diff --git a/PKG-INFO b/PKG-INFO index 06fbd34..bbc9a86 100644 --- a/PKG-INFO +++ b/PKG-INFO @@ -1,71 +1,71 @@ Metadata-Version: 2.1 Name: swh.indexer -Version: 0.1.1 +Version: 0.2.0 Summary: Software Heritage Content Indexer Home-page: https://forge.softwareheritage.org/diffusion/78/ Author: Software Heritage developers Author-email: swh-devel@inria.fr License: UNKNOWN Project-URL: Bug Reports, https://forge.softwareheritage.org/maniphest Project-URL: Funding, https://www.softwareheritage.org/donate Project-URL: Source, https://forge.softwareheritage.org/source/swh-indexer Project-URL: Documentation, https://docs.softwareheritage.org/devel/swh-indexer/ Description: swh-indexer ============ Tools to compute multiple indexes on SWH's raw contents: - content: - mimetype - ctags - language - fossology-license - metadata - revision: - metadata An indexer is in charge of: - looking up objects - extracting information from those objects - store those information in the swh-indexer db There are multiple indexers working on different object types: - content indexer: works with content sha1 hashes - revision indexer: works with revision sha1 hashes - origin indexer: works with origin identifiers Indexation procedure: - receive batch of ids - retrieve the associated data depending on object type - compute for that object some index - store the result to swh's storage Current content indexers: - mimetype (queue swh_indexer_content_mimetype): detect the encoding and mimetype - language (queue swh_indexer_content_language): detect the programming language - ctags (queue swh_indexer_content_ctags): compute tags information - fossology-license (queue swh_indexer_fossology_license): compute the license - metadata: translate file into translated_metadata dict Current revision indexers: - metadata: detects files containing metadata and retrieves translated_metadata in content_metadata table in storage or run content indexer to translate files. Platform: UNKNOWN Classifier: Programming Language :: Python :: 3 Classifier: Intended Audience :: Developers Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3) Classifier: Operating System :: OS Independent Classifier: Development Status :: 5 - Production/Stable Requires-Python: >=3.7 Description-Content-Type: text/markdown Provides-Extra: testing diff --git a/conftest.py b/conftest.py index de31662..d3cc5fd 100644 --- a/conftest.py +++ b/conftest.py @@ -1,19 +1,28 @@ # Copyright (C) 2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information +import pytest + from hypothesis import settings # define tests profile. Full documentation is at: # https://hypothesis.readthedocs.io/en/latest/settings.html#settings-profiles settings.register_profile("fast", max_examples=5, deadline=5000) settings.register_profile("slow", max_examples=20, deadline=5000) # Ignore the following modules because wsgi module fails as no # configuration file is found (--doctest-modules forces the module # loading) collect_ignore = ["swh/indexer/storage/api/wsgi.py"] # we use the swh_scheduler fixture pytest_plugins = ["swh.scheduler.pytest_plugin"] + + +@pytest.fixture(scope="session") +def swh_scheduler_celery_includes(swh_scheduler_celery_includes): + return swh_scheduler_celery_includes + [ + "swh.indexer.tasks", + ] diff --git a/requirements-swh.txt b/requirements-swh.txt index 0363717..39e073c 100644 --- a/requirements-swh.txt +++ b/requirements-swh.txt @@ -1,6 +1,6 @@ -swh.core[db,http] >= 0.0.87 +swh.core[db,http] >= 0.2.2 swh.model >= 0.0.15 swh.objstorage >= 0.0.43 -swh.scheduler >= 0.0.47 -swh.storage >= 0.8.0 +swh.scheduler >= 0.5.2 +swh.storage >= 0.12.0 swh.journal >= 0.1.0 diff --git a/requirements-test.txt b/requirements-test.txt index ac0c1f0..c1f90dc 100644 --- a/requirements-test.txt +++ b/requirements-test.txt @@ -1,5 +1,6 @@ confluent-kafka pytest +pytest-mock hypothesis>=3.11.0 swh.scheduler[testing] >= 0.5.0 swh.storage[testing] >= 0.10.0 diff --git a/swh.indexer.egg-info/PKG-INFO b/swh.indexer.egg-info/PKG-INFO index 06fbd34..bbc9a86 100644 --- a/swh.indexer.egg-info/PKG-INFO +++ b/swh.indexer.egg-info/PKG-INFO @@ -1,71 +1,71 @@ Metadata-Version: 2.1 Name: swh.indexer -Version: 0.1.1 +Version: 0.2.0 Summary: Software Heritage Content Indexer Home-page: https://forge.softwareheritage.org/diffusion/78/ Author: Software Heritage developers Author-email: swh-devel@inria.fr License: UNKNOWN Project-URL: Bug Reports, https://forge.softwareheritage.org/maniphest Project-URL: Funding, https://www.softwareheritage.org/donate Project-URL: Source, https://forge.softwareheritage.org/source/swh-indexer Project-URL: Documentation, https://docs.softwareheritage.org/devel/swh-indexer/ Description: swh-indexer ============ Tools to compute multiple indexes on SWH's raw contents: - content: - mimetype - ctags - language - fossology-license - metadata - revision: - metadata An indexer is in charge of: - looking up objects - extracting information from those objects - store those information in the swh-indexer db There are multiple indexers working on different object types: - content indexer: works with content sha1 hashes - revision indexer: works with revision sha1 hashes - origin indexer: works with origin identifiers Indexation procedure: - receive batch of ids - retrieve the associated data depending on object type - compute for that object some index - store the result to swh's storage Current content indexers: - mimetype (queue swh_indexer_content_mimetype): detect the encoding and mimetype - language (queue swh_indexer_content_language): detect the programming language - ctags (queue swh_indexer_content_ctags): compute tags information - fossology-license (queue swh_indexer_fossology_license): compute the license - metadata: translate file into translated_metadata dict Current revision indexers: - metadata: detects files containing metadata and retrieves translated_metadata in content_metadata table in storage or run content indexer to translate files. Platform: UNKNOWN Classifier: Programming Language :: Python :: 3 Classifier: Intended Audience :: Developers Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3) Classifier: Operating System :: OS Independent Classifier: Development Status :: 5 - Production/Stable Requires-Python: >=3.7 Description-Content-Type: text/markdown Provides-Extra: testing diff --git a/swh.indexer.egg-info/SOURCES.txt b/swh.indexer.egg-info/SOURCES.txt index 1dc3047..1440ceb 100644 --- a/swh.indexer.egg-info/SOURCES.txt +++ b/swh.indexer.egg-info/SOURCES.txt @@ -1,133 +1,134 @@ .gitignore .pre-commit-config.yaml AUTHORS CODE_OF_CONDUCT.md CONTRIBUTORS LICENSE MANIFEST.in Makefile Makefile.local README.md codemeta.json conftest.py mypy.ini pyproject.toml pytest.ini requirements-swh.txt requirements-test.txt requirements.txt setup.cfg setup.py tox.ini docs/.gitignore docs/Makefile docs/Makefile.local docs/README.md docs/conf.py docs/dev-info.rst docs/index.rst docs/metadata-workflow.rst docs/_static/.placeholder docs/_templates/.placeholder docs/images/.gitignore docs/images/Makefile docs/images/tasks-metadata-indexers.uml sql/bin/db-upgrade sql/bin/dot_add_content sql/doc/json sql/doc/json/.gitignore sql/doc/json/Makefile sql/doc/json/indexer_configuration.tool_configuration.schema.json sql/doc/json/revision_metadata.translated_metadata.json sql/json/.gitignore sql/json/Makefile sql/json/indexer_configuration.tool_configuration.schema.json sql/json/revision_metadata.translated_metadata.json sql/upgrades/115.sql sql/upgrades/116.sql sql/upgrades/117.sql sql/upgrades/118.sql sql/upgrades/119.sql sql/upgrades/120.sql sql/upgrades/121.sql sql/upgrades/122.sql sql/upgrades/123.sql sql/upgrades/124.sql sql/upgrades/125.sql sql/upgrades/126.sql sql/upgrades/127.sql sql/upgrades/128.sql sql/upgrades/129.sql sql/upgrades/130.sql sql/upgrades/131.sql sql/upgrades/132.sql swh/__init__.py swh.indexer.egg-info/PKG-INFO swh.indexer.egg-info/SOURCES.txt swh.indexer.egg-info/dependency_links.txt swh.indexer.egg-info/entry_points.txt swh.indexer.egg-info/requires.txt swh.indexer.egg-info/top_level.txt swh/indexer/__init__.py swh/indexer/cli.py swh/indexer/codemeta.py swh/indexer/ctags.py swh/indexer/fossology_license.py swh/indexer/indexer.py swh/indexer/journal_client.py swh/indexer/metadata.py swh/indexer/metadata_detector.py swh/indexer/mimetype.py swh/indexer/origin_head.py swh/indexer/py.typed swh/indexer/rehash.py swh/indexer/tasks.py swh/indexer/data/codemeta/CITATION swh/indexer/data/codemeta/LICENSE swh/indexer/data/codemeta/codemeta.jsonld swh/indexer/data/codemeta/crosswalk.csv swh/indexer/metadata_dictionary/__init__.py swh/indexer/metadata_dictionary/base.py swh/indexer/metadata_dictionary/codemeta.py swh/indexer/metadata_dictionary/maven.py swh/indexer/metadata_dictionary/npm.py swh/indexer/metadata_dictionary/python.py swh/indexer/metadata_dictionary/ruby.py swh/indexer/sql/10-swh-init.sql swh/indexer/sql/20-swh-enums.sql swh/indexer/sql/30-swh-schema.sql swh/indexer/sql/40-swh-func.sql swh/indexer/sql/50-swh-data.sql swh/indexer/sql/60-swh-indexes.sql swh/indexer/storage/__init__.py swh/indexer/storage/converters.py swh/indexer/storage/db.py swh/indexer/storage/exc.py swh/indexer/storage/in_memory.py swh/indexer/storage/interface.py swh/indexer/storage/metrics.py swh/indexer/storage/api/__init__.py swh/indexer/storage/api/client.py swh/indexer/storage/api/server.py swh/indexer/tests/__init__.py swh/indexer/tests/conftest.py swh/indexer/tests/tasks.py swh/indexer/tests/test_cli.py swh/indexer/tests/test_codemeta.py swh/indexer/tests/test_ctags.py swh/indexer/tests/test_fossology_license.py swh/indexer/tests/test_journal_client.py swh/indexer/tests/test_metadata.py swh/indexer/tests/test_mimetype.py swh/indexer/tests/test_origin_head.py