diff --git a/swh/loader/core/discovery.py b/swh/loader/core/discovery.py index 794e356..66bf265 100644 --- a/swh/loader/core/discovery.py +++ b/swh/loader/core/discovery.py @@ -1,207 +1,261 @@ # Copyright (C) 2022 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 """Primitives for finding the unknown parts of disk contents efficiently.""" +import abc from collections import namedtuple import itertools import logging import random from typing import Any, Iterable, List, Mapping, NamedTuple, Set, Union from swh.model.from_disk import model from swh.model.model import Sha1Git from swh.storage.interface import StorageInterface logger = logging.getLogger(__name__) # Maximum amount when sampling from the undecided set of directory entries SAMPLE_SIZE = 1000 # Sets of sha1 of contents, skipped contents and directories respectively Sample: NamedTuple = namedtuple( "Sample", ["contents", "skipped_contents", "directories"] ) +class ArchiveDiscoveryInterface(abc.ABC): + """Interface used in discovery code to abstract over ways of connecting to + the SWH archive (direct storage, web API, etc.) for all methods needed by + discovery algorithms.""" + + contents: List[model.Content] + skipped_contents: List[model.SkippedContent] + directories: List[model.Directory] + + def __init__( + self, + contents: List[model.Content], + skipped_contents: List[model.SkippedContent], + directories: List[model.Directory], + ) -> None: + self.contents = contents + self.skipped_contents = skipped_contents + self.directories = directories + + @abc.abstractmethod + async def content_missing(self, contents: List[Sha1Git]) -> Iterable[Sha1Git]: + """List content missing from the archive by sha1""" + + @abc.abstractmethod + async def skipped_content_missing( + self, skipped_contents: List[Sha1Git] + ) -> Iterable[Sha1Git]: + """List skipped content missing from the archive by sha1""" + + @abc.abstractmethod + async def directory_missing(self, directories: List[Sha1Git]) -> Iterable[Sha1Git]: + """List directories missing from the archive by sha1""" + + +class DiscoveryStorageConnection(ArchiveDiscoveryInterface): + """Use the storage APIs to query the archive""" + + def __init__( + self, + contents: List[model.Content], + skipped_contents: List[model.SkippedContent], + directories: List[model.Directory], + swh_storage: StorageInterface, + ) -> None: + super().__init__(contents, skipped_contents, directories) + self.storage = swh_storage + + async def content_missing(self, contents: List[Sha1Git]) -> Iterable[Sha1Git]: + """List content missing from the archive by sha1""" + return self.storage.content_missing_per_sha1_git(contents) + + async def skipped_content_missing( + self, skipped_contents: List[Sha1Git] + ) -> Iterable[Sha1Git]: + """List skipped content missing from the archive by sha1""" + contents = [ + {"sha1_git": s, "sha1": None, "sha256": None, "blake2s256": None} + for s in skipped_contents + ] + return (d["sha1_git"] for d in self.storage.skipped_content_missing(contents)) + + async def directory_missing(self, directories: List[Sha1Git]) -> Iterable[Sha1Git]: + """List directories missing from the archive by sha1""" + return self.storage.directory_missing(directories) + + class BaseDiscoveryGraph: """Creates the base structures and methods needed for discovery algorithms. Subclasses should override ``get_sample`` to affect how the discovery is made.""" def __init__(self, contents, skipped_contents, directories): self._all_contents: Mapping[ Sha1Git, Union[model.Content, model.SkippedContent] ] = {} self._undecided_directories: Set[Sha1Git] = set() self._children: Mapping[Sha1Git, model.DirectoryEntry] = {} self._parents: Mapping[model.DirectoryEntry, Sha1Git] = {} self.undecided: Set[Sha1Git] = set() for content in itertools.chain(contents, skipped_contents): self.undecided.add(content.sha1_git) self._all_contents[content.sha1_git] = content for directory in directories: self.undecided.add(directory.id) self._undecided_directories.add(directory.id) self._children[directory.id] = {c.target for c in directory.entries} for child in directory.entries: self._parents.setdefault(child.target, set()).add(directory.id) self.undecided |= self._undecided_directories self.known: Set[Sha1Git] = set() self.unknown: Set[Sha1Git] = set() def mark_known(self, entries: Iterable[Sha1Git]): """Mark ``entries`` and those they imply as known in the SWH archive""" self._mark_entries(entries, self._children, self.known) def mark_unknown(self, entries: Iterable[Sha1Git]): """Mark ``entries`` and those they imply as unknown in the SWH archive""" self._mark_entries(entries, self._parents, self.unknown) def _mark_entries( self, entries: Iterable[Sha1Git], transitive_mapping: Mapping[Any, Any], target_set: Set[Any], ): """Use Merkle graph properties to mark a