diff --git a/swh/indexer/indexer.py b/swh/indexer/indexer.py index 5d31a8e..2ace1d3 100644 --- a/swh/indexer/indexer.py +++ b/swh/indexer/indexer.py @@ -1,609 +1,613 @@ # Copyright (C) 2016-2020 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 abc from contextlib import contextmanager import logging import os import shutil import tempfile from typing import Any, Dict, Iterator, List, Optional, Set, Union from swh.core import utils from swh.core.config import load_from_envvar, merge_configs from swh.indexer.storage import INDEXER_CFG_KEY, PagedResult, Sha1, get_indexer_storage +from swh.indexer.storage.model import BaseRow from swh.model import hashutil from swh.model.model import Revision from swh.objstorage.exc import ObjNotFoundError from swh.objstorage.factory import get_objstorage from swh.scheduler import CONFIG as SWH_CONFIG from swh.storage import get_storage @contextmanager def write_to_temp(filename: str, data: bytes, working_directory: str) -> Iterator[str]: """Write the sha1's content in a temporary file. Args: filename: one of sha1's many filenames data: the sha1's content to write in temporary file working_directory: the directory into which the file is written Returns: The path to the temporary file created. That file is filled in with the raw content's data. """ os.makedirs(working_directory, exist_ok=True) temp_dir = tempfile.mkdtemp(dir=working_directory) content_path = os.path.join(temp_dir, filename) with open(content_path, "wb") as f: f.write(data) yield content_path shutil.rmtree(temp_dir) DEFAULT_CONFIG = { INDEXER_CFG_KEY: {"cls": "memory"}, "storage": {"cls": "memory"}, "objstorage": {"cls": "memory"}, } class BaseIndexer(metaclass=abc.ABCMeta): """Base class for indexers to inherit from. The main entry point is the :func:`run` function which is in charge of triggering the computations on the batch dict/ids received. Indexers can: - filter out ids whose data has already been indexed. - retrieve ids data from storage or objstorage - index this data depending on the object and store the result in storage. To implement a new object type indexer, inherit from the BaseIndexer and implement indexing: :meth:`~BaseIndexer.run`: object_ids are different depending on object. For example: sha1 for content, sha1_git for revision, directory, release, and id for origin To implement a new concrete indexer, inherit from the object level classes: :class:`ContentIndexer`, :class:`RevisionIndexer`, :class:`OriginIndexer`. Then you need to implement the following functions: :meth:`~BaseIndexer.filter`: filter out data already indexed (in storage). :meth:`~BaseIndexer.index_object`: compute index on id with data (retrieved from the storage or the objstorage by the id key) and return the resulting index computation. :meth:`~BaseIndexer.persist_index_computations`: persist the results of multiple index computations in the storage. The new indexer implementation can also override the following functions: :meth:`~BaseIndexer.prepare`: Configuration preparation for the indexer. When overriding, this must call the `super().prepare()` instruction. :meth:`~BaseIndexer.check`: Configuration check for the indexer. When overriding, this must call the `super().check()` instruction. :meth:`~BaseIndexer.register_tools`: This should return a dict of the tool(s) to use when indexing or filtering. """ - results: List[Dict] + results: List[Union[Dict, BaseRow]] USE_TOOLS = True catch_exceptions = True """Prevents exceptions in `index()` from raising too high. Set to False in tests to properly catch all exceptions.""" scheduler: Any def __init__(self, config=None, **kw) -> None: """Prepare and check that the indexer is ready to run. """ super().__init__() if config is not None: self.config = config elif SWH_CONFIG: self.config = SWH_CONFIG.copy() else: self.config = load_from_envvar() self.config = merge_configs(DEFAULT_CONFIG, self.config) self.prepare() self.check() self.log.debug("%s: config=%s", self, self.config) def prepare(self) -> None: """Prepare the indexer's needed runtime configuration. Without this step, the indexer cannot possibly run. """ config_storage = self.config.get("storage") if config_storage: self.storage = get_storage(**config_storage) objstorage = self.config["objstorage"] self.objstorage = get_objstorage(objstorage["cls"], objstorage["args"]) idx_storage = self.config[INDEXER_CFG_KEY] self.idx_storage = get_indexer_storage(**idx_storage) _log = logging.getLogger("requests.packages.urllib3.connectionpool") _log.setLevel(logging.WARN) self.log = logging.getLogger("swh.indexer") if self.USE_TOOLS: self.tools = list(self.register_tools(self.config.get("tools", []))) self.results = [] @property def tool(self) -> Dict: return self.tools[0] def check(self) -> None: """Check the indexer's configuration is ok before proceeding. If ok, does nothing. If not raise error. """ if self.USE_TOOLS and not self.tools: raise ValueError("Tools %s is unknown, cannot continue" % self.tools) def _prepare_tool(self, tool: Dict[str, Any]) -> Dict[str, Any]: """Prepare the tool dict to be compliant with the storage api. """ return {"tool_%s" % key: value for key, value in tool.items()} def register_tools( self, tools: Union[Dict[str, Any], List[Dict[str, Any]]] ) -> List[Dict[str, Any]]: """Permit to register tools to the storage. Add a sensible default which can be overridden if not sufficient. (For now, all indexers use only one tool) Expects the self.config['tools'] property to be set with one or more tools. Args: tools: Either a dict or a list of dict. Returns: list: List of dicts with additional id key. Raises: ValueError: if not a list nor a dict. """ if isinstance(tools, list): tools = list(map(self._prepare_tool, tools)) elif isinstance(tools, dict): tools = [self._prepare_tool(tools)] else: raise ValueError("Configuration tool(s) must be a dict or list!") if tools: return self.idx_storage.indexer_configuration_add(tools) else: return [] def index( self, id: Union[bytes, Dict, Revision], data: Optional[bytes] = None, **kwargs - ) -> Dict[str, Any]: + ) -> Union[Dict[str, Any], BaseRow]: """Index computation for the id and associated raw data. Args: id: identifier or Dict object data: id's data from storage or objstorage depending on object type Returns: dict: a dict that makes sense for the :meth:`.persist_index_computations` method. """ raise NotImplementedError() def filter(self, ids: List[bytes]) -> Iterator[bytes]: """Filter missing ids for that particular indexer. Args: ids: list of ids Yields: iterator of missing ids """ yield from ids @abc.abstractmethod def persist_index_computations(self, results, policy_update) -> Dict[str, int]: """Persist the computation resulting from the index. Args: results ([result]): List of results. One result is the result of the index function. policy_update ([str]): either 'update-dups' or 'ignore-dups' to respectively update duplicates or ignore them Returns: a summary dict of what has been inserted in the storage """ return {} class ContentIndexer(BaseIndexer): """A content indexer working on a list of ids directly. To work on indexer partition, use the :class:`ContentPartitionIndexer` instead. Note: :class:`ContentIndexer` is not an instantiable object. To use it, one should inherit from this class and override the methods mentioned in the :class:`BaseIndexer` class. """ def run( self, ids: Union[List[bytes], bytes, str], policy_update: str, **kwargs ) -> Dict: """Given a list of ids: - retrieve the content from the storage - execute the indexing computations - store the results (according to policy_update) Args: ids (Iterable[Union[bytes, str]]): sha1's identifier list policy_update (str): either 'update-dups' or 'ignore-dups' to respectively update duplicates or ignore them **kwargs: passed to the `index` method Returns: A summary Dict of the task's status """ sha1s = [ hashutil.hash_to_bytes(id_) if isinstance(id_, str) else id_ for id_ in ids ] results = [] summary: Dict = {"status": "uneventful"} try: for sha1 in sha1s: try: raw_content = self.objstorage.get(sha1) except ObjNotFoundError: self.log.warning( "Content %s not found in objstorage" % hashutil.hash_to_hex(sha1) ) continue res = self.index(sha1, raw_content, **kwargs) if res: # If no results, skip it results.append(res) summary["status"] = "eventful" summary = self.persist_index_computations(results, policy_update) self.results = results except Exception: if not self.catch_exceptions: raise self.log.exception("Problem when reading contents metadata.") summary["status"] = "failed" return summary class ContentPartitionIndexer(BaseIndexer): """A content partition indexer. This expects as input a partition_id and a nb_partitions. This will then index the contents within that partition. To work on a list of ids, use the :class:`ContentIndexer` instead. Note: :class:`ContentPartitionIndexer` is not an instantiable object. To use it, one should inherit from this class and override the methods mentioned in the :class:`BaseIndexer` class. """ @abc.abstractmethod def indexed_contents_in_partition( self, partition_id: int, nb_partitions: int, page_token: Optional[str] = None ) -> PagedResult[Sha1]: """Retrieve indexed contents within range [start, end]. Args: partition_id: Index of the partition to fetch nb_partitions: Total number of partitions to split into page_token: opaque token used for pagination Returns: PagedResult of Sha1. If next_page_token is None, there is no more data to fetch """ pass def _list_contents_to_index( self, partition_id: int, nb_partitions: int, indexed: Set[Sha1] ) -> Iterator[Sha1]: """Compute from storage the new contents to index in the partition_id . The already indexed contents are skipped. Args: partition_id: Index of the partition to fetch data from nb_partitions: Total number of partition indexed: Set of content already indexed. Yields: Sha1 id (bytes) of contents to index """ if not isinstance(partition_id, int) or not isinstance(nb_partitions, int): raise TypeError( f"identifiers must be int, not {partition_id!r} and {nb_partitions!r}." ) next_page_token = None while True: result = self.storage.content_get_partition( partition_id, nb_partitions, page_token=next_page_token ) contents = result.results for c in contents: _id = hashutil.hash_to_bytes(c.sha1) if _id in indexed: continue yield _id next_page_token = result.next_page_token if next_page_token is None: break def _index_contents( self, partition_id: int, nb_partitions: int, indexed: Set[Sha1], **kwargs: Any - ) -> Iterator[Dict]: + ) -> Iterator[Union[BaseRow, Dict]]: """Index the contents within the partition_id. Args: start: Starting bound from range identifier end: End range identifier indexed: Set of content already indexed. Yields: indexing result as dict to persist in the indexer backend """ for sha1 in self._list_contents_to_index(partition_id, nb_partitions, indexed): try: raw_content = self.objstorage.get(sha1) except ObjNotFoundError: self.log.warning(f"Content {sha1.hex()} not found in objstorage") continue res = self.index(sha1, raw_content, **kwargs) if res: - if not isinstance(res["id"], bytes): + # TODO: remove this check when all endpoints moved away from dicts. + if isinstance(res, dict) and not isinstance(res["id"], bytes): raise TypeError( "%r.index should return ids as bytes, not %r" % (self.__class__.