swh/indexer/tests/test_origin_metadata.py +swh/indexer/tests/test_tasks.py swh/indexer/tests/utils.py swh/indexer/tests/storage/__init__.py swh/indexer/tests/storage/conftest.py swh/indexer/tests/storage/generate_data_test.py swh/indexer/tests/storage/test_api_client.py swh/indexer/tests/storage/test_converters.py swh/indexer/tests/storage/test_in_memory.py swh/indexer/tests/storage/test_metrics.py swh/indexer/tests/storage/test_server.py swh/indexer/tests/storage/test_storage.py \ No newline at end of file diff --git a/swh.indexer.egg-info/requires.txt b/swh.indexer.egg-info/requires.txt index 69ab181..4d17096 100644 --- a/swh.indexer.egg-info/requires.txt +++ b/swh.indexer.egg-info/requires.txt @@ -1,18 +1,19 @@ vcversioner click python-magic>=0.4.13 pyld xmltodict -swh.core[db,http]>=0.0.87 +swh.core[db,http]>=0.2.2 swh.model>=0.0.15 swh.objstorage>=0.0.43 -swh.scheduler>=0.0.47 -swh.storage>=0.8.0 +swh.scheduler>=0.5.2 +swh.storage>=0.12.0 swh.journal>=0.1.0 [testing] confluent-kafka pytest +pytest-mock hypothesis>=3.11.0 swh.scheduler[testing]>=0.5.0 swh.storage[testing]>=0.10.0 diff --git a/swh/indexer/fossology_license.py b/swh/indexer/fossology_license.py index 1b4cffa..4fc15f1 100644 --- a/swh/indexer/fossology_license.py +++ b/swh/indexer/fossology_license.py @@ -1,185 +1,188 @@ # 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 logging import subprocess -from typing import Any, Dict, List, Optional +from typing import Any, Dict, List, Optional, Union from swh.model import hashutil -from .indexer import ContentIndexer, ContentRangeIndexer, write_to_temp +from .indexer import ContentIndexer, ContentPartitionIndexer, write_to_temp +from swh.indexer.storage.interface import PagedResult, Sha1 logger = logging.getLogger(__name__) def compute_license(path): """Determine license from file at path. Args: path: filepath to determine the license Returns: dict: A dict with the following keys: - licenses ([str]): associated detected licenses to path - path (bytes): content filepath """ try: properties = subprocess.check_output(["nomossa", path], universal_newlines=True) if properties: res = properties.rstrip().split(" contains license(s) ") licenses = res[1].split(",") else: licenses = [] return { "licenses": licenses, "path": path, } except subprocess.CalledProcessError: from os import path as __path logger.exception( "Problem during license detection for sha1 %s" % __path.basename(path) ) return { "licenses": [], "path": path, } class MixinFossologyLicenseIndexer: """Mixin fossology license indexer. See :class:`FossologyLicenseIndexer` and - :class:`FossologyLicenseRangeIndexer` + :class:`FossologyLicensePartitionIndexer` """ ADDITIONAL_CONFIG = { "workdir": ("str", "/tmp/swh/indexer.fossology.license"), "tools": ( "dict", { "name": "nomos", "version": "3.1.0rc2-31-ga2cbb8c", "configuration": {"command_line": "nomossa ",}, }, ), "write_batch_size": ("int", 1000), } CONFIG_BASE_FILENAME = "indexer/fossology_license" # type: Optional[str] tool: Any idx_storage: Any def prepare(self): super().prepare() self.working_directory = self.config["workdir"] def index( - self, id: bytes, data: Optional[bytes] = None, **kwargs + self, id: Union[bytes, Dict], data: Optional[bytes] = None, **kwargs ) -> Dict[str, Any]: """Index sha1s' content and store result. Args: id (bytes): content's identifier raw_content (bytes): associated raw content to content id Returns: dict: A dict, representing a content_license, with keys: - id (bytes): content's identifier (sha1) - license (bytes): license in bytes - path (bytes): path - indexer_configuration_id (int): tool used to compute the output """ assert isinstance(id, bytes) assert data is not None with write_to_temp( filename=hashutil.hash_to_hex(id), # use the id as pathname data=data, working_directory=self.working_directory, ) as content_path: properties = compute_license(path=content_path) properties.update( {"id": id, "indexer_configuration_id": self.tool["id"],} ) 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_license dict with the following keys: - id (bytes): content's identifier (sha1) - license (bytes): license in bytes - path (bytes): path policy_update: either 'update-dups' or 'ignore-dups' to respectively update duplicates or ignore them """ return self.idx_storage.content_fossology_license_add( results, conflict_update=(policy_update == "update-dups") ) class FossologyLicenseIndexer(MixinFossologyLicenseIndexer, ContentIndexer): """Indexer in charge of: - filtering out content already indexed - reading content from objstorage per the content's id (sha1) - computing {license, encoding} from that content - store result in storage """ def filter(self, ids): """Filter out known sha1s and return only missing ones. """ yield from self.idx_storage.content_fossology_license_missing( ({"id": sha1, "indexer_configuration_id": self.tool["id"],} for sha1 in ids) ) -class FossologyLicenseRangeIndexer(MixinFossologyLicenseIndexer, ContentRangeIndexer): - """FossologyLicense Range Indexer working on range of content identifiers. +class FossologyLicensePartitionIndexer( + MixinFossologyLicenseIndexer, ContentPartitionIndexer +): + """FossologyLicense Range Indexer working on range/partition of content identifiers. - filters out the non textual content - (optionally) filters out content already indexed (cf - :meth:`.indexed_contents_in_range`) + :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_range(self, start, end): - """Retrieve indexed content id within range [start, end]. + def indexed_contents_in_partition( + self, partition_id: int, nb_partitions: int, page_token: Optional[str] = None + ) -> PagedResult[Sha1]: + """Retrieve indexed content id within the partition id Args: - start (bytes): Starting bound from range identifier - end (bytes): End range identifier + 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: - dict: a dict with keys: - - - **ids** [bytes]: iterable of content ids within the range. - - **next** (Optional[bytes]): The next range of sha1 starts at - this sha1 if any + PagedResult of Sha1. If next_page_token is None, there is no more data + to fetch """ - return self.idx_storage.content_fossology_license_get_range( - start, end, self.tool["id"] + return self.idx_storage.content_fossology_license_get_partition( + self.tool["id"], partition_id, nb_partitions, page_token=page_token ) diff --git a/swh/indexer/indexer.py b/swh/indexer/indexer.py index 7e332a7..168639e 100644 --- a/swh/indexer/indexer.py +++ b/swh/indexer/indexer.py @@ -1,621 +1,633 @@ # 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 import os import logging import shutil import tempfile from contextlib import contextmanager from typing import Any, Dict, Iterator, List, Optional, Set, Tuple, Union from swh.scheduler import CONFIG as SWH_CONFIG from swh.storage import get_storage from swh.core.config import SWHConfig from swh.objstorage import get_objstorage from swh.objstorage.exc import ObjNotFoundError -from swh.indexer.storage import get_indexer_storage, INDEXER_CFG_KEY +from swh.indexer.storage import get_indexer_storage, INDEXER_CFG_KEY, PagedResult, Sha1 from swh.model import hashutil from swh.core import utils @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) class BaseIndexer(SWHConfig, 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] CONFIG = "indexer/base" DEFAULT_CONFIG = { INDEXER_CFG_KEY: ( "dict", {"cls": "remote", "args": {"url": "http://localhost:5007/"}}, ), "storage": ( "dict", {"cls": "remote", "args": {"url": "http://localhost:5002/",}}, ), "objstorage": ( "dict", {"cls": "remote", "args": {"url": "http://localhost:5003/",}}, ), } ADDITIONAL_CONFIG = {} # type: Dict[str, Tuple[str, Any]] 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: config_keys = ( "base_filename", "config_filename", "additional_configs", "global_config", ) config_args = {k: v for k, v in kw.items() if k in config_keys} if self.ADDITIONAL_CONFIG: config_args.setdefault("additional_configs", []).append( self.ADDITIONAL_CONFIG ) self.config = self.parse_config_file(**config_args) 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: bytes, data: Optional[bytes] = None, **kwargs + self, id: Union[bytes, Dict], data: Optional[bytes] = None, **kwargs ) -> Dict[str, Any]: """Index computation for the id and associated raw data. Args: - id: identifier + 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 range, use the :class:`ContentRangeIndexer` + 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 """ status = "uneventful" sha1s = [ hashutil.hash_to_bytes(id_) if isinstance(id_, str) else id_ for id_ in ids ] results = [] summary: Dict = {} 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) 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.") status = "failed" finally: summary["status"] = status return summary -class ContentRangeIndexer(BaseIndexer): - """A content range indexer. +class ContentPartitionIndexer(BaseIndexer): + """A content partition indexer. - This expects as input a range of ids to index. + 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:`ContentRangeIndexer` is not an instantiable + 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_range(self, start: bytes, end: bytes) -> Any: + 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: - start: Starting bound from range identifier - end: End range identifier + partition_id: Index of the partition to fetch + nb_partitions: Total number of partitions to split into + page_token: opaque token used for pagination - Yields: - bytes: Content identifier present in the range ``[start, end]`` + Returns: + PagedResult of Sha1. If next_page_token is None, there is no more data + to fetch """ pass def _list_contents_to_index( - self, start: bytes, end: bytes, indexed: Set[bytes] - ) -> Iterator[bytes]: - """Compute from storage the new contents to index in the range [start, - end]. The already indexed contents are skipped. + 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: - start: Starting bound from range identifier - end: End range identifier + partition_id: Index of the partition to fetch data from + nb_partitions: Total number of partition indexed: Set of content already indexed. Yields: - bytes: Identifier of contents to index. + Sha1 id (bytes) of contents to index """ - if not isinstance(start, bytes) or not isinstance(end, bytes): - raise TypeError("identifiers must be bytes, not %r and %r." % (start, end)) - while start: - result = self.storage.content_get_range(start, end) - contents = result["contents"] + 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"]) + _id = hashutil.hash_to_bytes(c.sha1) if _id in indexed: continue yield _id - start = result["next"] + next_page_token = result.next_page_token + if next_page_token is None: + break def _index_contents( - self, start: bytes, end: bytes, indexed: Set[bytes], **kwargs: Any + self, partition_id: int, nb_partitions: int, indexed: Set[Sha1], **kwargs: Any ) -> Iterator[Dict]: - """Index the contents from within range [start, end] + """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: - dict: Data indexed to persist using the indexer storage + indexing result as dict to persist in the indexer backend """ - for sha1 in self._list_contents_to_index(start, end, indexed): + 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( - "Content %s not found in objstorage" % hashutil.hash_to_hex(sha1) - ) + 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): 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, start: bytes, end: bytes + self, partition_id: int, nb_partitions: int ) -> Iterator[Dict]: - """Index not already indexed contents in range [start, end]. + """Index not already indexed contents within the partition partition_id Args: - start: Starting range identifier - end: Ending range identifier + partition_id: Index of the partition to fetch + nb_partitions: Total number of partitions to split into Yields: - dict: Content identifier present in the range - ``[start, end]`` which are not already indexed. + indexing result as dict to persist in the indexer backend """ - while start: - indexed_page = self.indexed_contents_in_range(start, end) - contents = indexed_page["ids"] - _end = contents[-1] if contents else end - yield from self._index_contents(start, _end, contents) - start = indexed_page["next"] + 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, - start: Union[bytes, str], - end: Union[bytes, str], + partition_id: int, + nb_partitions: int, skip_existing: bool = True, - **kwargs + **kwargs, ) -> Dict: - """Given a range of content ids, compute the indexing computations on - the contents within. Either the indexer is incremental - (filter out existing computed data) or not (compute - everything from scratch). + """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: - start: Starting range identifier - end: Ending range identifier + 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 + (default) or not **kwargs: passed to the `index` method Returns: - A dict with the task's status + dict with the indexing task status """ status = "uneventful" summary: Dict[str, Any] = {} count = 0 try: - range_start = ( - hashutil.hash_to_bytes(start) if isinstance(start, str) else start - ) - range_end = hashutil.hash_to_bytes(end) if isinstance(end, str) else end - if skip_existing: - gen = self._index_with_skipping_already_done(range_start, range_end) + gen = self._index_with_skipping_already_done( + partition_id, nb_partitions + ) else: - gen = self._index_contents(range_start, range_end, indexed=set([])) + 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( 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: status = "eventful" except Exception: if not self.catch_exceptions: raise self.log.exception("Problem when computing metadata.") status = "failed" finally: summary["status"] = status 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" results = self.index_list(origin_urls, **kwargs) summary_persist = self.persist_index_computations(results, policy_update) self.results = results if summary_persist: for value in summary_persist.values(): if value > 0: status = "eventful" summary.update(summary_persist) summary["status"] = status return summary def index_list(self, origins: List[Any], **kwargs: Any) -> List[Dict]: results = [] for origin in origins: try: res = self.index(origin, **kwargs) if res: # If no results, skip it results.append(res) except Exception: if not self.catch_exceptions: raise self.log.exception("Problem when processing origin %s", origin["id"]) 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 = [] - revs = self.storage.revision_get( - hashutil.hash_to_bytes(id_) if isinstance(id_, str) else id_ for id_ in ids - ) - for rev in revs: + 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") status = "failed" summary_persist = self.persist_index_computations(results, policy_update) if summary_persist: for value in summary_persist.values(): if value > 0: status = "eventful" summary.update(summary_persist) self.results = results summary["status"] = status return summary diff --git a/swh/indexer/mimetype.py b/swh/indexer/mimetype.py index 384e5bc..890dcd8 100644 --- a/swh/indexer/mimetype.py +++ b/swh/indexer/mimetype.py @@ -1,157 +1,161 @@ # 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 Optional, Dict, Any, List import magic -from .indexer import ContentIndexer, ContentRangeIndexer +from typing import Any, Optional, Dict, List, Union + +from swh.indexer.storage.interface import PagedResult, Sha1 + +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]: """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, } class MixinMimetypeIndexer: """Mixin mimetype indexer. - See :class:`MimetypeIndexer` and :class:`MimetypeRangeIndexer` + See :class:`MimetypeIndexer` and :class:`MimetypePartitionIndexer` """ tool: Any idx_storage: Any ADDITIONAL_CONFIG = { "tools": ( "dict", { "name": "file", "version": "1:5.30-1+deb9u1", "configuration": {"type": "library", "debian-package": "python3-magic"}, }, ), "write_batch_size": ("int", 1000), } CONFIG_BASE_FILENAME = "indexer/mimetype" # type: Optional[str] def index( - self, id: bytes, data: Optional[bytes] = None, **kwargs + self, id: Union[bytes, Dict], data: Optional[bytes] = None, **kwargs ) -> Dict[str, Any]: """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 data is not None properties = compute_mimetype_encoding(data) + assert isinstance(id, bytes) properties.update( {"id": id, "indexer_configuration_id": self.tool["id"],} ) 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 MimetypeRangeIndexer(MixinMimetypeIndexer, ContentRangeIndexer): +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_range`) + :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_range( - self, start: bytes, end: bytes - ) -> Dict[str, Optional[bytes]]: - """Retrieve indexed content id within range [start, end]. + 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: - start: Starting bound from range identifier - end: End range identifier + 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: - dict: a dict with keys: - - - ids: iterable of content ids within the range. - - next: The next range of sha1 starts at - this sha1 if any + PagedResult of Sha1. If next_page_token is None, there is no more data + to fetch """ - return self.idx_storage.content_mimetype_get_range(start, end, self.tool["id"]) + return self.idx_storage.content_mimetype_get_partition( + self.tool["id"], partition_id, nb_partitions, page_token=page_token + ) diff --git a/swh/indexer/rehash.py b/swh/indexer/rehash.py index 2593d67..06d907a 100644 --- a/swh/indexer/rehash.py +++ b/swh/indexer/rehash.py @@ -1,189 +1,192 @@ # 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 logging import itertools from collections import defaultdict from typing import Dict, Any, Tuple, List, Generator from swh.core import utils from swh.core.config import SWHConfig from swh.model import hashutil from swh.objstorage import get_objstorage from swh.objstorage.exc import ObjNotFoundError from swh.storage import get_storage class RecomputeChecksums(SWHConfig): """Class in charge of (re)computing content's hashes. Hashes to compute are defined across 2 configuration options: compute_checksums ([str]) list of hash algorithms that py:func:`swh.model.hashutil.MultiHash.from_data` function should be able to deal with. For variable-length checksums, a desired checksum length should also be provided. Their format is : e.g: blake2:512 recompute_checksums (bool) a boolean to notify that we also want to recompute potential existing hashes specified in compute_checksums. Default to False. """ DEFAULT_CONFIG = { # The storage to read from or update metadata to "storage": ( "dict", {"cls": "remote", "args": {"url": "http://localhost:5002/"},}, ), # The objstorage to read contents' data from "objstorage": ( "dict", { "cls": "pathslicing", "args": { "root": "/srv/softwareheritage/objects", "slicing": "0:2/2:4/4:6", }, }, ), # the set of checksums that should be computed. # Examples: 'sha1_git', 'blake2b512', 'blake2s256' "compute_checksums": ("list[str]", []), # whether checksums that already exist in the DB should be # recomputed/updated or left untouched "recompute_checksums": ("bool", False), # Number of contents to retrieve blobs at the same time "batch_size_retrieve_content": ("int", 10), # Number of contents to update at the same time "batch_size_update": ("int", 100), } CONFIG_BASE_FILENAME = "indexer/rehash" def __init__(self) -> None: self.config = self.parse_config_file() self.storage = get_storage(**self.config["storage"]) self.objstorage = get_objstorage(**self.config["objstorage"]) self.compute_checksums = self.config["compute_checksums"] self.recompute_checksums = self.config["recompute_checksums"] self.batch_size_retrieve_content = self.config["batch_size_retrieve_content"] self.batch_size_update = self.config["batch_size_update"] self.log = logging.getLogger("swh.indexer.rehash") if not self.compute_checksums: raise ValueError("Checksums list should not be empty.") def _read_content_ids( self, contents: List[Dict[str, Any]] ) -> Generator[bytes, Any, None]: """Read the content identifiers from the contents. """ for c in contents: h = c["sha1"] if isinstance(h, str): h = hashutil.hash_to_bytes(h) yield h def get_new_contents_metadata( self, all_contents: List[Dict[str, Any]] ) -> Generator[Tuple[Dict[str, Any], List[Any]], Any, None]: """Retrieve raw contents and compute new checksums on the contents. Unknown or corrupted contents are skipped. Args: all_contents: List of contents as dictionary with the necessary primary keys Yields: tuple: tuple of (content to update, list of checksums computed) """ content_ids = self._read_content_ids(all_contents) for contents in utils.grouper(content_ids, self.batch_size_retrieve_content): contents_iter = itertools.tee(contents, 2) try: - content_metadata = self.storage.content_get_metadata( + content_metadata: Dict[ + bytes, List[Dict] + ] = self.storage.content_get_metadata( # noqa [s for s in contents_iter[0]] ) except Exception: self.log.exception("Problem when reading contents metadata.") continue - for content in content_metadata: + for sha1, content_dicts in content_metadata.items(): + if not content_dicts: + continue + content: Dict = content_dicts[0] # Recompute checksums provided in compute_checksums options if self.recompute_checksums: checksums_to_compute = list(self.compute_checksums) else: # Compute checksums provided in compute_checksums # options not already defined for that content checksums_to_compute = [ h for h in self.compute_checksums if not content.get(h) ] if not checksums_to_compute: # Nothing to recompute continue try: - raw_content = self.objstorage.get(content["sha1"]) + raw_content = self.objstorage.get(sha1) except ObjNotFoundError: - self.log.warning( - "Content %s not found in objstorage!" % content["sha1"] - ) + self.log.warning("Content %s not found in objstorage!", sha1) continue content_hashes = hashutil.MultiHash.from_data( raw_content, hash_names=checksums_to_compute ).digest() content.update(content_hashes) yield content, checksums_to_compute def run(self, contents: List[Dict[str, Any]]) -> Dict: """Given a list of content: - (re)compute a given set of checksums on contents available in our object storage - update those contents with the new metadata Args: contents: contents as dictionary with necessary keys. key present in such dictionary should be the ones defined in the 'primary_key' option. Returns: A summary dict with key 'status', task' status and 'count' the number of updated contents. """ status = "uneventful" count = 0 for data in utils.grouper( self.get_new_contents_metadata(contents), self.batch_size_update ): groups: Dict[str, List[Any]] = defaultdict(list) for content, keys_to_update in data: - keys = ",".join(keys_to_update) - groups[keys].append(content) + keys_str = ",".join(keys_to_update) + groups[keys_str].append(content) for keys_to_update, contents in groups.items(): - keys = keys_to_update.split(",") + keys: List[str] = keys_to_update.split(",") try: self.storage.content_update(contents, keys=keys) count += len(contents) status = "eventful" except Exception: self.log.exception("Problem during update.") continue return { "status": status, "count": count, } diff --git a/swh/indexer/storage/__init__.py b/swh/indexer/storage/__init__.py index 9cad65d..c023d5d 100644 --- a/swh/indexer/storage/__init__.py +++ b/swh/indexer/storage/__init__.py @@ -1,599 +1,649 @@ # Copyright (C) 2015-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import json import psycopg2 import psycopg2.pool from collections import defaultdict, Counter -from typing import Dict, List +from typing import Dict, List, Optional +from swh.model.hashutil import hash_to_bytes, hash_to_hex +from swh.model.model import SHA1_SIZE from swh.storage.common import db_transaction_generator, db_transaction from swh.storage.exc import StorageDBError +from swh.storage.utils import get_partition_bounds_bytes + +from .interface import PagedResult, Sha1 from . import converters from .db import Db from .exc import IndexerStorageArgumentException, DuplicateId from .metrics import process_metrics, send_metric, timed 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 dictionaries in `data` have the same id, raises a `ValueError`. Values associated to the key must be hashable. Args: data (List[dict]): List of dictionaries to be inserted >>> check_id_duplicates([ ... {'id': 'foo', 'data': 'spam'}, ... {'id': 'bar', 'data': 'egg'}, ... ]) >>> check_id_duplicates([ ... {'id': 'foo', 'data': 'spam'}, ... {'id': 'foo', 'data': 'egg'}, ... ]) Traceback (most recent call last): ... swh.indexer.storage.exc.DuplicateId: ['foo'] """ counter = Counter(item["id"] for item in data) duplicates = [id_ for (id_, count) in counter.items() if count >= 2] if duplicates: raise DuplicateId(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): for obj in db.content_mimetype_missing_from_list(mimetypes, cur): yield obj[0] - def _content_get_range( + @timed + @db_transaction() + def get_partition( self, - content_type, - start, - end, - indexer_configuration_id, - limit=1000, + 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 content_type not in db.content_indexer_names: - err = "Wrong type. Should be one of [%s]" % ( - ",".join(db.content_indexer_names) - ) + 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) - ids = [] - next_id = None - for counter, obj in enumerate( - db.content_get_range( - content_type, + 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, ) - ): - _id = obj[0] - if counter >= limit: - next_id = _id - break + ] - ids.append(_id) + if len(ids) >= limit: + next_page_token = hash_to_hex(ids[-1]) + ids = ids[:limit] - return {"ids": ids, "next": next_id} + assert len(ids) <= limit + return PagedResult(results=ids, next_page_token=next_page_token) @timed @db_transaction() - def content_mimetype_get_range( - self, start, end, indexer_configuration_id, limit=1000, db=None, cur=None - ): - return self._content_get_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, + db=None, + cur=None, + ) -> PagedResult[Sha1]: + return self.get_partition( "mimetype", - start, - end, 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 ) -> 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(mimetypes) mimetypes.sort(key=lambda m: m["id"]) db.mktemp_content_mimetype(cur) db.copy_to( 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): for c in db.content_mimetype_get_from_list(ids, cur): yield 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(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]: check_id_duplicates(ctags) ctags.sort(key=lambda m: m["id"]) def _convert_ctags(__ctags): """Convert ctags dict to list of ctags. """ for ctags in __ctags: yield from converters.ctags_to_db(ctags) db.mktemp_content_ctags(cur) db.copy_to( list(_convert_ctags(ctags)), tblname="tmp_content_ctags", columns=["id", "name", "kind", "line", "lang", "indexer_configuration_id"], cur=cur, ) count = db.content_ctags_add_from_temp(conflict_update, cur) return {"content_ctags:add": count} @timed @db_transaction_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]: check_id_duplicates(licenses) licenses.sort(key=lambda m: m["id"]) db.mktemp_content_fossology_license(cur) db.copy_to( ( { "id": sha1["id"], "indexer_configuration_id": sha1["indexer_configuration_id"], "license": license, } for sha1 in licenses for license in sha1["licenses"] ), tblname="tmp_content_fossology_license", columns=["id", "license", "indexer_configuration_id"], cur=cur, ) count = db.content_fossology_license_add_from_temp(conflict_update, cur) return {"content_fossology_license:add": count} @timed @db_transaction() - def content_fossology_license_get_range( - self, start, end, indexer_configuration_id, limit=1000, db=None, cur=None - ): - return self._content_get_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, + db=None, + cur=None, + ) -> PagedResult[Sha1]: + return self.get_partition( "fossology_license", - start, - end, 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(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(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(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/in_memory.py b/swh/indexer/storage/in_memory.py index 1bcc46d..a70f212 100644 --- a/swh/indexer/storage/in_memory.py +++ b/swh/indexer/storage/in_memory.py @@ -1,455 +1,496 @@ # 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 -import bisect -from collections import defaultdict, Counter import itertools import json import operator import math import re -from typing import Any, Dict, List + +from collections import defaultdict, Counter +from typing import Any, Dict, List, Optional + +from swh.model.model import SHA1_SIZE +from swh.model.hashutil import hash_to_hex, hash_to_bytes +from swh.storage.utils import get_partition_bounds_bytes +from swh.storage.in_memory import SortedList from . import MAPPING_NAMES, check_id_duplicates from .exc import IndexerStorageArgumentException +from .interface import PagedResult, Sha1 + 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.") class SubStorage: """Implements common