directory entry as known or unknown. If an entry is known, then all of its descendants are known. If it's unknown, then all of its ancestors are unknown. - ``entries``: directory entries to mark along with their ancestors/descendants where applicable. - ``transitive_mapping``: mapping from an entry to the next entries to mark in the hierarchy, if any. - ``target_set``: set where marked entries will be added. """ to_process = set(entries) while to_process: current = to_process.pop() target_set.add(current) self.undecided.discard(current) self._undecided_directories.discard(current) next_entries = transitive_mapping.get(current, set()) & self.undecided to_process.update(next_entries) - def get_sample( + async def get_sample( self, ) -> Sample: """Return a three-tuple of samples from the undecided sets of contents, skipped contents and directories respectively. These samples will be queried against the storage which will tell us which are known.""" raise NotImplementedError() - def do_query(self, swh_storage: StorageInterface, sample: Sample) -> None: - """Given a three-tuple of samples, ask the storage which are known or + async def do_query( + self, archive: ArchiveDiscoveryInterface, sample: Sample + ) -> None: + """Given a three-tuple of samples, ask the archive which are known or unknown and mark them as such.""" - contents_sample, skipped_contents_sample, directories_sample = sample - # TODO unify those APIs, and make it so only have to manipulate hashes - if contents_sample: - known = set(contents_sample) - unknown = set(swh_storage.content_missing_per_sha1_git(contents_sample)) - known -= unknown - - self.mark_known(known) - self.mark_unknown(unknown) - - if skipped_contents_sample: - known = set(skipped_contents_sample) - as_dicts = [ - self._all_contents[s].to_dict() for s in skipped_contents_sample - ] - unknown = { - d["sha1_git"] for d in swh_storage.skipped_content_missing(as_dicts) - } - known -= unknown - - self.mark_known(known) - self.mark_unknown(unknown) + methods = ( + archive.content_missing, + archive.skipped_content_missing, + archive.directory_missing, + ) - if directories_sample: - known = set(directories_sample) - unknown = set(swh_storage.directory_missing(directories_sample)) + for sample_per_type, method in zip(sample, methods): + if not sample_per_type: + continue + known = set(sample_per_type) + unknown = set(await method(sample_per_type)) known -= unknown self.mark_known(known) self.mark_unknown(unknown) class RandomDirSamplingDiscoveryGraph(BaseDiscoveryGraph): """Use a random sampling using only directories. This allows us to find a statistically good spread of entries in the graph with a smaller population than using all types of entries. When there are no more directories, only contents or skipped contents are undecided if any are left: we send them directly to the storage since they should be few and their structure flat.""" - def get_sample(self) -> Sample: + async def get_sample(self) -> Sample: if self._undecided_directories: if len(self._undecided_directories) <= SAMPLE_SIZE: return Sample( contents=set(), skipped_contents=set(), directories=set(self._undecided_directories), ) sample = random.sample(self._undecided_directories, SAMPLE_SIZE) directories = {o for o in sample} return Sample( contents=set(), skipped_contents=set(), directories=directories ) contents = set() skipped_contents = set() for sha1 in self.undecided: obj = self._all_contents[sha1] obj_type = obj.object_type if obj_type == model.Content.object_type: contents.add(sha1) elif obj_type == model.SkippedContent.object_type: skipped_contents.add(sha1) else: raise TypeError(f"Unexpected object type {obj_type}") return Sample( contents=contents, skipped_contents=skipped_contents, directories=set() ) -def filter_known_objects( - swh_storage: StorageInterface, - contents: List[model.Content], - skipped_contents: List[model.SkippedContent], - directories: List[model.Directory], -): - """Filter ``contents``, ``skipped_contents`` and ``directories`` to only return - those that are unknown to the SWH archive using a discovery algorithm.""" +async def filter_known_objects(archive: ArchiveDiscoveryInterface): + """Filter ``archive``'s ``contents``, ``skipped_contents`` and ``directories`` + to only return those that are unknown to the SWH archive using a discovery + algorithm.""" + contents = archive.contents + skipped_contents = archive.skipped_contents + directories = archive.directories + contents_count = len(contents) skipped_contents_count = len(skipped_contents) directories_count = len(directories) graph = RandomDirSamplingDiscoveryGraph(contents, skipped_contents, directories) while graph.undecided: - sample = graph.get_sample() - graph.do_query(swh_storage, sample) + sample = await graph.get_sample() + await