__name__, res["id"]) ) yield res def _index_with_skipping_already_done( self, partition_id: int, nb_partitions: int - ) -> Iterator[Dict]: + ) -> Iterator[Union[BaseRow, Dict]]: """Index not already indexed contents within the partition partition_id Args: partition_id: Index of the partition to fetch nb_partitions: Total number of partitions to split into Yields: indexing result as dict to persist in the indexer backend """ next_page_token = None contents = set() while True: indexed_page = self.indexed_contents_in_partition( partition_id, nb_partitions, page_token=next_page_token ) for sha1 in indexed_page.results: contents.add(sha1) yield from self._index_contents(partition_id, nb_partitions, contents) next_page_token = indexed_page.next_page_token if next_page_token is None: break def run( self, partition_id: int, nb_partitions: int, skip_existing: bool = True, **kwargs, ) -> Dict: """Given a partition of content ids, index the contents within. Either the indexer is incremental (filter out existing computed data) or it computes everything from scratch. Args: partition_id: Index of the partition to fetch nb_partitions: Total number of partitions to split into skip_existing: Skip existing indexed data (default) or not **kwargs: passed to the `index` method Returns: dict with the indexing task status """ summary: Dict[str, Any] = {"status": "uneventful"} count = 0 try: if skip_existing: gen = self._index_with_skipping_already_done( partition_id, nb_partitions ) else: gen = self._index_contents(partition_id, nb_partitions, indexed=set([])) count_object_added_key: Optional[str] = None for contents in utils.grouper(gen, n=self.config["write_batch_size"]): res = self.persist_index_computations( list(contents), policy_update="update-dups" ) if not count_object_added_key: count_object_added_key = list(res.keys())[0] count += res[count_object_added_key] if count > 0: summary["status"] = "eventful" except Exception: if not self.catch_exceptions: raise self.log.exception("Problem when computing metadata.") summary["status"] = "failed" if count > 0 and count_object_added_key: summary[count_object_added_key] = count return summary class OriginIndexer(BaseIndexer): """An object type indexer, inherits from the :class:`BaseIndexer` and implements Origin indexing using the run method Note: the :class:`OriginIndexer` is not an instantiable object. To use it in another context one should inherit from this class and override the methods mentioned in the :class:`BaseIndexer` class. """ def run( self, origin_urls: List[str], policy_update: str = "update-dups", **kwargs ) -> Dict: """Given a list of origin urls: - retrieve origins from storage - execute the indexing computations - store the results (according to policy_update) Args: origin_urls: list of origin urls. policy_update: either 'update-dups' or 'ignore-dups' to respectively update duplicates (default) or ignore them **kwargs: passed to the `index` method """ summary: Dict[str, Any] = {"status": "uneventful"} try: results = self.index_list(origin_urls, **kwargs) except Exception: if not self.catch_exceptions: raise summary["status"] = "failed" return summary summary_persist = self.persist_index_computations(results, policy_update) self.results = results if summary_persist: for value in summary_persist.values(): if value > 0: summary["status"] = "eventful" summary.update(summary_persist) return summary - def index_list(self, origins: List[Any], **kwargs: Any) -> List[Dict]: + def index_list( + self, origins: List[Any], **kwargs: Any + ) -> List[Union[Dict, BaseRow]]: results = [] for origin in origins: try: res = self.index(origin, **kwargs) if res: # If no results, skip it results.append(res) except Exception: self.log.exception("Problem when processing origin %s", origin) raise return results class RevisionIndexer(BaseIndexer): """An object type indexer, inherits from the :class:`BaseIndexer` and implements Revision indexing using the run method Note: the :class:`RevisionIndexer` is not an instantiable object. To use it in another context one should inherit from this class and override the methods mentioned in the :class:`BaseIndexer` class. """ def run(self, ids: Union[str, bytes], policy_update: str) -> Dict: """Given a list of sha1_gits: - retrieve revisions from storage - execute the indexing computations - store the results (according to policy_update) Args: ids: sha1_git's identifier list policy_update: either 'update-dups' or 'ignore-dups' to respectively update duplicates or ignore them """ summary: Dict[str, Any] = {"status": "uneventful"} results = [] revision_ids = [ hashutil.hash_to_bytes(id_) if isinstance(id_, str) else id_ for id_ in ids ] for rev in self.storage.revision_get(revision_ids): if not rev: self.log.warning( "Revisions %s not found in storage" % list(map(hashutil.hash_to_hex, ids)) ) continue try: res = self.index(rev) if res: # If no results, skip it results.append(res) except Exception: if not self.catch_exceptions: raise self.log.exception("Problem when processing revision") summary["status"] = "failed" return summary summary_persist = self.persist_index_computations(results, policy_update) if summary_persist: for value in summary_persist.values(): if value > 0: summary["status"] = "eventful" summary.update(summary_persist) self.results = results return summary diff --git a/swh/indexer/metadata.py b/swh/indexer/metadata.py index 5863918..0cc443a 100644 --- a/swh/indexer/metadata.py +++ b/swh/indexer/metadata.py @@ -1,381 +1,382 @@ # Copyright (C) 2017-2020 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 copy import deepcopy from typing import Any, Callable, Dict, Iterator, List, Tuple from swh.core.config import merge_configs from swh.core.utils import grouper from swh.indexer.codemeta import merge_documents from swh.indexer.indexer import ContentIndexer, OriginIndexer, RevisionIndexer from swh.indexer.metadata_detector import detect_metadata from swh.indexer.metadata_dictionary import MAPPINGS from swh.indexer.origin_head import OriginHeadIndexer from swh.indexer.storage import INDEXER_CFG_KEY from swh.model import hashutil REVISION_GET_BATCH_SIZE = 10 ORIGIN_GET_BATCH_SIZE = 10 def call_with_batches( f: Callable[[List[Dict[str, Any]]], Dict["str", Any]], args: List[Dict[str, str]], batch_size: int, ) -> Iterator[str]: """Calls a function with batches of args, and concatenates the results. """ groups = grouper(args, batch_size) for group in groups: yield from f(list(group)) class ContentMetadataIndexer(ContentIndexer): """Content-level indexer This indexer is in charge of: - filtering out content already indexed in content_metadata - reading content from objstorage with the content's id sha1 - computing metadata by given context - using the metadata_dictionary as the 'swh-metadata-translator' tool - store result in content_metadata table """ def filter(self, ids): """Filter out known sha1s and return only missing ones. """ yield from self.idx_storage.content_metadata_missing( ({"id": sha1, "indexer_configuration_id": self.tool["id"],} for sha1 in ids) ) def index(self, id, data, log_suffix="unknown revision"): """Index sha1s' content and store result. Args: id (bytes): content's identifier data (bytes): raw content in bytes Returns: dict: dictionary representing a content_metadata. If the translation wasn't successful the metadata keys will be returned as None """ result = { "id": id, "indexer_configuration_id": self.tool["id"], "metadata": None, } try: mapping_name = self.tool["tool_configuration"]["context"] log_suffix += ", content_id=%s" % hashutil.hash_to_hex(id) result["metadata"] = MAPPINGS[mapping_name](log_suffix).translate(data) except Exception: self.log.exception( "Problem during metadata translation " "for content %s" % hashutil.hash_to_hex(id) ) if result["metadata"] is None: return None return result def persist_index_computations( self, results: List[Dict], policy_update: str ) -> Dict[str, int]: """Persist the results in storage. Args: results: list of content_metadata, dict with the following keys: - id (bytes): content's identifier (sha1) - metadata (jsonb): detected metadata policy_update: either 'update-dups' or 'ignore-dups' to respectively update duplicates or ignore them """ return self.idx_storage.content_metadata_add( results, conflict_update=(policy_update == "update-dups") ) DEFAULT_CONFIG: Dict[str, Any] = { "tools": { "name": "swh-metadata-detector", "version": "0.0.2", "configuration": {}, }, } class RevisionMetadataIndexer(RevisionIndexer): """Revision-level indexer This indexer is in charge of: - filtering revisions already indexed in revision_intrinsic_metadata table with defined computation tool - retrieve all entry_files in root directory - use metadata_detector for file_names containing metadata - compute metadata translation if necessary and possible (depends on tool) - send sha1s to content indexing if possible - store the results for revision """ def __init__(self, *args, **kwargs): super().__init__(*args, **kwargs) self.config = merge_configs(DEFAULT_CONFIG, self.config) def filter(self, sha1_gits): """Filter out known sha1s and return only missing ones. """ yield from self.idx_storage.revision_intrinsic_metadata_missing( ( {"id": sha1_git, "indexer_configuration_id": self.tool["id"],} for sha1_git in sha1_gits ) ) def index(self, rev): """Index rev by processing it and organizing result. use metadata_detector to iterate on filenames - if one filename detected -> sends file to content indexer - if multiple file detected -> translation needed at revision level Args: rev: revision model object from storage Returns: dict: dictionary representing a revision_intrinsic_metadata, with keys: - id (str): rev's identifier (sha1_git) - indexer_configuration_id (bytes): tool used - metadata: dict of retrieved metadata """ result = { "id": rev.id, "indexer_configuration_id": self.tool["id"], "mappings": None, "metadata": None, } try: root_dir = rev.directory dir_ls = list(self.storage.directory_ls(root_dir, recursive=False)) if [entry["type"] for entry in dir_ls] == ["dir"]: # If the root is just a single directory, recurse into it # eg. PyPI packages, GNU tarballs subdir = dir_ls[0]["target"] dir_ls = self.storage.directory_ls(subdir, recursive=False) files = [entry for entry in dir_ls if entry["type"] == "file"] detected_files = detect_metadata(files) (mappings, metadata) = self.translate_revision_intrinsic_metadata( detected_files, log_suffix="revision=%s" % hashutil.hash_to_hex(rev.id), ) result["mappings"] = mappings result["metadata"] = metadata except Exception as e: self.log.exception("Problem when indexing rev: %r", e) return result def persist_index_computations( self, results: List[Dict], 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) - mimetype (bytes): mimetype in bytes - encoding (bytes): encoding in bytes policy_update: either 'update-dups' or 'ignore-dups' to respectively update duplicates or ignore them """ # TODO: add functions in storage to keep data in # revision_intrinsic_metadata return self.idx_storage.revision_intrinsic_metadata_add( results, conflict_update=(policy_update == "update-dups") ) def translate_revision_intrinsic_metadata( self, detected_files: Dict[str, List[Any]], log_suffix: str ) -> Tuple[List[Any], List[Any]]: """ Determine plan of action to translate metadata when containing one or multiple detected files: Args: detected_files: dictionary mapping context names (e.g., "npm", "authors") to list of sha1 Returns: (List[str], dict): list of mappings used and dict with translated metadata according to the CodeMeta vocabulary """ used_mappings = [MAPPINGS[context].name for context in detected_files] metadata = [] tool = { "name": "swh-metadata-translator", "version": "0.0.2", "configuration": {}, } # TODO: iterate on each context, on each file # -> get raw_contents # -> translate each content config = {k: self.config[k] for k in [INDEXER_CFG_KEY, "objstorage", "storage"]} config["tools"] = [tool] for context in detected_files.keys(): cfg = deepcopy(config) cfg["tools"][0]["configuration"]["context"] = context c_metadata_indexer = ContentMetadataIndexer(config=cfg) # sha1s that are in content_metadata table sha1s_in_storage = [] metadata_generator = self.idx_storage.content_metadata_get( detected_files[context] ) for c in metadata_generator: # extracting metadata sha1 = c["id"] sha1s_in_storage.append(sha1) local_metadata = c["metadata"] # local metadata is aggregated if local_metadata: metadata.append(local_metadata) sha1s_filtered = [ item for item in detected_files[context] if item not in sha1s_in_storage ] if sha1s_filtered: # content indexing try: c_metadata_indexer.run( sha1s_filtered, policy_update="ignore-dups", log_suffix=log_suffix, ) # on the fly possibility: for result in c_metadata_indexer.results: + assert isinstance(result, dict) # TODO: remove this local_metadata = result["metadata"] metadata.append(local_metadata) except Exception: self.log.exception("Exception while indexing metadata on contents") metadata = merge_documents(metadata) return (used_mappings, metadata) class OriginMetadataIndexer(OriginIndexer): USE_TOOLS = False def __init__(self, config=None, **kwargs) -> None: super().__init__(config=config, **kwargs) self.origin_head_indexer = OriginHeadIndexer(config=config) self.revision_metadata_indexer = RevisionMetadataIndexer(config=config) def index_list(self, origin_urls, **kwargs): head_rev_ids = [] origins_with_head = [] origins = list( call_with_batches( self.storage.origin_get, origin_urls, ORIGIN_GET_BATCH_SIZE, ) ) for origin in origins: if origin is None: continue head_result = self.origin_head_indexer.index(origin.url) if head_result: origins_with_head.append(origin) head_rev_ids.append(head_result["revision_id"]) head_revs = list( call_with_batches( self.storage.revision_get, head_rev_ids, REVISION_GET_BATCH_SIZE ) ) assert len(head_revs) == len(head_rev_ids) results = [] for (origin, rev) in zip(origins_with_head, head_revs): if not rev: self.log.warning("Missing head revision of origin %r", origin.url) continue rev_metadata = self.revision_metadata_indexer.index(rev) orig_metadata = { "from_revision": rev_metadata["id"], "id": origin.url, "metadata": rev_metadata["metadata"], "mappings": rev_metadata["mappings"], "indexer_configuration_id": rev_metadata["indexer_configuration_id"], } results.append((orig_metadata, rev_metadata)) return results def persist_index_computations( self, results: List[Dict], policy_update: str ) -> Dict[str, int]: conflict_update = policy_update == "update-dups" # Deduplicate revisions rev_metadata: List[Any] = [] orig_metadata: List[Any] = [] revs_to_delete: List[Any] = [] origs_to_delete: List[Any] = [] summary: Dict = {} for (orig_item, rev_item) in results: assert rev_item["metadata"] == orig_item["metadata"] if not rev_item["metadata"] or rev_item["metadata"].keys() <= {"@context"}: # If we didn't find any metadata, don't store a DB record # (and delete existing ones, if any) if rev_item not in revs_to_delete: revs_to_delete.append(rev_item) if orig_item not in origs_to_delete: origs_to_delete.append(orig_item) else: if rev_item not in rev_metadata: rev_metadata.append(rev_item) if orig_item not in orig_metadata: orig_metadata.append(orig_item) if