missing/get/add logic for each indexer type.""" def __init__(self, tools): self._tools = tools - self._sorted_ids = [] + self._sorted_ids = SortedList[bytes, bytes]() self._data = {} # map (id_, tool_id) -> metadata_dict self._tools_per_id = defaultdict(set) # map id_ -> Set[tool_id] def missing(self, ids): """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 id_ in ids: tool_id = id_["indexer_configuration_id"] id_ = id_["id"] if tool_id not in self._tools_per_id.get(id_, set()): yield id_ def get(self, ids): """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 tool_id in self._tools_per_id.get(id_, set()): key = (id_, tool_id) yield { "id": id_, "tool": _transform_tool(self._tools[tool_id]), **self._data[key], } def get_all(self): yield from self.get(self._sorted_ids) - def get_range(self, start, end, indexer_configuration_id, limit): - """Retrieve data within range [start, end] bound by limit. + 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: - **start** (bytes): Starting identifier range (expected smaller - than end) - **end** (bytes): Ending identifier range (expected larger - than start) - **indexer_configuration_id** (int): The tool used to index data - **limit** (int): Limit result + **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 + IndexerStorageArgumentException for; + - limit to None + - wrong indexer_type provided Returns: - a dict with keys: - - **ids** [bytes]: iterable of content ids within the range. - - **next** (Optional[bytes]): The next range of sha1 starts at - this sha1 if any + 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") - from_index = bisect.bisect_left(self._sorted_ids, start) - to_index = bisect.bisect_right(self._sorted_ids, end, lo=from_index) - if to_index - from_index >= limit: - return { - "ids": self._sorted_ids[from_index : from_index + limit], - "next": self._sorted_ids[from_index + limit], - } - else: - return { - "ids": self._sorted_ids[from_index:to_index], - "next": 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: List[Dict], 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 item in data: item = item.copy() tool_id = item.pop("indexer_configuration_id") id_ = item.pop("id") data_item = item if not conflict_update and tool_id in self._tools_per_id.get(id_, set()): # Duplicate, should not be updated continue key = (id_, tool_id) self._data[key] = data_item self._tools_per_id[id_].add(tool_id) count += 1 if id_ not in self._sorted_ids: - bisect.insort(self._sorted_ids, id_) + self._sorted_ids.add(id_) return count def add_merge( self, new_data: List[Dict], conflict_update: bool, merged_key: str ) -> int: added = 0 all_subitems: List for new_item in new_data: id_ = new_item["id"] tool_id = new_item["indexer_configuration_id"] if conflict_update: all_subitems = [] else: existing = list(self.get([id_])) all_subitems = [ old_subitem for existing_item in existing if existing_item["tool"]["id"] == tool_id for old_subitem in existing_item[merged_key] ] for new_subitem in new_item[merged_key]: if new_subitem not in all_subitems: all_subitems.append(new_subitem) added += self.add( [ { "id": id_, "indexer_configuration_id": tool_id, merged_key: all_subitems, } ], conflict_update=True, ) if id_ not in self._sorted_ids: - bisect.insort(self._sorted_ids, id_) + self._sorted_ids.add(id_) return added 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"]) key = (id_, tool_id) if tool_id in self._tools_per_id[id_]: self._tools_per_id[id_].remove(tool_id) if key in self._data: deleted += 1 del self._data[key] return deleted class IndexerStorage: """In-memory SWH indexer storage.""" def __init__(self): self._tools = {} self._mimetypes = SubStorage(self._tools) self._languages = SubStorage(self._tools) self._content_ctags = SubStorage(self._tools) self._licenses = SubStorage(self._tools) self._content_metadata = SubStorage(self._tools) self._revision_intrinsic_metadata = SubStorage(self._tools) self._origin_intrinsic_metadata = SubStorage(self._tools) def check_config(self, *, check_write): return True def content_mimetype_missing(self, mimetypes): yield from self._mimetypes.missing(mimetypes) - def content_mimetype_get_range( - self, start, end, indexer_configuration_id, limit=1000 - ): - return self._mimetypes.get_range(start, end, indexer_configuration_id, limit) + 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 ) -> Dict[str, int]: check_id_types(mimetypes) added = self._mimetypes.add(mimetypes, conflict_update) return {"content_mimetype:add": added} def content_mimetype_get(self, ids): 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 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(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): for item_ctags_item in item["ctags"]: yield {"id": item["id"], "tool": item["tool"], **item_ctags_item} def content_ctags_add( self, ctags: List[Dict], conflict_update: bool = False ) -> Dict[str, int]: check_id_types(ctags) added = self._content_ctags.add_merge(ctags, conflict_update, "ctags") return {"content_ctags:add": added} def content_ctags_search(self, expression, limit=10, last_sha1=None): nb_matches = 0 for ((id_, tool_id), item) in sorted(self._content_ctags._data.items()): if id_ <= (last_sha1 or bytes(0 for _ in range(SHA1_DIGEST_SIZE))): continue for ctags_item in item["ctags"]: if ctags_item["name"] != expression: continue nb_matches += 1 yield { "id": id_, "tool": _transform_tool(self._tools[tool_id]), **ctags_item, } if nb_matches >= limit: return 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 res = {} for d in self._licenses.get(ids): res.setdefault(d.pop("id"), []).append(d) for (id_, facts) in res.items(): yield {id_: facts} def content_fossology_license_add( self, licenses: List[Dict], conflict_update: bool = False ) -> Dict[str, int]: check_id_types(licenses) added = self._licenses.add_merge(licenses, conflict_update, "licenses") return {"fossology_license_add:add": added} - def content_fossology_license_get_range( - self, start, end, indexer_configuration_id, limit=1000 - ): - return self._licenses.get_range(start, end, indexer_configuration_id, limit) + 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 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(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 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(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 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(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 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) 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 d059cf5..0fb6620 100644 --- a/swh/indexer/storage/interface.py +++ b/swh/indexer/storage/interface.py @@ -1,641 +1,617 @@ # 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 +from typing import Dict, List, Optional, TypeVar from swh.core.api import remote_api_endpoint +from swh.core.api.classes import PagedResult as CorePagedResult + + +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): """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 """ ... - def _content_get_range( + @remote_api_endpoint("content_mimetype/range") + def content_mimetype_get_partition( self, - content_type, - start, - end, - indexer_configuration_id, - limit=1000, - with_textual_data=False, - ): - """Retrieve ids of type content_type within range [start, end] bound - by limit. + 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: - **content_type** (str): content's type (mimetype, language, etc...) - **start** (bytes): Starting identifier range (expected smaller - than end) - **end** (bytes): Ending identifier range (expected larger - than start) - **indexer_configuration_id** (int): The tool used to index data - **limit** (int): Limit result (default to 1000) - **with_textual_data** (bool): Deal with only textual - content (True) or all - content (all contents by - defaults, False) + **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: - ValueError for; + IndexerStorageArgumentException for; - limit to None - - wrong content_type provided - - Returns: - a dict with keys: - - **ids** [bytes]: iterable of content ids within the range. - - **next** (Optional[bytes]): The next range of sha1 starts at - this sha1 if any - - """ - ... - - @remote_api_endpoint("content_mimetype/range") - def content_mimetype_get_range( - self, start, end, indexer_configuration_id, limit=1000 - ): - """Retrieve mimetypes within range [start, end] bound by limit. - - Args: - **start** (bytes): Starting identifier range (expected smaller - than end) - **end** (bytes): Ending identifier range (expected larger - than start) - **indexer_configuration_id** (int): The tool used to index data - **limit** (int): Limit result (default to 1000) - - Raises: - ValueError for limit to None + - wrong indexer_type provided Returns: - a dict with keys: - - **ids** [bytes]: iterable of content ids within the range. - - **next** (Optional[bytes]): The next range of sha1 starts at - this sha1 if any + 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 ) -> 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) Returns: Dict summary of number of rows added """ ... @remote_api_endpoint("content_mimetype") def content_mimetype_get(self, ids): """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_range( - self, start, end, indexer_configuration_id, limit=1000 - ): - """Retrieve licenses within range [start, end] bound by limit. + 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: - **start** (bytes): Starting identifier range (expected smaller - than end) - **end** (bytes): Ending identifier range (expected larger - than start) - **indexer_configuration_id** (int): The tool used to index data - **limit** (int): Limit result (default