graph.do_query(archive, sample) contents = [c for c in contents if c.sha1_git in graph.unknown] skipped_contents = [c for c in skipped_contents if c.sha1_git in graph.unknown] directories = [c for c in directories if c.id in graph.unknown] logger.debug( "Filtered out %d contents, %d skipped contents and %d directories", contents_count - len(contents), skipped_contents_count - len(skipped_contents), directories_count - len(directories), ) return (contents, skipped_contents, directories) diff --git a/swh/loader/package/loader.py b/swh/loader/package/loader.py index 52a083d..4d5add9 100644 --- a/swh/loader/package/loader.py +++ b/swh/loader/package/loader.py @@ -1,1118 +1,1123 @@ # Copyright (C) 2019-2021 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 asyncio import datetime import hashlib from itertools import islice import json import logging import os import string import sys import tempfile from typing import ( Any, Dict, Generic, Iterator, List, Mapping, Optional, Sequence, Set, Tuple, TypeVar, ) import attr from requests.exceptions import ContentDecodingError import sentry_sdk from swh.core.tarball import uncompress from swh.loader.core import discovery from swh.loader.core.loader import BaseLoader from swh.loader.exception import NotFound from swh.loader.package.utils import download from swh.model import from_disk from swh.model.hashutil import hash_to_hex from swh.model.model import ( ExtID, MetadataAuthority, MetadataAuthorityType, MetadataFetcher, ) from swh.model.model import ( Origin, OriginVisit, OriginVisitStatus, RawExtrinsicMetadata, Release, Revision, Sha1Git, Snapshot, ) from swh.model.model import ObjectType as ModelObjectType from swh.model.swhids import CoreSWHID, ExtendedObjectType, ExtendedSWHID, ObjectType from swh.storage.algos.snapshot import snapshot_get_latest from swh.storage.interface import StorageInterface from swh.storage.utils import now logger = logging.getLogger(__name__) SWH_METADATA_AUTHORITY = MetadataAuthority( type=MetadataAuthorityType.REGISTRY, url="https://softwareheritage.org/", metadata={}, ) """Metadata authority for extrinsic metadata generated by Software Heritage. Used for metadata on "original artifacts", ie. length, filename, and checksums of downloaded archive files.""" PartialExtID = Tuple[str, int, bytes] """The ``extid_type`` and ``extid`` fields of an :class:`ExtID` object.""" @attr.s class RawExtrinsicMetadataCore: """Contains the core of the metadata extracted by a loader, that will be used to build a full RawExtrinsicMetadata object by adding object identifier, context, and provenance information.""" format = attr.ib(type=str) metadata = attr.ib(type=bytes) discovery_date = attr.ib(type=Optional[datetime.datetime], default=None) """Defaults to the visit date.""" @attr.s class BasePackageInfo: """Compute the primary key for a dict using the id_keys as primary key composite. Args: d: A dict entry to compute the primary key on id_keys: Sequence of keys to use as primary key Returns: The identity for that dict entry """ url = attr.ib(type=str) filename = attr.ib(type=Optional[str]) version = attr.ib(type=str) """Version name/number.""" MANIFEST_FORMAT: Optional[string.Template] = None """If not None, used by the default extid() implementation to format a manifest, before hashing it to produce an ExtID.""" EXTID_TYPE: str = "package-manifest-sha256" EXTID_VERSION: int = 0 # The following attribute has kw_only=True in order to allow subclasses # to add attributes. Without kw_only, attributes without default values cannot # go after attributes with default values. # See directory_extrinsic_metadata = attr.ib( type=List[RawExtrinsicMetadataCore], default=[], kw_only=True, ) """:term:`extrinsic metadata` collected by the loader, that will be attached to the loaded directory and added to the Metadata storage.""" # TODO: add support for metadata for releases and contents def extid(self) -> Optional[PartialExtID]: """Returns a unique intrinsic identifier of this package info, or None if this package info is not 'deduplicatable' (meaning that we will always load it, instead of checking the ExtID storage to see if we already did)""" if self.MANIFEST_FORMAT is None: return None else: manifest = self.MANIFEST_FORMAT.substitute( {k: str(v) for (k, v) in attr.asdict(self).items()} ) return ( self.EXTID_TYPE, self.EXTID_VERSION, hashlib.sha256(manifest.encode()).digest(), ) TPackageInfo = TypeVar("TPackageInfo", bound=BasePackageInfo) class PackageLoader(BaseLoader, Generic[TPackageInfo]): def __init__(self, storage: StorageInterface, url: str, **kwargs: Any): """Loader's constructor. This raises exception if the minimal required configuration is missing (cf. fn:`check` method). Args: storage: Storage instance url: Origin url to load data from """ super().