rev_metadata: summary_rev = self.idx_storage.revision_intrinsic_metadata_add( rev_metadata, conflict_update=conflict_update ) summary.update(summary_rev) if orig_metadata: summary_ori = self.idx_storage.origin_intrinsic_metadata_add( orig_metadata, conflict_update=conflict_update ) summary.update(summary_ori) # revs_to_delete should always be empty unless we changed a mapping # to detect less files or less content. # However, origs_to_delete may be empty whenever an upstream deletes # a metadata file. if origs_to_delete: summary_ori = self.idx_storage.origin_intrinsic_metadata_delete( origs_to_delete ) summary.update(summary_ori) if revs_to_delete: summary_rev = self.idx_storage.revision_intrinsic_metadata_delete( revs_to_delete ) summary.update(summary_rev) return summary diff --git a/swh/indexer/mimetype.py b/swh/indexer/mimetype.py index d2cc437..92cf01e 100644 --- a/swh/indexer/mimetype.py +++ b/swh/indexer/mimetype.py @@ -1,164 +1,167 @@ # Copyright (C) 2016-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information from typing import Any, Dict, List, Optional, Union import magic from swh.core.config import merge_configs from swh.indexer.storage.interface import PagedResult, Sha1 +from swh.indexer.storage.model import ContentMimetypeRow from swh.model.model import Revision from .indexer import ContentIndexer, ContentPartitionIndexer if not hasattr(magic.Magic, "from_buffer"): raise ImportError( 'Expected "import magic" to import python-magic, but file_magic ' "was imported instead." ) -def compute_mimetype_encoding(raw_content: bytes) -> Dict[str, bytes]: +def compute_mimetype_encoding(raw_content: bytes) -> Dict[str, str]: """Determine mimetype and encoding from the raw content. Args: raw_content: content's raw data Returns: dict: mimetype and encoding key and corresponding values. """ m = magic.Magic(mime=True, mime_encoding=True) res = m.from_buffer(raw_content) try: mimetype, encoding = res.split("; charset=") except ValueError: mimetype, encoding = res, "" return { "mimetype": mimetype, "encoding": encoding, } DEFAULT_CONFIG: Dict[str, Any] = { "tools": { "name": "file", "version": "1:5.30-1+deb9u1", "configuration": {"type": "library", "debian-package": "python3-magic"}, }, "write_batch_size": 1000, } class MixinMimetypeIndexer: """Mixin mimetype indexer. See :class:`MimetypeIndexer` and :class:`MimetypePartitionIndexer` """ tool: Any idx_storage: Any def __init__(self, *args, **kwargs): super().__init__(*args, **kwargs) self.config = merge_configs(DEFAULT_CONFIG, self.config) def index( self, id: Union[bytes, Dict, Revision], data: Optional[bytes] = None, **kwargs - ) -> Dict[str, Any]: + ) -> ContentMimetypeRow: """Index sha1s' content and store result. Args: id: content's identifier data: raw content in bytes Returns: dict: content's mimetype; dict keys being - id: content's identifier (sha1) - mimetype: mimetype in bytes - encoding: encoding in bytes """ + assert isinstance(id, bytes) assert data is not None properties = compute_mimetype_encoding(data) - assert isinstance(id, bytes) - properties.update( - {"id": id, "indexer_configuration_id": self.tool["id"],} + return ContentMimetypeRow( + id=id, + indexer_configuration_id=self.tool["id"], + mimetype=properties["mimetype"], + encoding=properties["encoding"], ) - return properties def persist_index_computations( self, results: List[Dict], policy_update: str ) -> Dict[str, int]: """Persist the results in storage. Args: results: list of content's mimetype dicts (see :meth:`.index`) policy_update: either 'update-dups' or 'ignore-dups' to respectively update duplicates or ignore them """ return self.idx_storage.content_mimetype_add( results, conflict_update=(policy_update == "update-dups") ) class MimetypeIndexer(MixinMimetypeIndexer, ContentIndexer): """Mimetype Indexer working on list of content identifiers. It: - (optionally) filters out content already indexed (cf. :meth:`.filter`) - reads content from objstorage per the content's id (sha1) - computes {mimetype, encoding} from that content - stores result in storage """ def filter(self, ids): """Filter out known sha1s and return only missing ones. """ yield from self.idx_storage.content_mimetype_missing( ({"id": sha1, "indexer_configuration_id": self.tool["id"],} for sha1 in ids) ) class MimetypePartitionIndexer(MixinMimetypeIndexer, ContentPartitionIndexer): """Mimetype Range Indexer working on range of content identifiers. It: - (optionally) filters out content already indexed (cf :meth:`.indexed_contents_in_partition`) - reads content from objstorage per the content's id (sha1) - computes {mimetype, encoding} from that content - stores result in storage """ def indexed_contents_in_partition( self, partition_id: int, nb_partitions: int, page_token: Optional[str] = None, ) -> PagedResult[Sha1]: """Retrieve indexed content ids within partition_id. Args: partition_id: Index of the partition to fetch nb_partitions: Total number of partitions to split into page_token: opaque token used for pagination Returns: PagedResult of Sha1. If next_page_token is None, there is no more data to fetch """ return self.idx_storage.content_mimetype_get_partition( self.tool["id"], partition_id, nb_partitions, page_token=page_token ) diff --git a/swh/indexer/storage/__init__.py b/swh/indexer/storage/__init__.py index 4d7c3ad..2f4c285 100644 --- a/swh/indexer/storage/__init__.py +++ b/swh/indexer/storage/__init__.py @@ -1,648 +1,658 @@ # 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, defaultdict import itertools import json -from typing import Dict, List, Optional +from typing import Dict, Iterable, Iterator, List, Optional, Tuple import psycopg2 import psycopg2.pool from swh.core.db.common import db_transaction, db_transaction_generator 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_generator() - def content_mimetype_missing(self, mimetypes, db=None, cur=None): + def content_mimetype_missing( + self, mimetypes: Iterable[Dict], db=None, cur=None + ) -> Iterator[Tuple[Sha1, int]]: for obj in db.content_mimetype_missing_from_list(mimetypes, cur): yield obj[0] @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[Dict], conflict_update: bool = False, db=None, cur=None + self, + mimetypes: List[ContentMimetypeRow], + conflict_update: bool = False, + db=None, + cur=None, ) -> Dict[str, int]: """Add mimetypes to the storage (if conflict_update is True, this will override existing data if any). Returns: A dict with the number of new elements added to the storage. """ - check_id_duplicates(map(ContentMimetypeRow.from_dict, mimetypes)) - mimetypes.sort(key=lambda m: m["id"]) + check_id_duplicates(mimetypes) + mimetypes.sort(key=lambda m: m.id) db.mktemp_content_mimetype(cur) db.copy_to( - mimetypes, + [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_generator() - def content_mimetype_get(self, ids, db=None, cur=None): + def content_mimetype_get( + self, ids: Iterable[Sha1], db=None, cur=None + ) -> Iterator[ContentMimetypeRow]: for c in db.content_mimetype_get_from_list(ids, cur): - yield converters.db_to_mimetype(dict(zip(db.content_mimetype_cols, c))) + yield ContentMimetypeRow.from_dict( + converters.db_to_mimetype(dict(zip(db.content_mimetype_cols, c))) + ) @timed @db_transaction_generator() def content_language_missing(self, languages, db=None, cur=None): for obj in db.content_language_missing_from_list(languages, cur): yield obj[0] @timed @db_transaction_generator() def content_language_get(self, ids, db=None, cur=None): for c in db.content_language_get_from_list(ids, cur): yield converters.db_to_language(dict(zip(db.content_language_cols, c))) @timed @process_metrics @db_transaction() def content_language_add( self, languages: List[Dict], conflict_update: bool = False, db=None, cur=None ) -> Dict[str, int]: check_id_duplicates(map(ContentLanguageRow.from_dict, 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": "unknown" if not lang["lang"] else lang["lang"], "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_generator() def content_ctags_missing(self, ctags, db=None, cur=None): for obj in db.content_ctags_missing_from_list(ctags, cur): yield obj[0] @timed @db_transaction_generator() def content_ctags_get(self, ids, db=None, cur=None): for c in db.content_ctags_get_from_list(ids, cur): yield converters.db_to_ctags(dict(zip(db.content_ctags_cols, c))) @timed @process_metrics @db_transaction() def content_ctags_add( self, ctags: List[Dict], 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"]) db.mktemp_content_ctags(cur) db.copy_to( rows, 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_generator() def content_ctags_search( self, expression, limit=10, last_sha1=None, db=None, cur=None ): for obj in db.content_ctags_search(expression, last_sha1, limit, cur=cur): yield converters.db_to_ctags(dict(zip(db.content_ctags_cols, obj))) @timed @db_transaction_generator() def content_fossology_license_get(self, ids, db=None, cur=None): d = defaultdict(list) for c in db.content_fossology_license_get_from_list(ids, cur): license = dict(zip(db.content_fossology_license_cols, c)) id_ = license["id"] d[id_].append(converters.db_to_fossology_license(license)) for id_, facts in d.items(): yield {id_: facts} @timed @process_metrics @db_transaction() def content_fossology_license_add( self, licenses: List[Dict], conflict_update: bool = False, db=None, cur=None ) -> Dict[str, int]: rows = list( itertools.chain.from_iterable( map(converters.fossology_license_to_db, licenses) ) ) check_id_duplicates(map(ContentLicenseRow.from_dict, rows)) licenses.sort(key=lambda m: m["id"]) db.mktemp_content_fossology_license(cur) db.copy_to( rows, 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_generator() def content_metadata_missing(self, metadata, db=None, cur=None): for obj in db.content_metadata_missing_from_list(metadata, cur): yield obj[0] @timed @db_transaction_generator() def content_metadata_get(self, ids, db=None, cur=None): for c in db.content_metadata_get_from_list(ids, cur): yield converters.db_to_metadata(dict(zip(db.content_metadata_cols, c))) @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_generator() def revision_intrinsic_metadata_missing(self, metadata, db=None, cur=None): for obj in db.revision_intrinsic_metadata_missing_from_list(metadata, cur): yield obj[0] @timed @db_transaction_generator() def revision_intrinsic_metadata_get(self, ids, db=None, cur=None): for c in db.revision_intrinsic_metadata_get_from_list(ids, cur): yield converters.db_to_metadata( dict(zip(db.revision_intrinsic_metadata_cols, c)) ) @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_generator() def origin_intrinsic_metadata_get(self, ids, db=None, cur=None): for c in db.origin_intrinsic_metadata_get_from_list(ids, cur): yield converters.db_to_metadata( dict(zip(db.origin_intrinsic_metadata_cols, c)) ) @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_generator() def origin_intrinsic_metadata_search_fulltext( self, conjunction, limit=100, db=None, cur=None ): for c in db.origin_intrinsic_metadata_search_fulltext( conjunction, limit=limit, cur=cur ): yield converters.db_to_metadata( dict(zip(db.origin_intrinsic_metadata_cols, c)) ) @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_generator() 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) count = 0 for line in tools: yield dict(zip(db.indexer_configuration_cols, line)) count += 1 send_metric( "indexer_configuration:add", count, method_name="indexer_configuration_add" ) @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/api/client.py b/swh/indexer/storage/api/client.py index 1abebc3..a873c6b 100644 --- a/swh/indexer/storage/api/client.py +++ b/swh/indexer/storage/api/client.py @@ -1,21 +1,24 @@ # Copyright (C) 2015-2018 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 from swh.core.api import RPCClient from swh.indexer.storage.exc import ( DuplicateId, IndexerStorageAPIError, IndexerStorageArgumentException, ) from ..interface import IndexerStorageInterface +from .serializers import DECODERS, ENCODERS class RemoteStorage(RPCClient): """Proxy to a remote storage API""" backend_class = IndexerStorageInterface api_exception = IndexerStorageAPIError reraise_exceptions = [IndexerStorageArgumentException, DuplicateId] + extra_type_decoders = DECODERS + extra_type_encoders = ENCODERS diff --git a/swh/indexer/storage/api/serializers.py b/swh/indexer/storage/api/serializers.py new file mode 100644 index 0000000..92ca1b0 --- /dev/null +++ b/swh/indexer/storage/api/serializers.py @@ -0,0 +1,26 @@ +# 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 + +"""Decoder and encoders for swh-model objects.""" + +from typing import Callable, Dict, List, Tuple + +import swh.indexer.storage.model as idx_model + + +def _encode_model_object(obj): + d = obj.to_dict() + d["__type__"] = type(obj).