to 1000) + **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: - ValueError for limit to None + IndexerStorageArgumentException for; + - limit to None + - wrong indexer_type provided - Returns: - a dict with keys: - - **ids** [bytes]: iterable of content ids within the range. - - **next** (Optional[bytes]): The next range of sha1 starts at - this sha1 if any + 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/tasks.py b/swh/indexer/tasks.py index 2ca6cdd..51c07ef 100644 --- a/swh/indexer/tasks.py +++ b/swh/indexer/tasks.py @@ -1,48 +1,48 @@ # 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 celery import current_app as app +from celery import shared_task -from .mimetype import MimetypeIndexer, MimetypeRangeIndexer +from .mimetype import MimetypeIndexer, MimetypePartitionIndexer from .ctags import CtagsIndexer -from .fossology_license import FossologyLicenseIndexer, FossologyLicenseRangeIndexer +from .fossology_license import FossologyLicenseIndexer, FossologyLicensePartitionIndexer from .rehash import RecomputeChecksums from .metadata import OriginMetadataIndexer -@app.task(name=__name__ + ".OriginMetadata") +@shared_task(name=__name__ + ".OriginMetadata") def origin_metadata(*args, **kwargs): return OriginMetadataIndexer().run(*args, **kwargs) -@app.task(name=__name__ + ".Ctags") +@shared_task(name=__name__ + ".Ctags") def ctags(*args, **kwargs): return CtagsIndexer().run(*args, **kwargs) -@app.task(name=__name__ + ".ContentFossologyLicense") +@shared_task(name=__name__ + ".ContentFossologyLicense") def fossology_license(*args, **kwargs): return FossologyLicenseIndexer().run(*args, **kwargs) -@app.task(name=__name__ + ".RecomputeChecksums") +@shared_task(name=__name__ + ".RecomputeChecksums") def recompute_checksums(*args, **kwargs): return RecomputeChecksums().run(*args, **kwargs) -@app.task(name=__name__ + ".ContentMimetype") +@shared_task(name=__name__ + ".ContentMimetype") def mimetype(*args, **kwargs): return MimetypeIndexer().run(*args, **kwargs) -@app.task(name=__name__ + ".ContentRangeMimetype") -def range_mimetype(*args, **kwargs): - return MimetypeRangeIndexer().run(*args, **kwargs) +@shared_task(name=__name__ + ".ContentMimetypePartition") +def mimetype_partition(*args, **kwargs): + return MimetypePartitionIndexer().run(*args, **kwargs) -@app.task(name=__name__ + ".ContentRangeFossologyLicense") -def range_license(*args, **kwargs): - return FossologyLicenseRangeIndexer().run(*args, **kwargs) +@shared_task(name=__name__ + ".ContentFossologyLicensePartition") +def license_partition(*args, **kwargs): + return FossologyLicensePartitionIndexer().run(*args, **kwargs) diff --git a/swh/indexer/tests/conftest.py b/swh/indexer/tests/conftest.py index 1ba1528..6438b89 100644 --- a/swh/indexer/tests/conftest.py +++ b/swh/indexer/tests/conftest.py @@ -1,74 +1,101 @@ # Copyright (C) 2019-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information +import os + from datetime import timedelta from unittest.mock import patch +import yaml import pytest from swh.objstorage import get_objstorage from swh.storage import get_storage from swh.indexer.storage import get_indexer_storage from .utils import fill_storage, fill_obj_storage TASK_NAMES = ["revision_intrinsic_metadata", "origin_intrinsic_metadata"] @pytest.fixture def indexer_scheduler(swh_scheduler): for taskname in TASK_NAMES: swh_scheduler.create_task_type( { "type": taskname, "description": "The {} indexer testing task".format(taskname), "backend_name": "swh.indexer.tests.tasks.{}".format(taskname), "default_interval": timedelta(days=1), "min_interval": timedelta(hours=6), "max_interval": timedelta(days=12), "num_retries": 3, } ) return swh_scheduler @pytest.fixture def idx_storage(): """An instance of in-memory indexer storage that gets injected into all indexers classes. """ idx_storage = get_indexer_storage("memory", {}) with patch("swh.indexer.storage.in_memory.IndexerStorage") as idx_storage_mock: idx_storage_mock.return_value = idx_storage yield idx_storage @pytest.fixture def storage(): """An instance of in-memory storage that gets injected into all indexers classes. """ storage = get_storage(cls="memory") fill_storage(storage) with patch("swh.storage.in_memory.InMemoryStorage") as storage_mock: storage_mock.return_value = storage yield storage @pytest.fixture def obj_storage(): """An instance of in-memory objstorage that gets injected into all indexers classes. """ objstorage = get_objstorage("memory", {}) fill_obj_storage(objstorage) with patch.dict( "swh.objstorage.factory._STORAGE_CLASSES", {"memory": lambda: objstorage} ): yield objstorage + + +@pytest.fixture +def swh_indexer_config(): + return { + "storage": {"cls": "memory"}, + "objstorage": {"cls": "memory", "args": {},}, + "indexer_storage": {"cls": "memory", "args": {},}, + "tools": { + "name": "file", + "version": "1:5.30-1+deb9u1", + "configuration": {"type": "library", "debian-package": "python3-magic"}, + }, + "compute_checksums": ["blake2b512"], # for rehash indexer + } + + +@pytest.fixture +def swh_config(swh_indexer_config, monkeypatch, tmp_path): + conffile = os.path.join(str(tmp_path), "indexer.yml") + with open(conffile, "w") as f: + f.write(yaml.dump(swh_indexer_config)) + monkeypatch.setenv("SWH_CONFIG_FILENAME", conffile) + return conffile diff --git a/swh/indexer/tests/storage/test_storage.py b/swh/indexer/tests/storage/test_storage.py index 9c5b892..2101caa 100644 --- a/swh/indexer/tests/storage/test_storage.py +++ b/swh/indexer/tests/storage/test_storage.py @@ -1,1831 +1,1899 @@ # 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 import pytest from swh.model.hashutil import hash_to_bytes from swh.indexer.storage.exc import ( IndexerStorageArgumentException, DuplicateId, ) from swh.indexer.storage.interface import IndexerStorageInterface 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", + "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. """ 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, } ] ) 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]) 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], } ] 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]) 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]) 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,}] # 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) 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,}] # 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"], } for hash_ in hashes ] data_v2 = [ { "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 ] # 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"] ) expected_data_v2 = [ {"id": hash_, **self.example_data[1], "tool": tool,} for hash_ in hashes ] assert actual_data == expected_data_v2 def test_add__duplicate_twice(self, swh_indexer_storage_with_data): 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_rev2 = { "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,} ] 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"], } # 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,}] 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",}, ] - def test_generate_content_mimetype_get_range_limit_none(self, swh_indexer_storage): - """mimetype_get_range call with wrong limit input should fail""" + 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 - with pytest.raises(IndexerStorageArgumentException) as e: - storage.content_mimetype_get_range( - start=None, end=None, indexer_configuration_id=None, limit=None + 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 ) - assert e.value.args == ("limit should not be None",) - - def test_generate_content_mimetype_get_range_no_limit( + def test_generate_content_mimetype_get_partition_no_limit( self, swh_indexer_storage_with_data ): - """mimetype_get_range returns mimetypes within range provided""" + """get_partition should return result""" storage, data = swh_indexer_storage_with_data mimetypes = data.mimetypes - # All ids from the db - content_ids = sorted([c["id"] for c in mimetypes]) - - start = content_ids[0] - end = content_ids[-1] + expected_ids = set([c["id"] for c in mimetypes]) + indexer_configuration_id = mimetypes[0]["indexer_configuration_id"] - # retrieve mimetypes - tool_id = mimetypes[0]["indexer_configuration_id"] - actual_result = storage.content_mimetype_get_range( - start, end, indexer_configuration_id=tool_id - ) + assert len(mimetypes) == 16 + nb_partitions = 16 - actual_ids = actual_result["ids"] - actual_next = actual_result["next"] + 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(mimetypes) == len(actual_ids) - assert actual_next is None - assert content_ids == actual_ids + 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_range_limit( + def test_generate_content_mimetype_get_partition_full( self, swh_indexer_storage_with_data ): - """mimetype_get_range paginates results if limit exceeded""" - storage, data = swh_indexer_storage_with_data + """get_partition for a single partition should return available ids - indexer_configuration_id = data.tools["file"]["id"] - - # input the list of sha1s we want from storage - content_ids = sorted([c["id"] for c in data.mimetypes]) - mimetypes = list(storage.content_mimetype_get(content_ids)) - assert len(mimetypes) == len(data.mimetypes) + """ + 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"] - start = content_ids[0] - end = content_ids[-1] - # retrieve mimetypes limited to 10 results - actual_result = storage.content_mimetype_get_range( - start, end, indexer_configuration_id=indexer_configuration_id, limit=10 + 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 - assert actual_result - assert set(actual_result.keys()) == {"ids", "next"} - actual_ids = actual_result["ids"] - actual_next = actual_result["next"] + 