__init__(storage=storage, origin_url=url, **kwargs) self.status_load = "" self.status_visit = "" def load_status(self) -> Dict[str, str]: """Detailed loading status.""" return { "status": self.status_load, } def visit_status(self) -> str: """Detailed visit status.""" return self.status_visit def get_versions(self) -> Sequence[str]: """Return the list of all published package versions. Raises: class:`swh.loader.exception.NotFound` error when failing to read the published package versions. Returns: Sequence of published versions """ return [] def get_package_info(self, version: str) -> Iterator[Tuple[str, TPackageInfo]]: """Given a release version of a package, retrieve the associated package information for such version. Args: version: Package version Returns: (branch name, package metadata) """ yield from {} def build_release( self, p_info: TPackageInfo, uncompressed_path: str, directory: Sha1Git ) -> Optional[Release]: """Build the release from the archive metadata (extrinsic artifact metadata) and the intrinsic metadata. Args: p_info: Package information uncompressed_path: Artifact uncompressed path on disk """ raise NotImplementedError("build_release") def get_default_version(self) -> str: """Retrieve the latest release version if any. Returns: Latest version """ return "" def last_snapshot(self) -> Optional[Snapshot]: """Retrieve the last snapshot out of the last visit.""" return snapshot_get_latest(self.storage, self.origin.url) def new_packageinfo_to_extid(self, p_info: TPackageInfo) -> Optional[PartialExtID]: return p_info.extid() def _get_known_extids( self, packages_info: List[TPackageInfo] ) -> Dict[PartialExtID, List[CoreSWHID]]: """Compute the ExtIDs from new PackageInfo objects, searches which are already loaded in the archive, and returns them if any.""" # Compute the ExtIDs of all the new packages, grouped by extid type new_extids: Dict[Tuple[str, int], List[bytes]] = {} for p_info in packages_info: res = p_info.extid() if res is not None: (extid_type, extid_version, extid_extid) = res new_extids.setdefault((extid_type, extid_version), []).append( extid_extid ) # For each extid type, call extid_get_from_extid() with all the extids of # that type, and store them in the '(type, extid) -> target' map. known_extids: Dict[PartialExtID, List[CoreSWHID]] = {} for ((extid_type, extid_version), extids) in new_extids.items(): for extid in self.storage.extid_get_from_extid( extid_type, extids, version=extid_version ): if extid is not None: key = (extid.extid_type, extid_version, extid.extid) known_extids.setdefault(key, []).append(extid.target) return known_extids def resolve_object_from_extids( self, known_extids: Dict[PartialExtID, List[CoreSWHID]], p_info: TPackageInfo, whitelist: Set[Sha1Git], ) -> Optional[CoreSWHID]: """Resolve the revision/release from known ExtIDs and a package info object. If the artifact has already been downloaded, this will return the existing release (or revision) targeting that uncompressed artifact directory. Otherwise, this returns None. Args: known_extids: Dict built from a list of ExtID, with the target as value p_info: Package information whitelist: Any ExtID with target not in this set is filtered out Returns: None or release/revision SWHID """ new_extid = p_info.extid() if new_extid is None: return None extid_targets = set() for extid_target in known_extids.get(new_extid, []): if extid_target.object_id not in whitelist: # There is a known ExtID for this package, but its target is not # in the snapshot. # This can happen for three reasons: # # 1. a loader crashed after writing the ExtID, but before writing # the snapshot # 2. some other loader loaded the same artifact, but produced # a different revision, causing an additional ExtID object # to be written. We will probably find this loader's ExtID # in a future iteration of this loop. # Note that for now, this is impossible, as each loader has a # completely different extid_type, but this is an implementation # detail of each loader. # 3. we took a snapshot, then the package disappeared, # then we took another snapshot, and the package reappeared # # In case of 1, we must actually load the package now, # so let's do it. # TODO: detect when we are in case 3 using release_missing # or revision_missing instead of the snapshot. continue elif extid_target.object_type in (ObjectType.RELEASE, ObjectType.REVISION): extid_targets.add(extid_target) else: # Note that this case should never be reached unless there is a # collision between a revision hash and some non-revision object's # hash, but better safe than sorry. logger.warning( "%s is in the whitelist, but is not a revision/release.", hash_to_hex(extid_target.object_type), ) if extid_targets: # This is a known package version, as we have an extid to reference it. # Let's return one of them. # If there