__name__ + return d + + +ENCODERS: List[Tuple[type, str, Callable]] = [ + (idx_model.BaseRow, "idx_model", _encode_model_object), +] + + +DECODERS: Dict[str, Callable] = { + "idx_model": lambda d: getattr(idx_model, d.pop("__type__")).from_dict(d), +} diff --git a/swh/indexer/storage/api/server.py b/swh/indexer/storage/api/server.py index 04289c6..09979ca 100644 --- a/swh/indexer/storage/api/server.py +++ b/swh/indexer/storage/api/server.py @@ -1,111 +1,118 @@ # Copyright (C) 2015-2019 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import logging import os from swh.core import config from swh.core.api import RPCServerApp from swh.core.api import encode_data_server as encode_data from swh.core.api import error_handler from swh.indexer.storage import INDEXER_CFG_KEY, get_indexer_storage from swh.indexer.storage.exc import IndexerStorageArgumentException from swh.indexer.storage.interface import IndexerStorageInterface +from .serializers import DECODERS, ENCODERS + def get_storage(): global storage if not storage: storage = get_indexer_storage(**app.config[INDEXER_CFG_KEY]) return storage -app = RPCServerApp( +class IndexerStorageServerApp(RPCServerApp): + extra_type_decoders = DECODERS + extra_type_encoders = ENCODERS + + +app = IndexerStorageServerApp( __name__, backend_class=IndexerStorageInterface, backend_factory=get_storage ) storage = None @app.errorhandler(Exception) def my_error_handler(exception): return error_handler(exception, encode_data) @app.errorhandler(IndexerStorageArgumentException) def argument_error_handler(exception): return error_handler(exception, encode_data, status_code=400) @app.route("/") def index(): return "SWH Indexer Storage API server" api_cfg = None def load_and_check_config(config_file, type="local"): """Check the minimal configuration is set to run the api or raise an error explanation. Args: config_file (str): Path to the configuration file to load type (str): configuration type. For 'local' type, more checks are done. Raises: Error if the setup is not as expected Returns: configuration as a dict """ if not config_file: raise EnvironmentError("Configuration file must be defined") if not os.path.exists(config_file): raise FileNotFoundError("Configuration file %s does not exist" % (config_file,)) cfg = config.read(config_file) if "indexer_storage" not in cfg: raise KeyError("Missing '%indexer_storage' configuration") if type == "local": vcfg = cfg["indexer_storage"] cls = vcfg.get("cls") if cls != "local": raise ValueError( "The indexer_storage backend can only be started with a " "'local' configuration" ) args = vcfg["args"] if not args.get("db"): raise ValueError("Invalid configuration; missing 'db' config entry") return cfg def make_app_from_configfile(): """Run the WSGI app from the webserver, loading the configuration from a configuration file. SWH_CONFIG_FILENAME environment variable defines the configuration path to load. """ global api_cfg if not api_cfg: config_file = os.environ.get("SWH_CONFIG_FILENAME") api_cfg = load_and_check_config(config_file) app.config.update(api_cfg) handler = logging.StreamHandler() app.logger.addHandler(handler) return app if __name__ == "__main__": print("Deprecated. Use swh-indexer") diff --git a/swh/indexer/storage/db.py b/swh/indexer/storage/db.py index ef012dd..e6c3f13 100644 --- a/swh/indexer/storage/db.py +++ b/swh/indexer/storage/db.py @@ -1,558 +1,566 @@ # Copyright (C) 2015-2018 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 +from typing import Dict, Iterable, Iterator, List + from swh.core.db import BaseDb from swh.core.db.db_utils import execute_values_generator, stored_procedure from swh.model import hashutil +from .interface import Sha1 + class Db(BaseDb): """Proxy to the SWH Indexer DB, with wrappers around stored procedures """ content_mimetype_hash_keys = ["id", "indexer_configuration_id"] - def _missing_from_list(self, table, data, hash_keys, cur=None): + def _missing_from_list( + self, table: str, data: Iterable[Dict], hash_keys: List[str], cur=None + ): """Read from table the data with hash_keys that are missing. Args: - table (str): Table name (e.g content_mimetype, content_language, + table: Table name (e.g content_mimetype, content_language, etc...) - data (dict): Dict of data to read from - hash_keys ([str]): List of keys to read in the data dict. + data: Dict of data to read from + hash_keys: List of keys to read in the data dict. Yields: The data which is missing from the db. """ cur = self._cursor(cur) keys = ", ".join(hash_keys) equality = " AND ".join(("t.%s = c.%s" % (key, key)) for key in hash_keys) yield from execute_values_generator( cur, """ select %s from (values %%s) as t(%s) where not exists ( select 1 from %s c where %s ) """ % (keys, keys, table, equality), (tuple(m[k] for k in hash_keys) for m in data), ) - def content_mimetype_missing_from_list(self, mimetypes, cur=None): + def content_mimetype_missing_from_list( + self, mimetypes: Iterable[Dict], cur=None + ) -> Iterator[Sha1]: """List missing mimetypes. """ yield from self._missing_from_list( "content_mimetype", mimetypes, self.content_mimetype_hash_keys, cur=cur ) content_mimetype_cols = [ "id", "mimetype", "encoding", "tool_id", "tool_name", "tool_version", "tool_configuration", ] @stored_procedure("swh_mktemp_content_mimetype") def mktemp_content_mimetype(self, cur=None): pass def content_mimetype_add_from_temp(self, conflict_update, cur=None): cur = self._cursor(cur) cur.execute("select * from swh_content_mimetype_add(%s)", (conflict_update,)) return cur.fetchone()[0] def _convert_key(self, key, main_table="c"): """Convert keys according to specific use in the module. Args: key (str): Key expression to change according to the alias used in the query main_table (str): Alias to use for the main table. Default to c for content_{something}. Expected: Tables content_{something} being aliased as 'c' (something in {language, mimetype, ...}), table indexer_configuration being aliased as 'i'. """ if key == "id": return "%s.id" % main_table elif key == "tool_id": return "i.id as tool_id" elif key == "licenses": return ( """ array(select name from fossology_license where id = ANY( array_agg(%s.license_id))) as licenses""" % main_table ) return key def _get_from_list(self, table, ids, cols, cur=None, id_col="id"): """Fetches entries from the `table` such that their `id` field (or whatever is given to `id_col`) is in `ids`. Returns the columns `cols`. The `cur` parameter is used to connect to the database. """ cur = self._cursor(cur) keys = map(self._convert_key, cols) query = """ select {keys} from (values %s) as t(id) inner join {table} c on c.{id_col}=t.id inner join indexer_configuration i on c.indexer_configuration_id=i.id; """.format( keys=", ".join(keys), id_col=id_col, table=table ) yield from execute_values_generator(cur, query, ((_id,) for _id in ids)) content_indexer_names = { "mimetype": "content_mimetype", "fossology_license": "content_fossology_license", } def content_get_range( self, content_type, start, end, indexer_configuration_id, limit=1000, with_textual_data=False, cur=None, ): """Retrieve contents with content_type, within range [start, end] bound by limit and associated to the given indexer configuration id. When asking to work on textual content, that filters on the mimetype table with any mimetype that is not binary. """ cur = self._cursor(cur) table = self.content_indexer_names[content_type] if with_textual_data: extra = """inner join content_mimetype cm on (t.id=cm.id and cm.mimetype like 'text/%%' and %(start)s <= cm.id and cm.id <= %(end)s) """ else: extra = "" query = f"""select t.id from {table} t {extra} where t.indexer_configuration_id=%(tool_id)s and %(start)s <= t.id and t.id <= %(end)s order by t.indexer_configuration_id, t.id limit %(limit)s""" cur.execute( query, { "start": start, "end": end, "tool_id": indexer_configuration_id, "limit": limit, }, ) yield from cur def content_mimetype_get_from_list(self, ids, cur=None): yield from self._get_from_list( "content_mimetype", ids, self.content_mimetype_cols, cur=cur ) content_language_hash_keys = ["id", "indexer_configuration_id"] def content_language_missing_from_list(self, languages, cur=None): """List missing languages. """ yield from self._missing_from_list( "content_language", languages, self.content_language_hash_keys, cur=cur ) content_language_cols = [ "id", "lang", "tool_id", "tool_name", "tool_version", "tool_configuration", ] @stored_procedure("swh_mktemp_content_language") def mktemp_content_language(self, cur=None): pass def content_language_add_from_temp(self, conflict_update, cur=None): cur = self._cursor(cur) cur.execute("select * from swh_content_language_add(%s)", (conflict_update,)) return cur.fetchone()[0] def content_language_get_from_list(self, ids, cur=None): yield from self._get_from_list( "content_language", ids, self.content_language_cols, cur=cur ) content_ctags_hash_keys = ["id", "indexer_configuration_id"] def content_ctags_missing_from_list(self, ctags, cur=None): """List missing ctags. """ yield from self._missing_from_list( "content_ctags", ctags, self.content_ctags_hash_keys, cur=cur ) content_ctags_cols = [ "id", "name", "kind", "line", "lang", "tool_id", "tool_name", "tool_version", "tool_configuration", ] @stored_procedure("swh_mktemp_content_ctags") def mktemp_content_ctags(self, cur=None): pass def content_ctags_add_from_temp(self, conflict_update, cur=None): cur = self._cursor(cur) cur.execute("select * from swh_content_ctags_add(%s)", (conflict_update,)) return cur.fetchone()[0] def content_ctags_get_from_list(self, ids, cur=None): cur = self._cursor(cur) keys = map(self._convert_key, self.content_ctags_cols) yield from execute_values_generator( cur, """ select %s from (values %%s) as t(id) inner join content_ctags c on c.id=t.id inner join indexer_configuration i on c.indexer_configuration_id=i.id order by line """ % ", ".join(keys), ((_id,) for _id in ids), ) def content_ctags_search(self, expression, last_sha1, limit, cur=None): cur = self._cursor(cur) if not last_sha1: query = """SELECT %s FROM swh_content_ctags_search(%%s, %%s)""" % ( ",".join(self.content_ctags_cols) ) cur.execute(query, (expression, limit)) else: if last_sha1 and isinstance(last_sha1, bytes): last_sha1 = "\\x%s" % hashutil.hash_to_hex(last_sha1) elif last_sha1: last_sha1 = "\\x%s" % last_sha1 query = """SELECT %s FROM swh_content_ctags_search(%%s, %%s, %%s)""" % ( ",".join(self.content_ctags_cols) ) cur.execute(query, (expression, limit, last_sha1)) yield from cur content_fossology_license_cols = [ "id", "tool_id", "tool_name", "tool_version", "tool_configuration", "licenses", ] @stored_procedure("swh_mktemp_content_fossology_license") def mktemp_content_fossology_license(self, cur=None): pass def content_fossology_license_add_from_temp(self, conflict_update, cur=None): """Add new licenses per content. """ cur = self._cursor(cur) cur.execute( "select * from swh_content_fossology_license_add(%s)", (conflict_update,) ) return cur.fetchone()[0] def content_fossology_license_get_from_list(self, ids, cur=None): """Retrieve licenses per id. """ cur = self._cursor(cur) keys = map(self._convert_key, self.content_fossology_license_cols) yield from execute_values_generator( cur, """ select %s from (values %%s) as t(id) inner join content_fossology_license c on t.id=c.id inner join indexer_configuration i on i.id=c.indexer_configuration_id group by c.id, i.id, i.tool_name, i.tool_version, i.tool_configuration; """ % ", ".join(keys), ((_id,) for _id in ids), ) content_metadata_hash_keys = ["id", "indexer_configuration_id"] def content_metadata_missing_from_list(self, metadata, cur=None): """List missing metadata. """ yield from self._missing_from_list( "content_metadata", metadata, self.content_metadata_hash_keys, cur=cur ) content_metadata_cols = [ "id", "metadata", "tool_id", "tool_name", "tool_version", "tool_configuration", ] @stored_procedure("swh_mktemp_content_metadata") def mktemp_content_metadata(self, cur=None): pass def content_metadata_add_from_temp(self, conflict_update, cur=None): cur = self._cursor(cur) cur.execute("select * from swh_content_metadata_add(%s)", (conflict_update,)) return cur.fetchone()[0] def content_metadata_get_from_list(self, ids, cur=None): yield from self._get_from_list( "content_metadata", ids, self.content_metadata_cols, cur=cur ) revision_intrinsic_metadata_hash_keys = ["id", "indexer_configuration_id"] def revision_intrinsic_metadata_missing_from_list(self, metadata, cur=None): """List missing metadata. """ yield from self._missing_from_list( "revision_intrinsic_metadata", metadata, self.revision_intrinsic_metadata_hash_keys, cur=cur, ) revision_intrinsic_metadata_cols = [ "id", "metadata", "mappings", "tool_id", "tool_name", "tool_version", "tool_configuration", ] @stored_procedure("swh_mktemp_revision_intrinsic_metadata") def mktemp_revision_intrinsic_metadata(self, cur=None): pass def revision_intrinsic_metadata_add_from_temp(self, conflict_update, cur=None): cur = self._cursor(cur) cur.execute( "select * from swh_revision_intrinsic_metadata_add(%s)", (conflict_update,) ) return cur.fetchone()[0] def revision_intrinsic_metadata_delete(self, entries, cur=None): cur = self._cursor(cur) cur.execute( "DELETE from revision_intrinsic_metadata " "WHERE (id, indexer_configuration_id) IN " " (VALUES %s) " "RETURNING id" % (", ".join("%s" for _ in entries)), tuple((e["id"], e["indexer_configuration_id"]) for e in entries), ) return len(cur.fetchall()) def revision_intrinsic_metadata_get_from_list(self, ids, cur=None): yield from self._get_from_list( "revision_intrinsic_metadata", ids, self.revision_intrinsic_metadata_cols, cur=cur, ) origin_intrinsic_metadata_cols = [ "id", "metadata", "from_revision", "mappings", "tool_id", "tool_name", "tool_version", "tool_configuration", ] origin_intrinsic_metadata_regconfig = "pg_catalog.simple" """The dictionary used to normalize 'metadata' and queries. 