doe", } metadata2_rev = { "id": data.revision_id_2, "metadata": metadata2, "mappings": [], "indexer_configuration_id": tool_id, } metadata2_origin = { "id": data.origin_url_2, "metadata": metadata2, "mappings": [], "indexer_configuration_id": tool_id, "from_revision": data.revision_id_2, } # when storage.revision_intrinsic_metadata_add([metadata1_rev]) storage.origin_intrinsic_metadata_add([metadata1_origin]) storage.revision_intrinsic_metadata_add([metadata2_rev]) storage.origin_intrinsic_metadata_add([metadata2_origin]) # then search = storage.origin_intrinsic_metadata_search_fulltext assert set([res["id"] for res in search(["Doe"])]) == set( [data.origin_url_1, data.origin_url_2] ) assert [res["id"] for res in search(["John", "Doe"])] == [data.origin_url_1] assert [res["id"] for res in search(["John"])] == [data.origin_url_1] assert not list(search(["John", "Jane"])) def test_origin_intrinsic_metadata_search_fulltext_rank( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data # given tool_id = data.tools["swh-metadata-detector"]["id"] # The following authors have "Random Person" to add some more content # to the JSON data, to work around normalization quirks when there # are few words (rank/(1+ln(nb_words)) is very sensitive to nb_words # for small values of nb_words). metadata1 = {"author": ["Random Person", "John Doe", "Jane Doe",]} metadata1_rev = { "id": data.revision_id_1, "metadata": metadata1, "mappings": [], "indexer_configuration_id": tool_id, } metadata1_origin = { "id": data.origin_url_1, "metadata": metadata1, "mappings": [], "indexer_configuration_id": tool_id, "from_revision": data.revision_id_1, } metadata2 = {"author": ["Random Person", "Jane Doe",]} metadata2_rev = { "id": data.revision_id_2, "metadata": metadata2, "mappings": [], "indexer_configuration_id": tool_id, } metadata2_origin = { "id": data.origin_url_2, "metadata": metadata2, "mappings": [], "indexer_configuration_id": tool_id, "from_revision": data.revision_id_2, } # when storage.revision_intrinsic_metadata_add([metadata1_rev]) storage.origin_intrinsic_metadata_add([metadata1_origin]) storage.revision_intrinsic_metadata_add([metadata2_rev]) storage.origin_intrinsic_metadata_add([metadata2_origin]) # then search = storage.origin_intrinsic_metadata_search_fulltext assert [res["id"] for res in search(["Doe"])] == [ data.origin_url_1, data.origin_url_2, ] assert [res["id"] for res in search(["Doe"], limit=1)] == [data.origin_url_1] assert [res["id"] for res in search(["John"])] == [data.origin_url_1] assert [res["id"] for res in search(["Jane"])] == [ data.origin_url_2, data.origin_url_1, ] assert [res["id"] for res in search(["John", "Jane"])] == [data.origin_url_1] def _fill_origin_intrinsic_metadata( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data tool1_id = data.tools["swh-metadata-detector"]["id"] tool2_id = data.tools["swh-metadata-detector2"]["id"] metadata1 = { "@context": "foo", "author": "John Doe", } metadata1_rev = { "id": data.revision_id_1, "metadata": metadata1, "mappings": ["npm"], "indexer_configuration_id": tool1_id, } metadata1_origin = { "id": data.origin_url_1, "metadata": metadata1, "mappings": ["npm"], "indexer_configuration_id": tool1_id, "from_revision": data.revision_id_1, } metadata2 = { "@context": "foo", "author": "Jane Doe", } metadata2_rev = { "id": data.revision_id_2, "metadata": metadata2, "mappings": ["npm", "gemspec"], "indexer_configuration_id": tool2_id, } metadata2_origin = { "id": data.origin_url_2, "metadata": metadata2, "mappings": ["npm", "gemspec"], "indexer_configuration_id": tool2_id, "from_revision": data.revision_id_2, } metadata3 = { "@context": "foo", } metadata3_rev = { "id": data.revision_id_3, "metadata": metadata3, "mappings": ["npm", "gemspec"], "indexer_configuration_id": tool2_id, } metadata3_origin = { "id": data.origin_url_3, "metadata": metadata3, "mappings": ["pkg-info"], "indexer_configuration_id": tool2_id, "from_revision": data.revision_id_3, } storage.revision_intrinsic_metadata_add([metadata1_rev]) storage.origin_intrinsic_metadata_add([metadata1_origin]) storage.revision_intrinsic_metadata_add([metadata2_rev]) storage.origin_intrinsic_metadata_add([metadata2_origin]) storage.revision_intrinsic_metadata_add([metadata3_rev]) storage.origin_intrinsic_metadata_add([metadata3_origin]) def test_origin_intrinsic_metadata_search_by_producer( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data self._fill_origin_intrinsic_metadata(swh_indexer_storage_with_data) tool1 = data.tools["swh-metadata-detector"] tool2 = data.tools["swh-metadata-detector2"] endpoint = storage.origin_intrinsic_metadata_search_by_producer # test pagination # no 'page_token' param, return all origins result = endpoint(ids_only=True) assert result["origins"] == [ data.origin_url_1, data.origin_url_2, data.origin_url_3, ] assert "next_page_token" not in result # 'page_token' is < than origin_1, return everything result = endpoint(page_token=data.origin_url_1[:-1], ids_only=True) assert result["origins"] == [ data.origin_url_1, data.origin_url_2, data.origin_url_3, ] assert "next_page_token" not in result # 'page_token' is origin_3, return nothing result = endpoint(page_token=data.origin_url_3, ids_only=True) assert not result["origins"] assert "next_page_token" not in result # test limit argument result = endpoint(page_token=data.origin_url_1[:-1], limit=2, ids_only=True) assert result["origins"] == [data.origin_url_1, data.origin_url_2] assert result["next_page_token"] == result["origins"][-1] result = endpoint(page_token=data.origin_url_1, limit=2, ids_only=True) assert result["origins"] == [data.origin_url_2, data.origin_url_3] assert "next_page_token" not in result result = endpoint(page_token=data.origin_url_2, limit=2, ids_only=True) assert result["origins"] == [data.origin_url_3] assert "next_page_token" not in result # test mappings filtering result = endpoint(mappings=["npm"], ids_only=True) assert result["origins"] == [data.origin_url_1, data.origin_url_2] assert "next_page_token" not in result result = endpoint(mappings=["npm", "gemspec"], ids_only=True) assert result["origins"] == [data.origin_url_1, data.origin_url_2] assert "next_page_token" not in result result = endpoint(mappings=["gemspec"], ids_only=True) assert result["origins"] == [data.origin_url_2] assert "next_page_token" not in result result = endpoint(mappings=["pkg-info"], ids_only=True) assert result["origins"] == [data.origin_url_3] assert "next_page_token" not in result result = endpoint(mappings=["foobar"], ids_only=True) assert not result["origins"] assert "next_page_token" not in result # test pagination + mappings result = endpoint(mappings=["npm"], limit=1, ids_only=True) assert result["origins"] == [data.origin_url_1] assert result["next_page_token"] == result["origins"][-1] # test tool filtering result = endpoint(tool_ids=[tool1["id"]], ids_only=True) assert result["origins"] == [data.origin_url_1] assert "next_page_token" not in result result = endpoint(tool_ids=[tool2["id"]], ids_only=True) assert sorted(result["origins"]) == [data.origin_url_2, data.origin_url_3] assert "next_page_token" not in result result = endpoint(tool_ids=[tool1["id"], tool2["id"]], ids_only=True) assert sorted(result["origins"]) == [ data.origin_url_1, data.origin_url_2, data.origin_url_3, ] assert "next_page_token" not in result # test ids_only=False assert endpoint(mappings=["gemspec"])["origins"] == [ { "id": data.origin_url_2, "metadata": {"@context": "foo", "author": "Jane Doe",}, "mappings": ["npm", "gemspec"], "tool": tool2, "from_revision": data.revision_id_2, } ] def test_origin_intrinsic_metadata_stats( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data