is a release extid, return it. release_extid_targets = { extid_target for extid_target in extid_targets if extid_target.object_type == ObjectType.RELEASE } # Exclude missing targets missing_releases = { CoreSWHID(object_type=ObjectType.RELEASE, object_id=id_) for id_ in self.storage.release_missing( [swhid.object_id for swhid in release_extid_targets] ) } if missing_releases: err_message = "Found ExtIDs pointing to missing releases" logger.error(err_message + ": %s", missing_releases) with sentry_sdk.push_scope() as scope: scope.set_extra( "missing_releases", [str(x) for x in missing_releases] ) sentry_sdk.capture_message(err_message, "error") release_extid_targets -= missing_releases extid_target2 = self.select_extid_target(p_info, release_extid_targets) if extid_target2: return extid_target2 # If there is no release extid (ie. if the package was only loaded with # older versions of this loader, which produced revision objects instead # of releases), return a revision extid when possible. revision_extid_targets = { extid_target for extid_target in extid_targets if extid_target.object_type == ObjectType.REVISION } if revision_extid_targets: assert len(extid_targets) == 1, extid_targets extid_target = list(extid_targets)[0] return extid_target # No target found (this is probably a new package version) return None def select_extid_target( self, p_info: TPackageInfo, extid_targets: Set[CoreSWHID] ) -> Optional[CoreSWHID]: """Given a list of release extid targets, choses one appropriate for the given package info. Package loaders shyould implement this if their ExtIDs may map to multiple releases, so they can fetch releases from the storage and inspect their fields to select the right one for this ``p_info``. """ if extid_targets: # The base package loader does not have the domain-specific knowledge # to select the right release -> crash if there is more than one. assert len(extid_targets) == 1, extid_targets return list(extid_targets)[0] return None def download_package( self, p_info: TPackageInfo, tmpdir: str ) -> List[Tuple[str, Mapping]]: """Download artifacts for a specific package. All downloads happen in in the tmpdir folder. Default implementation expects the artifacts package info to be about one artifact per package. Note that most implementation have 1 artifact per package. But some implementation have multiple artifacts per package (debian), some have none, the package is the artifact (gnu). Args: artifacts_package_info: Information on the package artifacts to download (url, filename, etc...) tmpdir: Location to retrieve such artifacts Returns: List of (path, computed hashes) """ try: return [download(p_info.url, dest=tmpdir, filename=p_info.filename)] except ContentDecodingError: # package might be erroneously marked as gzip compressed while is is not, # try to download its raw bytes again without attempting to uncompress # the input stream return [ download( p_info.url, dest=tmpdir, filename=p_info.filename, extra_request_headers={"Accept-Encoding": "identity"}, ) ] def uncompress( self, dl_artifacts: List[Tuple[str, Mapping[str, Any]]], dest: str ) -> str: """Uncompress the artifact(s) in the destination folder dest. Optionally, this could need to use the p_info dict for some more information (debian). """ uncompressed_path = os.path.join(dest, "src") for a_path, _ in dl_artifacts: uncompress(a_path, dest=uncompressed_path) return uncompressed_path def extra_branches(self) -> Dict[bytes, Mapping[str, Any]]: """Return an extra dict of branches that are used to update the set of branches. """ return {} def finalize_visit( self, *, snapshot: Optional[Snapshot], visit: OriginVisit, status_visit: str, status_load: str, failed_branches: List[str], errors: Optional[List[str]] = None, ) -> Dict[str, Any]: """Finalize the visit: - flush eventual unflushed data to storage - update origin visit's status - return the task's status """ self.status_load = status_load self.status_visit = status_visit self.storage.flush() snapshot_id: Optional[bytes] = None if snapshot and snapshot.id: # to prevent the snapshot.id to b"" snapshot_id = snapshot.id assert visit.visit visit_status = OriginVisitStatus( origin=self.origin.url, visit=visit.visit, type=self.visit_type, date=now(), status=status_visit, snapshot=snapshot_id, ) self.storage.origin_visit_status_add([visit_status]) result: Dict[str, Any] = { "status": status_load, } if snapshot_id: result["snapshot_id"] = hash_to_hex(snapshot_id) if failed_branches: logger.warning("%d failed branches", len(failed_branches)) for i, urls in enumerate(islice(failed_branches, 50)): prefix_url = "Failed branches: " if i == 0 else "" logger.warning("%s%s", prefix_url, urls) return result def load(self) -> Dict: """Load for a specific origin the associated contents. 1. Get the list of versions in an origin. 