'pg_catalog.simple' provides no stopword, so it should be suitable for proper names and non-English content. When updating this value, make sure to add a new index on origin_intrinsic_metadata.metadata.""" @stored_procedure("swh_mktemp_origin_intrinsic_metadata") def mktemp_origin_intrinsic_metadata(self, cur=None): pass def origin_intrinsic_metadata_add_from_temp(self, conflict_update, cur=None): cur = self._cursor(cur) cur.execute( "select * from swh_origin_intrinsic_metadata_add(%s)", (conflict_update,) ) return cur.fetchone()[0] def origin_intrinsic_metadata_delete(self, entries, cur=None): cur = self._cursor(cur) cur.execute( "DELETE from origin_intrinsic_metadata " "WHERE (id, indexer_configuration_id) IN" " (VALUES %s) " "RETURNING id" % (", ".join("%s" for _ in entries)), tuple((e["id"], e["indexer_configuration_id"]) for e in entries), ) return len(cur.fetchall()) def origin_intrinsic_metadata_get_from_list(self, ids, cur=None): yield from self._get_from_list( "origin_intrinsic_metadata", ids, self.origin_intrinsic_metadata_cols, cur=cur, id_col="id", ) def origin_intrinsic_metadata_search_fulltext(self, terms, *, limit, cur): regconfig = self.origin_intrinsic_metadata_regconfig tsquery_template = " && ".join( "plainto_tsquery('%s', %%s)" % regconfig for _ in terms ) tsquery_args = [(term,) for term in terms] keys = ( self._convert_key(col, "oim") for col in self.origin_intrinsic_metadata_cols ) query = ( "SELECT {keys} FROM origin_intrinsic_metadata AS oim " "INNER JOIN indexer_configuration AS i " "ON oim.indexer_configuration_id=i.id " "JOIN LATERAL (SELECT {tsquery_template}) AS s(tsq) ON true " "WHERE oim.metadata_tsvector @@ tsq " "ORDER BY ts_rank(oim.metadata_tsvector, tsq, 1) DESC " "LIMIT %s;" ).format(keys=", ".join(keys), tsquery_template=tsquery_template) cur.execute(query, tsquery_args + [limit]) yield from cur def origin_intrinsic_metadata_search_by_producer( self, last, limit, ids_only, mappings, tool_ids, cur ): if ids_only: keys = "oim.id" else: keys = ", ".join( ( self._convert_key(col, "oim") for col in self.origin_intrinsic_metadata_cols ) ) query_parts = [ "SELECT %s" % keys, "FROM origin_intrinsic_metadata AS oim", "INNER JOIN indexer_configuration AS i", "ON oim.indexer_configuration_id=i.id", ] args = [] where = [] if last: where.append("oim.id > %s") args.append(last) if mappings is not None: where.append("oim.mappings && %s") args.append(mappings) if tool_ids is not None: where.append("oim.indexer_configuration_id = ANY(%s)") args.append(tool_ids) if where: query_parts.append("WHERE") query_parts.append(" AND ".join(where)) if limit: query_parts.append("LIMIT %s") args.append(limit) cur.execute(" ".join(query_parts), args) yield from cur indexer_configuration_cols = [ "id", "tool_name", "tool_version", "tool_configuration", ] @stored_procedure("swh_mktemp_indexer_configuration") def mktemp_indexer_configuration(self, cur=None): pass def indexer_configuration_add_from_temp(self, cur=None): cur = self._cursor(cur) cur.execute( "SELECT %s from swh_indexer_configuration_add()" % (",".join(self.indexer_configuration_cols),) ) yield from cur def indexer_configuration_get( self, tool_name, tool_version, tool_configuration, cur=None ): cur = self._cursor(cur) cur.execute( """select %s from indexer_configuration where tool_name=%%s and tool_version=%%s and tool_configuration=%%s""" % (",".join(self.indexer_configuration_cols)), (tool_name, tool_version, tool_configuration), ) return cur.fetchone() diff --git a/swh/indexer/storage/in_memory.py b/swh/indexer/storage/in_memory.py index b839215..245ab24 100644 --- a/swh/indexer/storage/in_memory.py +++ b/swh/indexer/storage/in_memory.py @@ -1,540 +1,539 @@ # 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, Iterator, 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 .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]) -> Iterator[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 """ 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()): yield id_ def get(self, ids: Iterable[Sha1]) -> Iterator[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`) """ for id_ in ids: for entry in self._data[id_].values(): entry = entry.copy() tool_id = entry.pop("indexer_configuration_id") yield self.row_class( id=id_, tool=_transform_tool(self._tools[tool_id]), **entry, ) def get_all(self) -> Iterator[TValue]: yield from 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): + def content_mimetype_missing( + self, mimetypes: Iterable[Dict] + ) -> Iterator[Tuple[Sha1, int]]: yield from 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[Dict], conflict_update: bool = False + self, mimetypes: List[ContentMimetypeRow], conflict_update: bool = False ) -> Dict[str, int]: - check_id_types(mimetypes) - added = self._mimetypes.add( - map(ContentMimetypeRow.from_dict, mimetypes), conflict_update - ) + added = self._mimetypes.add(mimetypes, conflict_update) return {"content_mimetype:add": added} - def content_mimetype_get(self, ids): - yield from (obj.to_dict() for obj in self._mimetypes.get(ids)) + def content_mimetype_get(self, ids: Iterable[Sha1]) -> Iterator[ContentMimetypeRow]: + yield from self._mimetypes.get(ids) def content_language_missing(self, languages): yield from self._languages.missing(languages) def content_language_get(self, ids): yield from (obj.to_dict() for obj in self._languages.get(ids)) def content_language_add( self, languages: List[Dict], conflict_update: bool = False ) -> Dict[str, int]: check_id_types(languages) added = self._languages.add( map(ContentLanguageRow.from_dict, languages), conflict_update ) return {"content_language:add": added} def content_ctags_missing(self, ctags): yield from self._content_ctags.missing(ctags) def content_ctags_get(self, ids): for item in self._content_ctags.get(ids): yield {"id": item.id, "tool": item.tool, **item.to_dict()} def content_ctags_add( self, ctags: List[Dict], 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, ) return {"content_ctags:add": added} def content_ctags_search(self, expression, limit=10, last_sha1=None): 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) 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: yield {"id": id_, "tool": tool, **ctag} def content_fossology_license_get(self, ids): # Rewrites the output of SubStorage.get from the old format to # the new one. SubStorage.get should be updated once all other # *_get methods use the new format. # See: https://forge.softwareheritage.org/T1433 for id_ in ids: items = {} for obj in self._licenses.get([id_]): items.setdefault(obj.tool["id"], (obj.tool, []))[1].append(obj.license) if items: yield { id_: [ {"tool": tool, "licenses": licenses} for (tool, licenses) in items.values() ] } def content_fossology_license_add( self, licenses: List[Dict], conflict_update: bool = False ) -> Dict[str, int]: check_id_types(licenses) added = self._licenses.add( map( ContentLicenseRow.from_dict, itertools.chain.from_iterable( map(converters.fossology_license_to_db, 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): yield from self._content_metadata.missing(metadata) def content_metadata_get(self, ids): yield from (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): yield from self._revision_intrinsic_metadata.missing(metadata) def revision_intrinsic_metadata_get(self, ids): yield from (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): yield from (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' ) for (rank_, result) in results[:limit]: yield result.to_dict() 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 c8bcca1..87f1399 100644 --- a/swh/indexer/storage/interface.py +++ b/swh/indexer/storage/interface.py @@ -1,616 +1,613 @@ # 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, List, Optional, TypeVar +from typing import Dict, Iterable, Iterator, 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 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): + def content_mimetype_missing( + self, mimetypes: Iterable[Dict] + ) -> Iterator[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 Yields: tuple (id, indexer_configuration_id): missing id """ ... @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[Dict], conflict_update: bool = False + self, mimetypes: List[ContentMimetypeRow], conflict_update: bool = False ) -> Dict[str, int]: """Add mimetypes not present in storage. Args: - mimetypes (iterable): dictionaries with keys: - - - **id** (bytes): sha1 identifier - - **mimetype** (bytes): raw content's mimetype - - **encoding** (bytes): raw content's encoding - - **indexer_configuration_id** (int): tool's id used to - compute the results - - **conflict_update** (bool): Flag to determine if we want to - overwrite (``True``) or skip duplicates (``False``, the - default) + mimetypes: mimetype rows to be added + 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): + def content_mimetype_get(self, ids: Iterable[Sha1]) -> Iterator[ContentMimetypeRow]: """Retrieve full content mimetype per ids. Args: ids (iterable): sha1 identifier Yields: mimetypes (iterable): dictionaries with keys: - **id** (bytes): sha1 identifier - **mimetype** (bytes): raw content's mimetype - **encoding** (bytes): raw content's encoding - **tool** (dict): Tool used to compute the language """ ... @remote_api_endpoint("content_language/missing") def content_language_missing(self, languages): """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 Yields: an iterable of missing id for the tuple (id, indexer_configuration_id) """ ... @remote_api_endpoint("content_language") def content_language_get(self, ids): """Retrieve full content language per ids. Args: ids (iterable): sha1 identifier Yields: languages (iterable): dictionaries with keys: - **id** (bytes): sha1 identifier - **lang** (bytes): raw content's language - **tool** (dict): Tool used to compute the language """ ... @remote_api_endpoint("content_language/add") def content_language_add( self, languages: List[Dict], conflict_update: bool = False ) -> Dict[str, int]: """Add languages not present in storage. Args: languages (iterable): dictionaries with keys: - **id** (bytes): sha1 - **lang** (bytes): language detected 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): """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, indexer_configuration_id) """ ... @remote_api_endpoint("content/ctags") def content_ctags_get(self, ids): """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 """ ... @remote_api_endpoint("content/ctags/add") def content_ctags_add( self, ctags: List[Dict], 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): """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: rows of ctags including id, name, lang, kind, line, etc... """ ... @remote_api_endpoint("content/fossology_license") def content_fossology_license_get(self, ids): """Retrieve licenses per id. Args: ids (iterable): sha1 checksums Yields: dict: ``{id: facts}`` where ``facts`` is a dict with the following keys: - **licenses** ([str]): associated licenses for that content - **tool** (dict): Tool used to compute the license """ ... @remote_api_endpoint("content/fossology_license/add") def content_fossology_license_add( self, licenses: List[Dict], conflict_update: bool = False ) -> Dict[str, int]: """Add licenses not present in storage. Args: licenses (iterable): dictionaries with keys: - **id**: sha1 - **licenses** ([bytes]): List of licenses associated to sha1 - **tool** (str): nomossa 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/tests/storage/conftest.py b/swh/indexer/tests/storage/conftest.py index 98c9340..6bcc638 100644 --- a/swh/indexer/tests/storage/conftest.py +++ b/swh/indexer/tests/storage/conftest.py @@ -1,74 +1,75 @@ # Copyright (C) 2015-2019 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information from os.path import join import pytest from swh.indexer.storage import get_indexer_storage +from swh.indexer.storage.model import ContentMimetypeRow from swh.model.hashutil import hash_to_bytes from swh.storage.pytest_plugin import postgresql_fact from . import SQL_DIR from .generate_data_test import FOSSOLOGY_LICENSES, MIMETYPE_OBJECTS, TOOLS DUMP_FILES = join(SQL_DIR, "*.sql") class DataObj(dict): def __getattr__(self, key): return self.__getitem__(key) def __setattr__(self, key, value): return self.