self._fill_origin_intrinsic_metadata(swh_indexer_storage_with_data) result = storage.origin_intrinsic_metadata_stats() assert result == { "per_mapping": { "gemspec": 1, "npm": 2, "pkg-info": 1, "codemeta": 0, "maven": 0, }, "total": 3, "non_empty": 2, } class TestIndexerStorageIndexerCondifuration: def test_indexer_configuration_add( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data tool = { "tool_name": "some-unknown-tool", "tool_version": "some-version", "tool_configuration": {"debian-package": "some-package"}, } actual_tool = storage.indexer_configuration_get(tool) assert actual_tool is None # does not exist # add it actual_tools = list(storage.indexer_configuration_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 = list(storage.indexer_configuration_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_indexer_configuration_add_multiple( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data tool = { "tool_name": "some-unknown-tool", "tool_version": "some-version", "tool_configuration": {"debian-package": "some-package"}, } actual_tools = list(storage.indexer_configuration_add([tool])) assert len(actual_tools) == 1 new_tools = [ tool, { "tool_name": "yet-another-tool", "tool_version": "version", "tool_configuration": {}, }, ] actual_tools = list(storage.indexer_configuration_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_indexer_configuration_get_missing( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data tool = { "tool_name": "unknown-tool", "tool_version": "3.1.0rc2-31-ga2cbb8c", "tool_configuration": {"command_line": "nomossa "}, } actual_tool = storage.indexer_configuration_get(tool) assert actual_tool is None def test_indexer_configuration_get( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data tool = { "tool_name": "nomos", "tool_version": "3.1.0rc2-31-ga2cbb8c", "tool_configuration": {"command_line": "nomossa "}, } actual_tool = storage.indexer_configuration_get(tool) assert actual_tool expected_tool = tool.copy() del actual_tool["id"] assert expected_tool == actual_tool def test_indexer_configuration_metadata_get_missing_context( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data tool = { "tool_name": "swh-metadata-translator", "tool_version": "0.0.1", "tool_configuration": {"context": "unknown-context"}, } actual_tool = storage.indexer_configuration_get(tool) assert actual_tool is None def test_indexer_configuration_metadata_get( self, swh_indexer_storage_with_data: Tuple[IndexerStorageInterface, Any] ) -> None: storage, data = swh_indexer_storage_with_data tool = { "tool_name": "swh-metadata-translator", "tool_version": "0.0.1", "tool_configuration": {"type": "local", "context": "NpmMapping"}, } storage.indexer_configuration_add([tool]) actual_tool = storage.indexer_configuration_get(tool) assert actual_tool expected_tool = tool.copy() expected_tool["id"] = actual_tool["id"] assert expected_tool == actual_tool diff --git a/swh/indexer/tests/test_ctags.py b/swh/indexer/tests/test_ctags.py index 4bb2139..bfb7f53 100644 --- a/swh/indexer/tests/test_ctags.py +++ b/swh/indexer/tests/test_ctags.py @@ -1,164 +1,162 @@ # Copyright (C) 2017-2018 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 json import unittest from unittest.mock import patch import pytest import swh.indexer.ctags from swh.indexer.ctags import CtagsIndexer, run_ctags +from swh.indexer.storage.model import ContentCtagsRow from swh.indexer.tests.utils import ( BASE_TEST_CONFIG, OBJ_STORAGE_DATA, SHA1_TO_CTAGS, CommonContentIndexerTest, fill_obj_storage, fill_storage, filter_dict, ) from swh.model.hashutil import hash_to_bytes class BasicTest(unittest.TestCase): @patch("swh.indexer.ctags.subprocess") def test_run_ctags(self, mock_subprocess): """Computing licenses from a raw content should return results """ output0 = """ {"name":"defun","kind":"function","line":1,"language":"scheme"} {"name":"name","kind":"symbol","line":5,"language":"else"}""" output1 = """ {"name":"let","kind":"var","line":10,"language":"something"}""" expected_result0 = [ {"name": "defun", "kind": "function", "line": 1, "lang": "scheme"}, {"name": "name", "kind": "symbol", "line": 5, "lang": "else"}, ] expected_result1 = [ {"name": "let", "kind": "var", "line": 10, "lang": "something"} ] for path, lang, intermediary_result, expected_result in [ (b"some/path", "lisp", output0, expected_result0), (b"some/path/2", "markdown", output1, expected_result1), ]: mock_subprocess.check_output.return_value = intermediary_result actual_result = list(run_ctags(path, lang=lang)) self.assertEqual(actual_result, expected_result) class InjectCtagsIndexer: """Override ctags computations. """ def compute_ctags(self, path, lang): """Inject fake ctags given path (sha1 identifier). """ return {"lang": lang, **SHA1_TO_CTAGS.get(path)} CONFIG = { **BASE_TEST_CONFIG, "tools": { "name": "universal-ctags", "version": "~git7859817b", "configuration": { "command_line": """ctags --fields=+lnz --sort=no """ """ --links=no """, "max_content_size": 1000, }, }, "languages": {"python": "python", "haskell": "haskell", "bar": "bar",}, "workdir": "/tmp", } class TestCtagsIndexer(CommonContentIndexerTest, unittest.TestCase): """Ctags indexer test scenarios: - Known sha1s in the input list have their data indexed - Unknown sha1 in the input list are not indexed """ def get_indexer_results(self, ids): yield from self.idx_storage.content_ctags_get(ids) def setUp(self): super().setUp() self.indexer = CtagsIndexer(config=CONFIG) self.indexer.catch_exceptions = False self.idx_storage = self.indexer.idx_storage fill_storage(self.indexer.storage) fill_obj_storage(self.indexer.objstorage) # Prepare test input self.id0 = "01c9379dfc33803963d07c1ccc748d3fe4c96bb5" self.id1 = "d4c647f0fc257591cc9ba1722484229780d1c607" self.id2 = "688a5ef812c53907562fe379d4b3851e69c7cb15" tool = {k.replace("tool_", ""): v for (k, v) in self.indexer.tool.items()} self.expected_results = [ - { - "id": hash_to_bytes(self.id0), - "tool": tool, - **SHA1_TO_CTAGS[self.id0][0], - }, - { - "id": hash_to_bytes(self.id1), - "tool": tool, - **SHA1_TO_CTAGS[self.id1][0], - }, - { - "id": hash_to_bytes(self.id2), - "tool": tool, - **SHA1_TO_CTAGS[self.id2][0], - }, + *[ + ContentCtagsRow(id=hash_to_bytes(self.id0), tool=tool, **kwargs,) + for kwargs in SHA1_TO_CTAGS[self.id0] + ], + *[ + ContentCtagsRow(id=hash_to_bytes(self.id1), tool=tool, **kwargs,) + for kwargs in SHA1_TO_CTAGS[self.id1] + ], + *[ + ContentCtagsRow(id=hash_to_bytes(self.id2), tool=tool, **kwargs,) + for kwargs in SHA1_TO_CTAGS[self.id2] + ], ] self._set_mocks() def _set_mocks(self): def find_ctags_for_content(raw_content): for (sha1, ctags) in SHA1_TO_CTAGS.items(): if OBJ_STORAGE_DATA[sha1] == raw_content: return ctags else: raise ValueError( ("%r not found in objstorage, can't mock its ctags.") % raw_content ) def fake_language(raw_content, *args, **kwargs): ctags = find_ctags_for_content(raw_content) return {"lang": ctags[0]["lang"]} self._real_compute_language = swh.indexer.ctags.compute_language swh.indexer.ctags.compute_language = fake_language def fake_check_output(cmd, *args, **kwargs): id_ = cmd[-1].split("/")[-1] return "\n".join( json.dumps({"language": ctag["lang"], **ctag}) for ctag in SHA1_TO_CTAGS[id_] ) self._real_check_output = swh.indexer.ctags.subprocess.check_output swh.indexer.ctags.subprocess.check_output = fake_check_output def tearDown(self): swh.indexer.ctags.compute_language = self._real_compute_language swh.indexer.ctags.subprocess.check_output = self._real_check_output super().tearDown() def test_ctags_w_no_tool(): with pytest.raises(ValueError): CtagsIndexer(config=filter_dict(CONFIG, "tools"))