2. Get the snapshot from the previous run of the loader, and filter out versions that were already loaded, if their :term:`extids ` match Then, for each remaining version in the origin 3. Fetch the files for one package version By default, this can be implemented as a simple HTTP request. Loaders with more specific requirements can override this, e.g.: the PyPI loader checks the integrity of the downloaded files; the Debian loader has to download and check several files for one package version. 4. Extract the downloaded files. By default, this would be a universal archive/tarball extraction. Loaders for specific formats can override this method (for instance, the Debian loader uses dpkg-source -x). 5. Convert the extracted directory to a set of Software Heritage objects Using swh.model.from_disk. 6. Extract the metadata from the unpacked directories This would only be applicable for "smart" loaders like npm (parsing the package.json), PyPI (parsing the PKG-INFO file) or Debian (parsing debian/changelog and debian/control). On "minimal-metadata" sources such as the GNU archive, the lister should provide the minimal set of metadata needed to populate the revision/release objects (authors, dates) as an argument to the task. 7. Generate the revision/release objects for the given version. From the data generated at steps 3 and 4. end for each 8. Generate and load the snapshot for the visit Using the revisions/releases collected at step 7., and the branch information from step 2., generate a snapshot and load it into the Software Heritage archive """ self.status_load = "uneventful" # either: eventful, uneventful, failed self.status_visit = "full" # see swh.model.model.OriginVisitStatus snapshot = None failed_branches: List[str] = [] # Prepare origin and origin_visit origin = Origin(url=self.origin.url) try: self.storage.origin_add([origin]) visit = list( self.storage.origin_visit_add( [ OriginVisit( origin=self.origin.url, date=self.visit_date, type=self.visit_type, ) ] ) )[0] except Exception as e: logger.exception( "Failed to initialize origin_visit for %s", self.origin.url ) sentry_sdk.capture_exception(e) self.status_load = self.status_visit = "failed" return {"status": "failed"} # Get the previous snapshot for this origin. It is then used to see which # of the package's versions are already loaded in the archive. try: last_snapshot = self.last_snapshot() logger.debug("last snapshot: %s", last_snapshot) except Exception as e: logger.exception("Failed to get previous state for %s", self.origin.url) sentry_sdk.capture_exception(e) return self.finalize_visit( snapshot=snapshot, visit=visit, failed_branches=failed_branches, status_visit="failed", status_load="failed", errors=[str(e)], ) load_exceptions: List[Exception] = [] # Get the list of all version names try: versions = self.get_versions() except NotFound as e: return self.finalize_visit( snapshot=snapshot, visit=visit, failed_branches=failed_branches, status_visit="not_found", status_load="failed", errors=[str(e)], ) except Exception as e: logger.exception("Failed to get list of versions for %s", self.origin.url) sentry_sdk.capture_exception(e) return self.finalize_visit( snapshot=snapshot, visit=visit, failed_branches=failed_branches, status_visit="failed", status_load="failed", errors=[str(e)], ) errors = [] # Get the metadata of each version's package packages_info: List[Tuple[str, TPackageInfo]] = [] for version in versions: try: for branch_name, p_info in self.get_package_info(version): packages_info.append((branch_name, p_info)) except Exception as e: load_exceptions.append(e) sentry_sdk.capture_exception(e) error = f"Failed to get package info for version {version} of {self.origin.url}" logger.exception(error) errors.append(f"{error}: {e}") # Compute the ExtID of each of these packages known_extids = self._get_known_extids([p_info for (_, p_info) in packages_info]) if last_snapshot is None: last_snapshot_targets: Set[Sha1Git] = set() else: last_snapshot_targets = { branch.target for branch in last_snapshot.branches.values() } new_extids: Set[ExtID] = set() tmp_releases: Dict[str, List[Tuple[str, Sha1Git]]] = { version: [] for version in versions } for (branch_name, p_info) in packages_info: logger.debug("package_info: %s", p_info) # Check if the package was already loaded, using its ExtID swhid = self.resolve_object_from_extids( known_extids, p_info, last_snapshot_targets ) if swhid is not None and swhid.object_type == ObjectType.REVISION: # This package was already loaded, but by an older version # of this loader, which produced revisions instead of releases. # Let's fetch the revision's data, and "upgrade" it into a release. (rev,) = self.storage.revision_get([swhid.object_id]) if not rev: logger.error( "Failed to upgrade branch %s from