__setitem__(key, value) @pytest.fixture def swh_indexer_storage_with_data(swh_indexer_storage): data = DataObj() tools = { tool["tool_name"]: { "id": tool["id"], "name": tool["tool_name"], "version": tool["tool_version"], "configuration": tool["tool_configuration"], } for tool in swh_indexer_storage.indexer_configuration_add(TOOLS) } data.tools = tools data.sha1_1 = hash_to_bytes("34973274ccef6ab4dfaaf86599792fa9c3fe4689") data.sha1_2 = hash_to_bytes("61c2b3a30496d329e21af70dd2d7e097046d07b7") data.revision_id_1 = hash_to_bytes("7026b7c1a2af56521e951c01ed20f255fa054238") data.revision_id_2 = hash_to_bytes("7026b7c1a2af56521e9587659012345678904321") data.revision_id_3 = hash_to_bytes("7026b7c1a2af56521e9587659012345678904320") data.origin_url_1 = "file:///dev/0/zero" # 44434341 data.origin_url_2 = "file:///dev/1/one" # 44434342 data.origin_url_3 = "file:///dev/2/two" # 54974445 data.mimetypes = [ - {**mimetype_obj, "indexer_configuration_id": tools["file"]["id"]} + ContentMimetypeRow(indexer_configuration_id=tools["file"]["id"], **mimetype_obj) for mimetype_obj in MIMETYPE_OBJECTS ] swh_indexer_storage.content_mimetype_add(data.mimetypes) data.fossology_licenses = [ {**fossology_obj, "indexer_configuration_id": tools["nomos"]["id"]} for fossology_obj in FOSSOLOGY_LICENSES ] swh_indexer_storage._test_data = data return (swh_indexer_storage, data) swh_indexer_storage_postgresql = postgresql_fact( "postgresql_proc", dump_files=DUMP_FILES ) @pytest.fixture def swh_indexer_storage(swh_indexer_storage_postgresql): storage_config = { "cls": "local", "args": {"db": swh_indexer_storage_postgresql.dsn,}, } return get_indexer_storage(**storage_config) diff --git a/swh/indexer/tests/storage/test_storage.py b/swh/indexer/tests/storage/test_storage.py index f2bbf79..d2fb89b 100644 --- a/swh/indexer/tests/storage/test_storage.py +++ b/swh/indexer/tests/storage/test_storage.py @@ -1,1947 +1,1994 @@ # 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 Dict +from typing import Callable, Dict, Union import pytest from swh.indexer.storage.exc import DuplicateId, IndexerStorageArgumentException from swh.indexer.storage.interface import IndexerStorageInterface +from swh.indexer.storage.model import BaseRow, ContentMimetypeRow from swh.model.hashutil import hash_to_bytes def prepare_mimetypes_from(fossology_licenses): """Fossology license needs some consistent data in db to run. """ mimetypes = [] for c in fossology_licenses: mimetypes.append( - { - "id": c["id"], - "mimetype": "text/plain", # for filtering on textual data to work - "encoding": "utf-8", - "indexer_configuration_id": c["indexer_configuration_id"], - } + 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): assert swh_indexer_storage.check_config(check_write=True) assert swh_indexer_storage.check_config(check_write=False) def test_types(swh_indexer_storage): """Checks all methods of StorageInterface are implemented by this backend, and that they have the same signature.""" # Create an instance of the protocol (which cannot be instantiated # directly, so this creates a subclass, then instantiates it) interface = type("_", (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. """ + # For old endpoints (which use dicts), this is just the identity function. + # For new endpoints, it returns a BaseRow instance. + row_from_dict: Callable[[Dict], Union[Dict, BaseRow]] = ( + staticmethod(lambda x: x) # type: ignore + ) + + # Inverse function of row_from_dict + # TODO: remove this once all endpoints are migrated to rows + dict_from_row: Callable[[Union[Dict, BaseRow]], Dict] = ( + staticmethod(lambda x: x) # type: ignore + ) + def test_missing(self, swh_indexer_storage_with_data): 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")( [ - { - "id": data.sha1_2, - **self.example_data[0], - "indexer_configuration_id": tool_id, - } + 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): 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")([data_v1]) + 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 = [ - { - "id": data.sha1_2, - **self.example_data[0], - "tool": data.tools[self.tool_name], - } + 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")([data_v2]) + 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): 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")([data_v1]) + 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 = [{"id": data.sha1_2, **self.example_data[0], "tool": tool,}] + 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")([data_v2], conflict_update=True) + 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 = [{"id": data.sha1_2, **self.example_data[1], "tool": tool,}] + 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): 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 = [ - { - "id": hash_, - **self.example_data[0], - "indexer_configuration_id": tool["id"], - } + self.row_from_dict( + { + "id": hash_, + **self.example_data[0], + "indexer_configuration_id": tool["id"], + } + ) for hash_ in hashes ] data_v2 = [ - { - "id": hash_, - **self.example_data[1], - "indexer_configuration_id": tool["id"], - } + 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 = [ - {"id": hash_, **self.example_data[0], "tool": tool,} for hash_ in hashes + 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(): endpoint(storage, etype, "add")(data_v2a, conflict_update=True) def f2(): 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( - endpoint(storage, etype, "get")(hashes), key=lambda x: x["id"] + 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 + {"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): storage, data = swh_indexer_storage_with_data etype = self.endpoint_type tool = data.tools[self.tool_name] - data_rev1 = { - "id": data.revision_id_2, - **self.example_data[0], - "indexer_configuration_id": tool["id"], - } + data_rev1 = self.row_from_dict( + { + "id": data.revision_id_2, + **self.example_data[0], + "indexer_configuration_id": tool["id"], + } + ) - data_rev2 = { - "id": data.revision_id_2, - **self.example_data[1], - "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 = [ - {"id": data.revision_id_2, **self.example_data[0], "tool": tool,} + 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): 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"], - } + 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 = [{"id": data.sha1_2, **self.example_data[0], "tool": tool,}] + 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): """get_partition call with wrong limit input should fail""" storage = swh_indexer_storage indexer_configuration_id = None with pytest.raises( IndexerStorageArgumentException, match="limit should not be None" ): storage.content_mimetype_get_partition( indexer_configuration_id, 0, 3, limit=None ) def test_generate_content_mimetype_get_partition_no_limit( self, swh_indexer_storage_with_data ): """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"] + 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 ): """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"] + 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 ): """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"] + 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 ): """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"] + 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",}, ] 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",}, ] }, ] # 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): 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": [ {"name": "hello", "kind": "function", "line": 133, "lang": "Python",}, {"name": "counter", "kind": "variable", "line": 119, "lang": "Python",}, {"name": "hello", "kind": "variable", "line": 210, "lang": "Python",}, ], } ctag2 = { "id": data.sha1_2, "indexer_configuration_id": tool_id, "ctags": [ {"name": "hello", "kind": "variable", "line": 100, "lang": "C",}, {"name": "result", "kind": "variable", "line": 120, "lang": "C",}, ], } storage.content_ctags_add([ctag1, ctag2]) # 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", } ] # 2. when actual_ctags = list( storage.content_ctags_search("hello", limit=1, last_sha1=ctag1["id"]) ) # 2. then assert actual_ctags == [ { "id": ctag2["id"], "tool": tool, "name": "hello", "kind": "variable", "line": 100, "lang": "C", } ] # 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", }, ] # 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", } ] # 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", } ] def test_content_ctags_search_no_result(self, swh_indexer_storage): 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 ): 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",} ], } # given storage.content_ctags_add([ctag_v1]) storage.content_ctags_add([ctag_v1]) # 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 # given ctag_v2 = ctag_v1.copy() ctag_v2.update( { "ctags": [ {"name": "defn", "kind": "function", "line": 120, "lang": "Scheme",} ] } ) 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, }, ] actual_ctags = list(storage.content_ctags_get([data.sha1_2])) assert actual_ctags == expected_ctags def test_content_ctags_add__update_in_place(self, swh_indexer_storage_with_data): 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",} ], } # given storage.content_ctags_add([ctag_v1]) # 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 # 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", }, ] } ) storage.content_ctags_add([ctag_v2], 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 def test_add_empty(self, swh_indexer_storage_with_data): (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": [],}] ) 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): (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): 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 ): 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" def test_content_fossology_license_add__new_license_added( self, swh_indexer_storage_with_data ): storage, data = swh_indexer_storage_with_data # given tool = data.tools["nomos"] tool_id = tool["id"] license_v1 = { "id": data.sha1_1, "licenses": ["Apache-2.0"], "indexer_configuration_id": tool_id, } # given storage.content_fossology_license_add([license_v1]) # conflict does nothing storage.content_fossology_license_add([license_v1]) # when actual_licenses = list(storage.content_fossology_license_get([data.sha1_1])) # then expected_license = {data.sha1_1: [{"licenses": ["Apache-2.0"], "tool": tool,}]} assert actual_licenses == [expected_license] # given license_v2 = license_v1.copy() license_v2.update( {"licenses": ["BSD-2-Clause"],} ) storage.content_fossology_license_add([license_v2]) actual_licenses = list(storage.content_fossology_license_get([data.sha1_1])) expected_license = { data.sha1_1: [{"licenses": ["Apache-2.0", "BSD-2-Clause"], "tool": tool}] } # license did not change as the v2 was dropped. assert actual_licenses == [expected_license] def test_generate_content_fossology_license_get_partition_failure( self, swh_indexer_storage_with_data ): """get_partition call with wrong limit input should fail""" storage, data = swh_indexer_storage_with_data indexer_configuration_id = None with pytest.raises( IndexerStorageArgumentException, match="limit should not be None" ): storage.content_fossology_license_get_partition( indexer_configuration_id, 0, 3, limit=None, ) def test_generate_content_fossology_license_get_partition_no_limit( self, swh_indexer_storage_with_data ): """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(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 ): """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(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 ): """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(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 ): """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(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): (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"], "licenses": [], } ] ) 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): (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): 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): 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 ): 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 ): storage, data = swh_indexer_storage_with_data # given tool_id = data.tools["swh-metadata-detector"]["id"] metadata_v1 = { "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 ): storage, data = swh_indexer_storage_with_data # given tool_id = data.tools["swh-metadata-detector"]["id"] metadata_v1 = { "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 ): 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(): storage.origin_intrinsic_metadata_add(data_v2a, conflict_update=True) def f2(): 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 ): 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 ): 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 ): 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): 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 ): 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): 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): 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): 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): 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): 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 ): 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): 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_fossology_license.py b/swh/indexer/tests/test_fossology_license.py index 771c352..8a6246a 100644 --- a/swh/indexer/tests/test_fossology_license.py +++ b/swh/indexer/tests/test_fossology_license.py @@ -1,147 +1,163 @@ # 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 from typing import Any, Dict import unittest from unittest.mock import patch import pytest from swh.indexer import fossology_license from swh.indexer.fossology_license import ( FossologyLicenseIndexer, FossologyLicensePartitionIndexer, compute_license, ) +from swh.indexer.storage.model import ContentLicenseRow from