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"] + + # 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, + ) - assert len(actual_ids) == 10 - assert actual_next is not None - assert actual_next == content_ids[10] + for actual_id in actual_result.results: + seen_ids.append(actual_id) - expected_mimetypes = content_ids[:10] - assert expected_mimetypes == actual_ids + # Limit is higher than the max number of results + assert actual_result.next_page_token is None - # retrieve next part - actual_result = storage.content_mimetype_get_range( - start=end, end=end, indexer_configuration_id=indexer_configuration_id - ) - assert set(actual_result.keys()) == {"ids", "next"} - actual_ids = actual_result["ids"] - actual_next = actual_result["next"] + assert set(seen_ids) == expected_ids - assert actual_next is None - expected_mimetypes = [content_ids[-1]] - assert expected_mimetypes == actual_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"] + + 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 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 TestIndexerStorageContentFossologyLicence: 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_range_limit_none( + 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 - """license_get_range call with wrong limit input should fail""" - with pytest.raises(IndexerStorageArgumentException) as e: - storage.content_fossology_license_get_range( - start=None, end=None, indexer_configuration_id=None, limit=None + 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, ) - assert e.value.args == ("limit should not be None",) - - def test_generate_content_fossology_license_get_range_no_limit( + def test_generate_content_fossology_license_get_partition_no_limit( self, swh_indexer_storage_with_data ): - """license_get_range returns licenses within range provided""" + """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 - content_ids = sorted([c["id"] for c in fossology_licenses]) + expected_ids = set([c["id"] for c in fossology_licenses]) - start = content_ids[0] - end = content_ids[-1] + assert len(fossology_licenses) == 10 + assert len(mimetypes) == 10 + nb_partitions = 4 - # retrieve fossology_licenses - tool_id = fossology_licenses[0]["indexer_configuration_id"] - actual_result = storage.content_fossology_license_get_range( - start, end, indexer_configuration_id=tool_id - ) + actual_ids = [] + for partition_id in range(nb_partitions): - actual_ids = actual_result["ids"] - actual_next = actual_result["next"] + 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(fossology_licenses) == len(actual_ids) - assert actual_next is None - assert content_ids == actual_ids + 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_range_no_limit_with_filter( + def test_generate_content_fossology_license_get_partition_full( self, swh_indexer_storage_with_data ): - """This filters non textual, then returns results within range""" - storage, data = swh_indexer_storage_with_data - fossology_licenses = data.fossology_licenses - mimetypes = data.mimetypes + """get_partition for a single partition should return available ids + """ + storage, data = swh_indexer_storage_with_data # craft some consistent mimetypes - _mimetypes = prepare_mimetypes_from(fossology_licenses) - # add binary mimetypes which will get filtered out in results - for m in mimetypes: - _mimetypes.append( - {"mimetype": "binary", **m,} - ) + 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) + 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 - content_ids = sorted([c["id"] for c in fossology_licenses]) + expected_ids = set([c["id"] for c in fossology_licenses]) - start = content_ids[0] - end = content_ids[-1] - - # retrieve fossology_licenses - tool_id = fossology_licenses[0]["indexer_configuration_id"] - actual_result = storage.content_fossology_license_get_range( - start, end, indexer_configuration_id=tool_id + 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 - actual_ids = actual_result["ids"] - actual_next = actual_result["next"] - - assert len(fossology_licenses) == len(actual_ids) - assert actual_next is None - assert content_ids == actual_ids - - def test_generate_fossology_license_get_range_limit( + def test_generate_content_fossology_license_get_partition_empty( self, swh_indexer_storage_with_data ): - """fossology_license_get_range paginates results if limit exceeded""" + """get_partition when at least one of the partitions is empty""" storage, data = swh_indexer_storage_with_data - fossology_licenses = data.fossology_licenses - # craft some consistent mimetypes + fossology_licenses = data.fossology_licenses mimetypes = prepare_mimetypes_from(fossology_licenses) + indexer_configuration_id = fossology_licenses[0]["indexer_configuration_id"] - # add fossology_licenses to storage storage.content_mimetype_add(mimetypes, conflict_update=True) + # add fossology_licenses to storage storage.content_fossology_license_add(fossology_licenses) - # input the list of sha1s we want from storage - content_ids = sorted([c["id"] for c in fossology_licenses]) - start = content_ids[0] - end = content_ids[-1] + # All ids from the db + expected_ids = set([c["id"] for c in fossology_licenses]) - # retrieve fossology_licenses limited to 3 results - limited_results = len(fossology_licenses) - 1 - tool_id = fossology_licenses[0]["indexer_configuration_id"] - actual_result = storage.content_fossology_license_get_range( - start, end, indexer_configuration_id=tool_id, limit=limited_results - ) + # 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) - actual_ids = actual_result["ids"] - actual_next = actual_result["next"] + seen_ids = [] - assert limited_results == len(actual_ids) - assert actual_next is not None - assert actual_next == content_ids[-1] + 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, + ) - expected_fossology_licenses = content_ids[:-1] - assert expected_fossology_licenses == actual_ids + for actual_id in actual_result.results: + seen_ids.append(actual_id) - # retrieve next part - actual_results2 = storage.content_fossology_license_get_range( - start=end, end=end, indexer_configuration_id=tool_id - ) - actual_ids2 = actual_results2["ids"] - actual_next2 = actual_results2["next"] + # Limit is higher than the max number of results + assert actual_result.next_page_token is None - assert actual_next2 is None - expected_fossology_licenses2 = [content_ids[-1]] - assert expected_fossology_licenses2 == actual_ids2 + 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 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 5ee22e3..6cd0d9d 100644 --- a/swh/indexer/tests/test_fossology_license.py +++ b/swh/indexer/tests/test_fossology_license.py @@ -1,170 +1,148 @@ # Copyright (C) 2017-2018 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import unittest import pytest from unittest.mock import patch from typing import Any, Dict from swh.indexer import fossology_license from swh.indexer.fossology_license import ( FossologyLicenseIndexer, - FossologyLicenseRangeIndexer, + FossologyLicensePartitionIndexer, compute_license, ) from swh.indexer.tests.utils import ( SHA1_TO_LICENSES, CommonContentIndexerTest, - CommonContentIndexerRangeTest, + CommonContentIndexerPartitionTest, BASE_TEST_CONFIG, fill_storage, fill_obj_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: {"tool": tool, "licenses": SHA1_TO_LICENSES[self.id2],}, } def tearDown(self): super().tearDown() fossology_license.compute_license = self.orig_compute_license -class TestFossologyLicenseRangeIndexer( - CommonContentIndexerRangeTest, unittest.TestCase +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 = FossologyLicenseRangeIndexer(config=RANGE_CONFIG) + self.indexer = FossologyLicensePartitionIndexer(config=RANGE_CONFIG) self.indexer.catch_exceptions = False fill_storage(self.indexer.storage) fill_obj_storage(self.indexer.objstorage) - self.id0 = "01c9379dfc33803963d07c1ccc748d3fe4c96bb5" - self.id1 = "02fb2c89e14f7fab46701478c83779c7beb7b069" - self.id2 = "103bc087db1d26afc3a0283f38663d081e9b01e6" - tool_id = self.indexer.tool["id"] - self.expected_results = { - self.id0: { - "id": self.id0, - "indexer_configuration_id": tool_id, - "licenses": SHA1_TO_LICENSES[self.id0], - }, - self.id1: { - "id": self.id1, - "indexer_configuration_id": tool_id, - "licenses": SHA1_TO_LICENSES[self.id1], - }, - self.id2: { - "id": self.id2, - "indexer_configuration_id": tool_id, - "licenses": SHA1_TO_LICENSES[self.id2], - }, - } - def tearDown(self): super().tearDown() fossology_license.compute_license = self.orig_compute_license 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): - FossologyLicenseRangeIndexer(config=filter_dict(RANGE_CONFIG, "tools")) + FossologyLicensePartitionIndexer(config=filter_dict(RANGE_CONFIG, "tools")) diff --git a/swh/indexer/tests/test_mimetype.py b/swh/indexer/tests/test_mimetype.py index 9a9e3b1..483743f 100644 --- a/swh/indexer/tests/test_mimetype.py +++ b/swh/indexer/tests/test_mimetype.py @@ -1,150 +1,126 @@ # 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 pytest import unittest from typing import Any, Dict from swh.indexer.mimetype import ( MimetypeIndexer, - MimetypeRangeIndexer, + MimetypePartitionIndexer, compute_mimetype_encoding, ) from swh.indexer.tests.utils import ( CommonContentIndexerTest, - CommonContentIndexerRangeTest, + CommonContentIndexerPartitionTest, BASE_TEST_CONFIG, fill_storage, fill_obj_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) 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 TestMimetypeRangeIndexer(CommonContentIndexerRangeTest, unittest.TestCase): +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 """ def setUp(self): super().setUp() - self.indexer = MimetypeRangeIndexer(config=RANGE_CONFIG) + self.indexer = MimetypePartitionIndexer(config=RANGE_CONFIG) self.indexer.catch_exceptions = False fill_storage(self.indexer.storage) fill_obj_storage(self.indexer.objstorage) - self.id0 = "01c9379dfc33803963d07c1ccc748d3fe4c96bb5" - self.id1 = "02fb2c89e14f7fab46701478c83779c7beb7b069" - self.id2 = "103bc087db1d26afc3a0283f38663d081e9b01e6" - tool_id = self.indexer.tool["id"] - - self.expected_results = { - self.id0: { - "encoding": "us-ascii", - "id": self.id0, - "indexer_configuration_id": tool_id, - "mimetype": "text/plain", - }, - self.id1: { - "encoding": "us-ascii", - "id": self.id1, - "indexer_configuration_id": tool_id, - "mimetype": "text/x-python", - }, - self.id2: { - "encoding": "us-ascii", - "id": self.id2, - "indexer_configuration_id": tool_id, - "mimetype": "text/plain", - }, - } - 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): - MimetypeRangeIndexer(config=filter_dict(CONFIG, "tools")) + MimetypePartitionIndexer(config=filter_dict(CONFIG, "tools")) diff --git a/swh/indexer/tests/test_tasks.py b/swh/indexer/tests/test_tasks.py new file mode 100644 index 0000000..1058f10 --- /dev/null +++ b/swh/indexer/tests/test_tasks.py @@ -0,0 +1,123 @@ +# 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 + + +def test_task_origin_metadata( + mocker, swh_scheduler_celery_app, swh_scheduler_celery_worker, swh_config +): + + mock_indexer = mocker.patch("swh.indexer.tasks.OriginMetadataIndexer.run") + mock_indexer.return_value = {"status": "eventful"} + + res = swh_scheduler_celery_app.send_task( + "swh.indexer.tasks.OriginMetadata", args=["origin-url"], + ) + assert res + res.wait() + assert res.successful() + + assert res.result == {"status": "eventful"} + + +def test_task_ctags( + mocker, swh_scheduler_celery_app, swh_scheduler_celery_worker, swh_config +): + + mock_indexer = mocker.patch("swh.indexer.tasks.CtagsIndexer.run") + mock_indexer.return_value = {"status": "eventful"} + + res = swh_scheduler_celery_app.send_task("swh.indexer.tasks.Ctags", args=["id0"],) + assert res + res.wait() + assert res.successful() + + assert res.result == {"status": "eventful"} + + +def test_task_fossology_license( + mocker, swh_scheduler_celery_app, swh_scheduler_celery_worker, swh_config +): + + mock_indexer = mocker.patch("swh.indexer.tasks.FossologyLicenseIndexer.run") + mock_indexer.return_value = {"status": "eventful"} + + res = swh_scheduler_celery_app.send_task( + "swh.indexer.tasks.ContentFossologyLicense", args=["id0"], + ) + assert res + res.wait() + assert res.successful() + + assert res.result == {"status": "eventful"} + + +def test_task_recompute_checksums( + mocker, swh_scheduler_celery_app, swh_scheduler_celery_worker, swh_config +): + + mock_indexer = mocker.patch("swh.indexer.tasks.RecomputeChecksums.run") + mock_indexer.return_value = {"status": "eventful"} + + res = swh_scheduler_celery_app.send_task( + "swh.indexer.tasks.RecomputeChecksums", args=[[{"blake2b256": "id"}]], + ) + assert res + res.wait() + assert res.successful() + + assert res.result == {"status": "eventful"} + + +def test_task_mimetype( + mocker, swh_scheduler_celery_app, swh_scheduler_celery_worker, swh_config +): + + mock_indexer = mocker.patch("swh.indexer.tasks.MimetypeIndexer.run") + mock_indexer.return_value = {"status": "eventful"} + + res = swh_scheduler_celery_app.send_task( + "swh.indexer.tasks.ContentMimetype", args=["id0"], + ) + assert res + res.wait() + assert res.successful() + + assert res.result == {"status": "eventful"} + + +def test_task_mimetype_partition( + mocker, swh_scheduler_celery_app, swh_scheduler_celery_worker, swh_config +): + + mock_indexer = mocker.patch("swh.indexer.tasks.MimetypePartitionIndexer.run") + mock_indexer.return_value = {"status": "eventful"} + + res = swh_scheduler_celery_app.send_task( + "swh.indexer.tasks.ContentMimetypePartition", args=[0, 4], + ) + assert res + res.wait() + assert res.successful() + + assert res.result == {"status": "eventful"} + + +def test_task_license_partition( + mocker, swh_scheduler_celery_app, swh_scheduler_celery_worker, swh_config +): + + mock_indexer = mocker.patch( + "swh.indexer.tasks.FossologyLicensePartitionIndexer.run" + ) + mock_indexer.return_value = {"status": "eventful"} + + res = swh_scheduler_celery_app.send_task( + "swh.indexer.tasks.ContentFossologyLicensePartition", args=[0, 4], + ) + assert res + res.wait() + assert res.successful() + + assert res.result == {"status": "eventful"} diff --git a/swh/indexer/tests/utils.py b/swh/indexer/tests/utils.py index b3f0612..04a34db 100644 --- a/swh/indexer/tests/utils.py +++ b/swh/indexer/tests/utils.py @@ -1,774 +1,770 @@ # 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 Dict, Any import unittest from hypothesis import strategies +from swh.core.api.classes import stream_results from swh.model import hashutil -from swh.model.hashutil import hash_to_bytes, hash_to_hex +from swh.model.hashutil import hash_to_bytes from swh.model.model import ( Content, Directory, DirectoryEntry, Origin, OriginVisit, OriginVisitStatus, Person, Revision, RevisionType, + SHA1_SIZE, Snapshot, SnapshotBranch, TargetType, Timestamp, TimestampWithTimezone, ) -from swh.storage.utils import now +from swh.storage.utils import now, get_partition_bounds_bytes from swh.indexer.storage import INDEXER_CFG_KEY 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( len(expected_results), len(actual_results), (expected_results, actual_results), ) for indexed_data in actual_results: (_id, indexed_data) = list(indexed_data.items())[0] 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 CommonContentIndexerRangeTest: +class CommonContentIndexerPartitionTest: """Allows to factorize tests on range indexer. """ def setUp(self): self.contents = sorted(OBJ_STORAGE_DATA) - def assert_results_ok(self, start, end, actual_results, expected_results=None): - if expected_results is None: - expected_results = self.expected_results + 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, + ) + ] + + start, end = get_partition_bounds_bytes(partition_id, nb_partitions, SHA1_SIZE) actual_results = list(actual_results) for indexed_data in actual_results: _id = indexed_data["id"] assert isinstance(_id, bytes) - indexed_data = indexed_data.copy() - indexed_data["id"] = hash_to_hex(indexed_data["id"]) - self.assertEqual(indexed_data, expected_results[hash_to_hex(_id)]) - self.assertTrue(start <= _id <= end) + assert _id in expected_ids + + assert start <= _id + if end: + assert _id <= end + _tool_id = indexed_data["indexer_configuration_id"] - self.assertEqual(_tool_id, self.indexer.tool["id"]) + assert _tool_id == self.indexer.tool["id"] def test__index_contents(self): """Indexing contents without existing data results in indexed data """ - _start, _end = [self.contents[0], self.contents[2]] # output hex ids - start, end = map(hashutil.hash_to_bytes, (_start, _end)) - # given - actual_results = list(self.indexer._index_contents(start, end, indexed={})) + partition_id = 0 + nb_partitions = 4 + + actual_results = list( + self.indexer._index_contents(partition_id, nb_partitions, indexed={}) + ) - self.assert_results_ok(start, end, actual_results) + 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 """ - _start, _end = [self.contents[0], self.contents[2]] # output hex ids - start, end = map(hashutil.hash_to_bytes, (_start, _end)) - data_indexed = [self.id0, self.id2] + partition_id = 3 + nb_partitions = 4 - # given - actual_results = self.indexer._index_contents( - start, end, indexed=set(map(hash_to_bytes, data_indexed)) + # first pass + actual_results = list( + self.indexer._index_contents(partition_id, nb_partitions, indexed={}) ) - # craft the expected results - expected_results = self.expected_results.copy() - for already_indexed_key in data_indexed: - expected_results.pop(already_indexed_key) + self.assert_results_ok(partition_id, nb_partitions, actual_results) - self.assert_results_ok(start, end, actual_results, expected_results) - - def test_generate_content_get(self): - """Optimal indexing should result in indexed data + indexed_ids = set(res["id"] for res in actual_results) - """ - _start, _end = [self.contents[0], self.contents[2]] # output hex ids - start, end = map(hashutil.hash_to_bytes, (_start, _end)) - - # given - actual_results = self.indexer.run(start, end) + actual_results = list( + self.indexer._index_contents( + partition_id, nb_partitions, indexed=indexed_ids + ) + ) - # then - self.assertEqual(actual_results, {"status": "uneventful"}) + # already indexed, so nothing new + assert actual_results == [] - def test_generate_content_get_input_as_bytes(self): + def test_generate_content_get(self): """Optimal indexing should result in indexed data - Input are in bytes here. - """ - _start, _end = [self.contents[0], self.contents[2]] # output hex ids - start, end = map(hashutil.hash_to_bytes, (_start, _end)) + partition_id = 0 + nb_partitions = 4 - # given - actual_results = self.indexer.run(start, end, skip_existing=False) - # no already indexed data so same result as prior test + actual_results = self.indexer.run( + partition_id, nb_partitions, skip_existing=False + ) - # then - self.assertEqual(actual_results, {"status": "uneventful"}) + assert actual_results == {"status": "uneventful"} # why? def test_generate_content_get_no_result(self): """No result indexed returns False""" - _start, _end = [ - "0000000000000000000000000000000000000000", - "0000000000000000000000000000000000000001", - ] - start, end = map(hashutil.hash_to_bytes, (_start, _end)) - # given - actual_results = self.indexer.run(start, end, incremental=False) + actual_results = self.indexer.run(0, 0, incremental=False) - # then - self.assertEqual(actual_results, {"status": "uneventful"}) + assert actual_results == {"status": "uneventful"}