revision to " "release, %s is missing from the storage. " "Falling back to re-loading from the origin.", branch_name, swhid, ) else: rev = None if swhid is None or (swhid.object_type == ObjectType.REVISION and not rev): # No matching revision or release found in the last snapshot, load it. release_id = None try: res = self._load_release(p_info, origin) if res: (release_id, directory_id) = res assert release_id assert directory_id self._load_extrinsic_directory_metadata( p_info, release_id, directory_id ) self.storage.flush() self.status_load = "eventful" except Exception as e: self.storage.clear_buffers() load_exceptions.append(e) sentry_sdk.capture_exception(e) error = f"Failed to load branch {branch_name} for {self.origin.url}" logger.exception(error) failed_branches.append(branch_name) errors.append(f"{error}: {e}") continue if release_id is None: continue add_extid = True elif swhid.object_type == ObjectType.REVISION: # If 'rev' was None, the previous block would have run. assert rev is not None rel = rev2rel(rev, p_info.version) self.storage.release_add([rel]) logger.debug("Upgraded %s to %s", swhid, rel.swhid()) release_id = rel.id # Create a new extid for this package, so the next run of this loader # will be able to find the new release, and use it (instead of the # old revision) add_extid = True elif swhid.object_type == ObjectType.RELEASE: # This package was already loaded, nothing to do. release_id = swhid.object_id add_extid = False else: assert False, f"Unexpected object type: {swhid}" assert release_id is not None if add_extid: partial_extid = p_info.extid() if partial_extid is not None: (extid_type, extid_version, extid) = partial_extid release_swhid = CoreSWHID( object_type=ObjectType.RELEASE, object_id=release_id ) new_extids.add( ExtID( extid_type=extid_type, extid_version=extid_version, extid=extid, target=release_swhid, ) ) tmp_releases[p_info.version].append((branch_name, release_id)) if load_exceptions: self.status_visit = "partial" if not tmp_releases: # We could not load any releases; fail completely logger.error("Failed to load any release for %s", self.origin.url) return self.finalize_visit( snapshot=snapshot, visit=visit, failed_branches=failed_branches, status_visit="failed", status_load="failed", errors=errors, ) try: # Retrieve the default release version (the "latest" one) default_version = self.get_default_version() logger.debug("default version: %s", default_version) # Retrieve extra branches extra_branches = self.extra_branches() logger.debug("extra branches: %s", extra_branches) snapshot = self._load_snapshot( default_version, tmp_releases, extra_branches ) self.storage.flush() except Exception as e: error = f"Failed to build snapshot for origin {self.origin.url}" logger.exception(error) errors.append(f"{error}: {e}") sentry_sdk.capture_exception(e) self.status_visit = "failed" self.status_load = "failed" if snapshot: try: metadata_objects = self.build_extrinsic_snapshot_metadata(snapshot.id) self.load_metadata_objects(metadata_objects) except Exception as e: error = ( f"Failed to load extrinsic snapshot metadata for {self.origin.url}" ) logger.exception(error) errors.append(f"{error}: {e}") sentry_sdk.capture_exception(e) self.status_visit = "partial" self.status_load = "failed" try: metadata_objects = self.build_extrinsic_origin_metadata() self.load_metadata_objects(metadata_objects) except Exception as e: error = f"Failed to load extrinsic origin metadata for {self.origin.url}" logger.exception(error) errors.append(f"{error}: {e}") sentry_sdk.capture_exception(e) self.status_visit = "partial" self.status_load = "failed" if self.status_load != "failed": self._load_extids(new_extids) return self.finalize_visit( snapshot=snapshot, visit=visit, failed_branches=failed_branches, status_visit=self.status_visit, status_load=self.status_load, errors=errors, ) def _load_directory( self, dl_artifacts: List[Tuple[str, Mapping[str, Any]]], tmpdir: str ) -> Tuple[str, from_disk.Directory]: uncompressed_path = self.uncompress(dl_artifacts, dest=tmpdir) logger.debug("uncompressed_path: %s", uncompressed_path) directory = from_disk.Directory.from_disk( path=uncompressed_path.encode("utf-8"), max_content_length=self.max_content_size, ) contents, skipped_contents, directories = from_disk.iter_directory(directory) # Instead of sending everything from the bottom up to the storage, # use a Merkle graph discovery algorithm to filter out known objects. - contents, skipped_contents, directories = discovery.filter_known_objects( - self.storage, contents, skipped_contents, directories + contents, skipped_contents, directories = asyncio.run( + discovery.filter_known_objects( + discovery.DiscoveryStorageConnection( + contents, skipped_contents, directories, self.storage + ), + ) ) logger.debug("Number of