swh.indexer.tests.utils import ( BASE_TEST_CONFIG, SHA1_TO_LICENSES, CommonContentIndexerPartitionTest, CommonContentIndexerTest, fill_obj_storage, fill_storage, filter_dict, ) class BasicTest(unittest.TestCase): @patch("swh.indexer.fossology_license.subprocess") def test_compute_license(self, mock_subprocess): """Computing licenses from a raw content should return results """ for path, intermediary_result, output in [ (b"some/path", None, []), (b"some/path/2", [], []), (b"other/path", " contains license(s) GPL,AGPL", ["GPL", "AGPL"]), ]: mock_subprocess.check_output.return_value = intermediary_result actual_result = compute_license(path) self.assertEqual(actual_result, {"licenses": output, "path": path,}) def mock_compute_license(path): """path is the content identifier """ if isinstance(id, bytes): path = path.decode("utf-8") # path is something like /tmp/tmpXXX/ so we keep only the sha1 part path = path.split("/")[-1] return {"licenses": SHA1_TO_LICENSES.get(path, [])} CONFIG = { **BASE_TEST_CONFIG, "workdir": "/tmp", "tools": { "name": "nomos", "version": "3.1.0rc2-31-ga2cbb8c", "configuration": {"command_line": "nomossa ",}, }, } # type: Dict[str, Any] RANGE_CONFIG = dict(list(CONFIG.items()) + [("write_batch_size", 100)]) class TestFossologyLicenseIndexer(CommonContentIndexerTest, unittest.TestCase): """Language 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_fossology_license_get(ids) def setUp(self): super().setUp() # replace actual license computation with a mock self.orig_compute_license = fossology_license.compute_license fossology_license.compute_license = mock_compute_license self.indexer = FossologyLicenseIndexer(CONFIG) self.indexer.catch_exceptions = False self.idx_storage = self.indexer.idx_storage fill_storage(self.indexer.storage) fill_obj_storage(self.indexer.objstorage) self.id0 = "01c9379dfc33803963d07c1ccc748d3fe4c96bb5" self.id1 = "688a5ef812c53907562fe379d4b3851e69c7cb15" self.id2 = "da39a3ee5e6b4b0d3255bfef95601890afd80709" # empty content tool = {k.replace("tool_", ""): v for (k, v) in self.indexer.tool.items()} # then self.expected_results = { self.id0: {"tool": tool, "licenses": SHA1_TO_LICENSES[self.id0],}, self.id1: {"tool": tool, "licenses": SHA1_TO_LICENSES[self.id1],}, self.id2: None, } def tearDown(self): super().tearDown() fossology_license.compute_license = self.orig_compute_license class TestFossologyLicensePartitionIndexer( CommonContentIndexerPartitionTest, unittest.TestCase ): """Range Fossology License Indexer tests. - new data within range are indexed - no data outside a range are indexed - with filtering existing indexed data prior to compute new index - without filtering existing indexed data prior to compute new index """ def setUp(self): super().setUp() # replace actual license computation with a mock self.orig_compute_license = fossology_license.compute_license fossology_license.compute_license = mock_compute_license self.indexer = FossologyLicensePartitionIndexer(config=RANGE_CONFIG) self.indexer.catch_exceptions = False fill_storage(self.indexer.storage) fill_obj_storage(self.indexer.objstorage) def tearDown(self): super().tearDown() fossology_license.compute_license = self.orig_compute_license + def assert_results_ok(self, partition_id, nb_partitions, actual_results): + # TODO: remove this method when fossology_license endpoints moved away + # from dicts. + actual_result_rows = [] + for res in actual_results: + for license in res["licenses"]: + actual_result_rows.append( + ContentLicenseRow( + id=res["id"], + indexer_configuration_id=res["indexer_configuration_id"], + license=license, + ) + ) + super().assert_results_ok(partition_id, nb_partitions, actual_result_rows) + def test_fossology_w_no_tool(): with pytest.raises(ValueError): FossologyLicenseIndexer(config=filter_dict(CONFIG, "tools")) def test_fossology_range_w_no_tool(): with pytest.raises(ValueError): FossologyLicensePartitionIndexer(config=filter_dict(RANGE_CONFIG, "tools")) diff --git a/swh/indexer/tests/test_mimetype.py b/swh/indexer/tests/test_mimetype.py index b401374..75b1f04 100644 --- a/swh/indexer/tests/test_mimetype.py +++ b/swh/indexer/tests/test_mimetype.py @@ -1,125 +1,127 @@ # Copyright (C) 2017-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information from typing import Any, Dict import unittest import pytest from swh.indexer.mimetype import ( MimetypeIndexer, MimetypePartitionIndexer, compute_mimetype_encoding, ) from swh.indexer.tests.utils import ( BASE_TEST_CONFIG, CommonContentIndexerPartitionTest, CommonContentIndexerTest, fill_obj_storage, fill_storage, filter_dict, ) def test_compute_mimetype_encoding(): """Compute mimetype encoding should return results""" for _input, _mimetype, _encoding in [ ("du français".encode(), "text/plain", "utf-8"), (b"def __init__(self):", "text/x-python", "us-ascii"), (b"\xff\xfe\x00\x00\x00\x00\xff\xfe\xff\xff", "application/octet-stream", ""), ]: actual_result = compute_mimetype_encoding(_input) assert actual_result == {"mimetype": _mimetype, "encoding": _encoding} CONFIG = { **BASE_TEST_CONFIG, "tools": { "name": "file", "version": "1:5.30-1+deb9u1", "configuration": {"type": "library", "debian-package": "python3-magic"}, }, } # type: Dict[str, Any] class TestMimetypeIndexer(CommonContentIndexerTest, unittest.TestCase): """Mimetype indexer test scenarios: - Known sha1s in the input list have their data indexed - Unknown sha1 in the input list are not indexed """ legacy_get_format = True def get_indexer_results(self, ids): - yield from self.idx_storage.content_mimetype_get(ids) + yield from (x.to_dict() for x in self.idx_storage.content_mimetype_get(ids)) def setUp(self): self.indexer = MimetypeIndexer(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) self.id0 = "01c9379dfc33803963d07c1ccc748d3fe4c96bb5" self.id1 = "688a5ef812c53907562fe379d4b3851e69c7cb15" self.id2 = "da39a3ee5e6b4b0d3255bfef95601890afd80709" tool = {k.replace("tool_", ""): v for (k, v) in self.indexer.tool.items()} self.expected_results = { self.id0: { "id": self.id0, "tool": tool, "mimetype": "text/plain", "encoding": "us-ascii", }, self.id1: { "id": self.id1, "tool": tool, "mimetype": "text/plain", "encoding": "us-ascii", }, self.id2: { "id": self.id2, "tool": tool, "mimetype": "application/x-empty", "encoding": "binary", }, } RANGE_CONFIG = dict(list(CONFIG.items()) + [("write_batch_size", 100)]) class TestMimetypePartitionIndexer( CommonContentIndexerPartitionTest, unittest.TestCase ): """Range Mimetype Indexer tests. - new data within range are indexed - no data outside a range are indexed - with filtering existing indexed data prior to compute new index - without filtering existing indexed data prior to compute new index """ + row_from_dict = staticmethod(lambda x: x) # type: ignore + def setUp(self): super().setUp() self.indexer = MimetypePartitionIndexer(config=RANGE_CONFIG) self.indexer.catch_exceptions = False fill_storage(self.indexer.storage) fill_obj_storage(self.indexer.objstorage) def test_mimetype_w_no_tool(): with pytest.raises(ValueError): MimetypeIndexer(config=filter_dict(CONFIG, "tools")) def test_mimetype_range_w_no_tool(): with pytest.raises(ValueError): MimetypePartitionIndexer(config=filter_dict(CONFIG, "tools")) diff --git a/swh/indexer/tests/utils.py b/swh/indexer/tests/utils.py index e13e2a5..df3d53d 100644 --- a/swh/indexer/tests/utils.py +++ b/swh/indexer/tests/utils.py @@ -1,764 +1,768 @@ # Copyright (C) 2017-2020 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 abc import functools -from typing import Any, Dict +from typing import Any, Callable, Dict, Union import unittest from hypothesis import strategies from swh.core.api.classes import stream_results from swh.indexer.storage import INDEXER_CFG_KEY +from swh.indexer.storage.model import BaseRow from swh.model import hashutil from swh.model.hashutil import hash_to_bytes from swh.model.model import ( Content, Directory, DirectoryEntry, Origin, OriginVisit, OriginVisitStatus, Person, Revision, RevisionType, Snapshot, SnapshotBranch, TargetType, Timestamp, TimestampWithTimezone, ) from swh.storage.utils import now BASE_TEST_CONFIG: Dict[str, Dict[str, Any]] = { "storage": {"cls": "memory"}, "objstorage": {"cls": "memory", "args": {},}, INDEXER_CFG_KEY: {"cls": "memory", "args": {},}, } ORIGINS = [ Origin(url="https://github.com/SoftwareHeritage/swh-storage"), Origin(url="rsync://ftp.gnu.org/gnu/3dldf"), Origin(url="https://forge.softwareheritage.org/source/jesuisgpl/"), Origin(url="https://pypi.org/project/limnoria/"), Origin(url="http://0-512-md.googlecode.com/svn/"), Origin(url="https://github.com/librariesio/yarn-parser"), Origin(url="https://github.com/librariesio/yarn-parser.git"), ] ORIGIN_VISITS = [ {"type": "git", "origin": ORIGINS[0].url}, {"type": "ftp", "origin": ORIGINS[1].url}, {"type": "deposit", "origin": ORIGINS[2].url}, {"type": "pypi", "origin": ORIGINS[3].url}, {"type": "svn", "origin": ORIGINS[4].url}, {"type": "git", "origin": ORIGINS[5].url}, {"type": "git", "origin": ORIGINS[6].url}, ] DIRECTORY = Directory( id=hash_to_bytes("34f335a750111ca0a8b64d8034faec9eedc396be"), entries=( DirectoryEntry( name=b"index.js", type="file", target=hash_to_bytes("01c9379dfc33803963d07c1ccc748d3fe4c96bb5"), perms=0o100644, ), DirectoryEntry( name=b"package.json", type="file", target=hash_to_bytes("26a9f72a7c87cc9205725cfd879f514ff4f3d8d5"), perms=0o100644, ), DirectoryEntry( name=b".github", type="dir", target=Directory(entries=()).id, perms=0o040000, ), ), ) DIRECTORY2 = Directory( id=b"\xf8zz\xa1\x12`<1$\xfav\xf9\x01\xfd5\x85F`\xf2\xb6", entries=( DirectoryEntry( name=b"package.json", type="file", target=hash_to_bytes("f5305243b3ce7ef8dc864ebc73794da304025beb"), perms=0o100644, ), ), ) REVISION = Revision( id=hash_to_bytes("c6201cb1b9b9df9a7542f9665c3b5dfab85e9775"), message=b"Improve search functionality", author=Person( name=b"Andrew Nesbitt", fullname=b"Andrew Nesbitt ", email=b"andrewnez@gmail.com", ), committer=Person( name=b"Andrew Nesbitt", fullname=b"Andrew Nesbitt ", email=b"andrewnez@gmail.com", ), committer_date=TimestampWithTimezone( timestamp=Timestamp(seconds=1380883849, microseconds=0,), offset=120, negative_utc=False, ), type=RevisionType.GIT, synthetic=False, date=TimestampWithTimezone( timestamp=Timestamp(seconds=1487596456, microseconds=0,), offset=0, negative_utc=False, ), directory=DIRECTORY2.id, parents=(), ) REVISIONS = [REVISION] SNAPSHOTS = [ Snapshot( id=hash_to_bytes("a50fde72265343b7d28cecf6db20d98a81d21965"), branches={ b"refs/heads/add-revision-origin-cache": SnapshotBranch( target=b'L[\xce\x1c\x88\x8eF\t\xf1"\x19\x1e\xfb\xc0s\xe7/\xe9l\x1e', target_type=TargetType.REVISION, ), b"refs/head/master": SnapshotBranch( target=b"8K\x12\x00d\x03\xcc\xe4]bS\xe3\x8f{\xd7}\xac\xefrm", target_type=TargetType.REVISION, ), b"HEAD": SnapshotBranch( target=b"refs/head/master", target_type=TargetType.ALIAS ), b"refs/tags/v0.0.103": SnapshotBranch( target=b'\xb6"Im{\xfdLb\xb0\x94N\xea\x96m\x13x\x88+\x0f\xdd', target_type=TargetType.RELEASE, ), }, ), Snapshot( id=hash_to_bytes("2c67f69a416bca4e1f3fcd848c588fab88ad0642"), branches={ b"3DLDF-1.1.4.tar.gz": SnapshotBranch( target=b'dJ\xfb\x1c\x91\xf4\x82B%]6\xa2\x90|\xd3\xfc"G\x99\x11', target_type=TargetType.REVISION, ), b"3DLDF-2.0.2.tar.gz": SnapshotBranch( target=b"\xb6\x0e\xe7\x9e9\xac\xaa\x19\x9e=\xd1\xc5\x00\\\xc6\xfc\xe0\xa6\xb4V", # noqa target_type=TargetType.REVISION, ), b"3DLDF-2.0.3-examples.tar.gz": SnapshotBranch( target=b"!H\x19\xc0\xee\x82-\x12F1\xbd\x97\xfe\xadZ\x80\x80\xc1\x83\xff", # noqa target_type=TargetType.REVISION, ), b"3DLDF-2.0.3.tar.gz": SnapshotBranch( target=b"\x8e\xa9\x8e/\xea}\x9feF\xf4\x9f\xfd\xee\xcc\x1a\xb4`\x8c\x8by", # noqa target_type=TargetType.REVISION, ), b"3DLDF-2.0.tar.gz": SnapshotBranch( target=b"F6*\xff(?\x19a\xef\xb6\xc2\x1fv$S\xe3G\xd3\xd1m", target_type=TargetType.REVISION, ), }, ), Snapshot( id=hash_to_bytes("68c0d26104d47e278dd6be07ed61fafb561d0d20"), branches={ b"master": SnapshotBranch( target=b"\xe7n\xa4\x9c\x9f\xfb\xb7\xf76\x11\x08{\xa6\xe9\x99\xb1\x9e]q\xeb", # noqa target_type=TargetType.REVISION, ) }, ), Snapshot( id=hash_to_bytes("f255245269e15fc99d284affd79f766668de0b67"), branches={ b"HEAD": SnapshotBranch( target=b"releases/2018.09.09", target_type=TargetType.ALIAS ), b"releases/2018.09.01": SnapshotBranch( target=b"<\xee1(\xe8\x8d_\xc1\xc9\xa6rT\xf1\x1d\xbb\xdfF\xfdw\xcf", target_type=TargetType.REVISION, ), b"releases/2018.09.09": SnapshotBranch( target=b"\x83\xb9\xb6\xc7\x05\xb1%\xd0\xfem\xd8kA\x10\x9d\xc5\xfa2\xf8t", # noqa target_type=TargetType.REVISION, ), }, ), Snapshot( id=hash_to_bytes("a1a28c0ab387a8f9e0618cb705eab81fc448f473"), branches={ b"master": SnapshotBranch( target=b"\xe4?r\xe1,\x88\xab\xec\xe7\x9a\x87\xb8\xc9\xad#.\x1bw=\x18", target_type=TargetType.REVISION, ) }, ), Snapshot( id=hash_to_bytes("bb4fd3a836930ce629d912864319637040ff3040"), branches={ b"HEAD": SnapshotBranch( target=REVISION.id, target_type=TargetType.REVISION, ) }, ), Snapshot( id=hash_to_bytes("bb4fd3a836930ce629d912864319637040ff3040"), branches={ b"HEAD": SnapshotBranch( target=REVISION.id, target_type=TargetType.REVISION, ) }, ), ] SHA1_TO_LICENSES = { "01c9379dfc33803963d07c1ccc748d3fe4c96bb5": ["GPL"], "02fb2c89e14f7fab46701478c83779c7beb7b069": ["Apache2.0"], "103bc087db1d26afc3a0283f38663d081e9b01e6": ["MIT"], "688a5ef812c53907562fe379d4b3851e69c7cb15": ["AGPL"], "da39a3ee5e6b4b0d3255bfef95601890afd80709": [], } SHA1_TO_CTAGS = { "01c9379dfc33803963d07c1ccc748d3fe4c96bb5": [ {"name": "foo", "kind": "str", "line": 10, "lang": "bar",} ], "d4c647f0fc257591cc9ba1722484229780d1c607": [ {"name": "let", "kind": "int", "line": 100, "lang": "haskell",} ], "688a5ef812c53907562fe379d4b3851e69c7cb15": [ {"name": "symbol", "kind": "float", "line": 99, "lang": "python",} ], } OBJ_STORAGE_DATA = { "01c9379dfc33803963d07c1ccc748d3fe4c96bb5": b"this is some text", "688a5ef812c53907562fe379d4b3851e69c7cb15": b"another text", "8986af901dd2043044ce8f0d8fc039153641cf17": b"yet another text", "02fb2c89e14f7fab46701478c83779c7beb7b069": b""" import unittest import logging from swh.indexer.mimetype import MimetypeIndexer from swh.indexer.tests.test_utils import MockObjStorage class MockStorage(): def content_mimetype_add(self, mimetypes): self.state = mimetypes self.conflict_update = conflict_update def indexer_configuration_add(self, tools): return [{ 'id': 10, }] """, "103bc087db1d26afc3a0283f38663d081e9b01e6": b""" #ifndef __AVL__ #define __AVL__ typedef struct _avl_tree avl_tree; typedef struct _data_t { int content; } data_t; """, "93666f74f1cf635c8c8ac118879da6ec5623c410": b""" (should 'pygments (recognize 'lisp 'easily)) """, "26a9f72a7c87cc9205725cfd879f514ff4f3d8d5": b""" { "name": "test_metadata", "version": "0.0.1", "description": "Simple package.json test for indexer", "repository": { "type": "git", "url": "https://github.com/moranegg/metadata_test" } } """, "d4c647f0fc257591cc9ba1722484229780d1c607": b""" { "version": "5.0.3", "name": "npm", "description": "a package manager for JavaScript", "keywords": [ "install", "modules", "package manager", "package.json" ], "preferGlobal": true, "config": { "publishtest": false }, "homepage": "https://docs.npmjs.com/", "author": "Isaac Z. Schlueter (http://blog.izs.me)", "repository": { "type": "git", "url": "https://github.com/npm/npm" }, "bugs": { "url": "https://github.com/npm/npm/issues" }, "dependencies": { "JSONStream": "~1.3.1", "abbrev": "~1.1.0", "ansi-regex": "~2.1.1", "ansicolors": "~0.3.2", "ansistyles": "~0.1.3" }, "devDependencies": { "tacks": "~1.2.6", "tap": "~10.3.2" }, "license": "Artistic-2.0" } """, "a7ab314d8a11d2c93e3dcf528ca294e7b431c449": b""" """, "da39a3ee5e6b4b0d3255bfef95601890afd80709": b"", # was 626364 / b'bcd' "e3e40fee6ff8a52f06c3b428bfe7c0ed2ef56e92": b"unimportant content for bcd", # was 636465 / b'cde' now yarn-parser package.json "f5305243b3ce7ef8dc864ebc73794da304025beb": b""" { "name": "yarn-parser", "version": "1.0.0", "description": "Tiny web service for parsing yarn.lock files", "main": "index.js", "scripts": { "start": "node index.js", "test": "mocha" }, "engines": { "node": "9.8.0" }, "repository": { "type": "git", "url": "git+https://github.com/librariesio/yarn-parser.git" }, "keywords": [ "yarn", "parse", "lock", "dependencies" ], "author": "Andrew Nesbitt", "license": "AGPL-3.0", "bugs": { "url": "https://github.com/librariesio/yarn-parser/issues" }, "homepage": "https://github.com/librariesio/yarn-parser#readme", "dependencies": { "@yarnpkg/lockfile": "^1.0.0", "body-parser": "^1.15.2", "express": "^4.14.0" }, "devDependencies": { "chai": "^4.1.2", "mocha": "^5.2.0", "request": "^2.87.0", "test": "^0.6.0" } } """, } YARN_PARSER_METADATA = { "@context": "https://doi.org/10.5063/schema/codemeta-2.0", "url": "https://github.com/librariesio/yarn-parser#readme", "codeRepository": "git+git+https://github.com/librariesio/yarn-parser.git", "author": [{"type": "Person", "name": "Andrew Nesbitt"}], "license": "https://spdx.org/licenses/AGPL-3.0", "version": "1.0.0", "description": "Tiny web service for parsing yarn.lock files", "issueTracker": "https://github.com/librariesio/yarn-parser/issues", "name": "yarn-parser", "keywords": ["yarn", "parse", "lock", "dependencies"], "type": "SoftwareSourceCode", } json_dict_keys = strategies.one_of( strategies.characters(), strategies.just("type"), strategies.just("url"), strategies.just("name"), strategies.just("email"), strategies.just("@id"), strategies.just("@context"), strategies.just("repository"), strategies.just("license"), strategies.just("repositories"), strategies.just("licenses"), ) """Hypothesis strategy that generates strings, with an emphasis on those that are often used as dictionary keys in metadata files.""" generic_json_document = strategies.recursive( strategies.none() | strategies.booleans() | strategies.floats() | strategies.characters(), lambda children: ( strategies.lists(children, min_size=1) | strategies.dictionaries(json_dict_keys, children, min_size=1) ), ) """Hypothesis strategy that generates possible values for values of JSON metadata files.""" def json_document_strategy(keys=None): """Generates an hypothesis strategy that generates metadata files for a JSON-based format that uses the given keys.""" if keys is None: keys = strategies.characters() else: keys = strategies.one_of(map(strategies.just, keys)) return strategies.dictionaries(keys, generic_json_document, min_size=1) def _tree_to_xml(root, xmlns, data): def encode(s): "Skips unpaired surrogates generated by json_document_strategy" return s.encode("utf8", "replace") def to_xml(data, indent=b" "): if data is None: return b"" elif isinstance(data, (bool, str, int, float)): return indent + encode(str(data)) elif isinstance(data, list): return b"\n".join(to_xml(v, indent=indent) for v in data) elif isinstance(data, dict): lines = [] for (key, value) in data.items(): lines.append(indent + encode("<{}>".format(key))) lines.append(to_xml(value, indent=indent + b" ")) lines.append(indent + encode("".format(key))) return b"\n".join(lines) else: raise TypeError(data) return b"\n".join( [ '<{} xmlns="{}">'.format(root, xmlns).encode(), to_xml(data), "".format(root).encode(), ] ) class TreeToXmlTest(unittest.TestCase): def test_leaves(self): self.assertEqual( _tree_to_xml("root", "http://example.com", None), b'\n\n', ) self.assertEqual( _tree_to_xml("root", "http://example.com", True), b'\n True\n', ) self.assertEqual( _tree_to_xml("root", "http://example.com", "abc"), b'\n abc\n', ) self.assertEqual( _tree_to_xml("root", "http://example.com", 42), b'\n 42\n', ) self.assertEqual( _tree_to_xml("root", "http://example.com", 3.14), b'\n 3.14\n', ) def test_dict(self): self.assertIn( _tree_to_xml("root", "http://example.com", {"foo": "bar", "baz": "qux"}), [ b'\n' b" \n bar\n \n" b" \n qux\n \n" b"", b'\n' b" \n qux\n \n" b" \n bar\n \n" b"", ], ) def test_list(self): self.assertEqual( _tree_to_xml( "root", "http://example.com", [{"foo": "bar"}, {"foo": "baz"},] ), b'\n' b" \n bar\n \n" b" \n baz\n \n" b"", ) def xml_document_strategy(keys, root, xmlns): """Generates an hypothesis strategy that generates metadata files for an XML format that uses the given keys.""" return strategies.builds( functools.partial(_tree_to_xml, root, xmlns), json_document_strategy(keys) ) def filter_dict(d, keys): "return a copy of the dict with keys deleted" if not isinstance(keys, (list, tuple)): keys = (keys,) return dict((k, v) for (k, v) in d.items() if k not in keys) def fill_obj_storage(obj_storage): """Add some content in an object storage.""" for (obj_id, content) in OBJ_STORAGE_DATA.items(): obj_storage.add(content, obj_id=hash_to_bytes(obj_id)) def fill_storage(storage): storage.origin_add(ORIGINS) storage.directory_add([DIRECTORY, DIRECTORY2]) storage.revision_add(REVISIONS) storage.snapshot_add(SNAPSHOTS) for visit, snapshot in zip(ORIGIN_VISITS, SNAPSHOTS): assert snapshot.id is not None visit = storage.origin_visit_add( [OriginVisit(origin=visit["origin"], date=now(), type=visit["type"])] )[0] visit_status = OriginVisitStatus( origin=visit.origin, visit=visit.visit, date=now(), status="full", snapshot=snapshot.id, ) storage.origin_visit_status_add([visit_status]) contents = [] for (obj_id, content) in OBJ_STORAGE_DATA.items(): content_hashes = hashutil.MultiHash.from_data(content).digest() contents.append( Content( data=content, length=len(content), status="visible", sha1=hash_to_bytes(obj_id), sha1_git=hash_to_bytes(obj_id), sha256=content_hashes["sha256"], blake2s256=content_hashes["blake2s256"], ) ) storage.content_add(contents) class CommonContentIndexerTest(metaclass=abc.ABCMeta): legacy_get_format = False """True if and only if the tested indexer uses the legacy format. see: https://forge.softwareheritage.org/T1433 """ def get_indexer_results(self, ids): """Override this for indexers that don't have a mock storage.""" return self.indexer.idx_storage.state def assert_legacy_results_ok(self, sha1s, expected_results=None): # XXX old format, remove this when all endpoints are # updated to the new one # see: https://forge.softwareheritage.org/T1433 sha1s = [ sha1 if isinstance(sha1, bytes) else hash_to_bytes(sha1) for sha1 in sha1s ] actual_results = list(self.get_indexer_results(sha1s)) if expected_results is None: expected_results = self.expected_results self.assertEqual( len(expected_results), len(actual_results), (expected_results, actual_results), ) for indexed_data in actual_results: _id = indexed_data["id"] expected_data = expected_results[hashutil.hash_to_hex(_id)].copy() expected_data["id"] = _id self.assertEqual(indexed_data, expected_data) def assert_results_ok(self, sha1s, expected_results=None): if self.legacy_get_format: self.assert_legacy_results_ok(sha1s, expected_results) return sha1s = [ sha1 if isinstance(sha1, bytes) else hash_to_bytes(sha1) for sha1 in sha1s ] actual_results = list(self.get_indexer_results(sha1s)) if expected_results is None: expected_results = self.expected_results self.assertEqual( sum(res is not None for res in expected_results.values()), sum(sum(map(len, res.values())) for res in actual_results), (expected_results, actual_results), ) for indexed_data in actual_results: (_id, indexed_data) = list(indexed_data.items())[0] if expected_results.get(hashutil.hash_to_hex(_id)) is None: self.assertEqual(indexed_data, []) else: expected_data = expected_results[hashutil.hash_to_hex(_id)].copy() expected_data = [expected_data] self.assertEqual(indexed_data, expected_data) def test_index(self): """Known sha1 have their data indexed """ sha1s = [self.id0, self.id1, self.id2] # when self.indexer.run(sha1s, policy_update="update-dups") self.assert_results_ok(sha1s) # 2nd pass self.indexer.run(sha1s, policy_update="ignore-dups") self.assert_results_ok(sha1s) def test_index_one_unknown_sha1(self): """Unknown sha1 are not indexed""" sha1s = [ self.id1, "799a5ef812c53907562fe379d4b3851e69c7cb15", # unknown "800a5ef812c53907562fe379d4b3851e69c7cb15", ] # unknown # when self.indexer.run(sha1s, policy_update="update-dups") # then expected_results = { k: v for k, v in self.expected_results.items() if k in sha1s } self.assert_results_ok(sha1s, expected_results) class CommonContentIndexerPartitionTest: """Allows to factorize tests on range indexer. """ + # TODO: remove this when all endpoints moved away from dicts + row_from_dict: Callable[[Union[Dict, BaseRow]], BaseRow] + def setUp(self): self.contents = sorted(OBJ_STORAGE_DATA) def assert_results_ok(self, partition_id, nb_partitions, actual_results): expected_ids = [ c.sha1 for c in stream_results( self.indexer.storage.content_get_partition, partition_id=partition_id, nb_partitions=nb_partitions, ) ] actual_results = list(actual_results) for indexed_data in actual_results: - _id = indexed_data["id"] - assert isinstance(_id, bytes) + _id = indexed_data.id assert _id in expected_ids - _tool_id = indexed_data["indexer_configuration_id"] + _tool_id = indexed_data.indexer_configuration_id assert _tool_id == self.indexer.tool["id"] def test__index_contents(self): """Indexing contents without existing data results in indexed data """ partition_id = 0 nb_partitions = 4 actual_results = list( self.indexer._index_contents(partition_id, nb_partitions, indexed={}) ) self.assert_results_ok(partition_id, nb_partitions, actual_results) def test__index_contents_with_indexed_data(self): """Indexing contents with existing data results in less indexed data """ partition_id = 3 nb_partitions = 4 # first pass actual_results = list( - self.indexer._index_contents(partition_id, nb_partitions, indexed={}) + self.indexer._index_contents(partition_id, nb_partitions, indexed={}), ) self.assert_results_ok(partition_id, nb_partitions, actual_results) - indexed_ids = set(res["id"] for res in actual_results) + # TODO: unconditionally use res.id when all endpoints moved away from dicts + indexed_ids = {getattr(res, "id", None) or res["id"] for res in actual_results} actual_results = list( self.indexer._index_contents( partition_id, nb_partitions, indexed=indexed_ids ) ) # already indexed, so nothing new assert actual_results == [] def test_generate_content_get(self): """Optimal indexing should result in indexed data """ partition_id = 0 nb_partitions = 1 actual_results = self.indexer.run( partition_id, nb_partitions, skip_existing=False ) assert actual_results["status"] == "eventful", actual_results def test_generate_content_get_no_result(self): """No result indexed returns False""" actual_results = self.indexer.run(1, 2 ** 512, incremental=False) assert actual_results == {"status": "uneventful"}