skipped contents: %s", len(skipped_contents)) self.storage.skipped_content_add(skipped_contents) logger.debug("Number of contents: %s", len(contents)) self.storage.content_add(contents) logger.debug("Number of directories: %s", len(directories)) self.storage.directory_add(directories) return (uncompressed_path, directory) def _load_release( self, p_info: TPackageInfo, origin ) -> Optional[Tuple[Sha1Git, Sha1Git]]: """Does all the loading of a release itself: * downloads a package and uncompresses it * loads it from disk * adds contents, directories, and release to self.storage * returns (release_id, directory_id) Raises exception when unable to download or uncompress artifacts """ with tempfile.TemporaryDirectory() as tmpdir: dl_artifacts = self.download_package(p_info, tmpdir) (uncompressed_path, directory) = self._load_directory(dl_artifacts, tmpdir) # FIXME: This should be release. cf. D409 release = self.build_release( p_info, uncompressed_path, directory=directory.hash ) if not release: # Some artifacts are missing intrinsic metadata # skipping those return None metadata = [metadata for (filepath, metadata) in dl_artifacts] assert release.target is not None, release assert release.target_type == ModelObjectType.DIRECTORY, release metadata_target = ExtendedSWHID( object_type=ExtendedObjectType.DIRECTORY, object_id=release.target ) original_artifact_metadata = RawExtrinsicMetadata( target=metadata_target, discovery_date=self.visit_date, authority=SWH_METADATA_AUTHORITY, fetcher=self.get_metadata_fetcher(), format="original-artifacts-json", metadata=json.dumps(metadata).encode(), origin=self.origin.url, release=release.swhid(), ) self.load_metadata_objects([original_artifact_metadata]) logger.debug("Release: %s", release) self.storage.release_add([release]) assert directory.hash return (release.id, directory.hash) def _load_snapshot( self, default_version: str, releases: Dict[str, List[Tuple[str, bytes]]], extra_branches: Dict[bytes, Mapping[str, Any]], ) -> Optional[Snapshot]: """Build snapshot out of the current releases stored and extra branches. Then load it in the storage. """ logger.debug("releases: %s", releases) # Build and load the snapshot branches = {} # type: Dict[bytes, Mapping[str, Any]] for version, branch_name_releases in releases.items(): if version == default_version and len(branch_name_releases) == 1: # only 1 branch (no ambiguity), we can create an alias # branch 'HEAD' branch_name, _ = branch_name_releases[0] # except for some corner case (deposit) if branch_name != "HEAD": branches[b"HEAD"] = { "target_type": "alias", "target": branch_name.encode("utf-8"), } for branch_name, target in branch_name_releases: branches[branch_name.encode("utf-8")] = { "target_type": "release", "target": target, } # Deal with extra-branches for name, branch_target in extra_branches.items(): if name in branches: error_message = f"Extra branch '{name!r}' has been ignored" logger.error(error_message) sentry_sdk.capture_message(error_message, "error") else: branches[name] = branch_target snapshot_data = {"branches": branches} logger.debug("snapshot: %s", snapshot_data) snapshot = Snapshot.from_dict(snapshot_data) logger.debug("snapshot: %s", snapshot) self.storage.snapshot_add([snapshot]) return snapshot def get_loader_name(self) -> str: """Returns a fully qualified name of this loader.""" return f"{self.__class__.__module__}.{self.__class__.__name__}" def get_loader_version(self) -> str: """Returns the version of the current loader.""" module_name = self.__class__.__module__ or "" module_name_parts = module_name.split(".") # Iterate rootward through the package hierarchy until we find a parent of this # loader's module with a __version__ attribute. for prefix_size in range(len(module_name_parts), 0, -1): package_name = ".".join(module_name_parts[0:prefix_size]) module = sys.modules[package_name] if hasattr(module, "__version__"): return module.__version__ # If this loader's class has no parent package with a __version__, # it should implement it itself. raise NotImplementedError( f"Could not dynamically find the version of {self.get_loader_name()}." ) def get_metadata_fetcher(self) -> MetadataFetcher: """Returns a MetadataFetcher instance representing this package loader; which is used to for adding provenance information to extracted extrinsic metadata, if any.""" return MetadataFetcher( name=self.get_loader_name(), version=self.get_loader_version(), metadata={}, ) def get_metadata_authority(self) -> MetadataAuthority: """For package loaders that get extrinsic metadata, returns the authority the metadata are coming from. """ raise NotImplementedError("get_metadata_authority") def get_extrinsic_origin_metadata(self) -> List[RawExtrinsicMetadataCore]: """Returns metadata items, used by build_extrinsic_origin_metadata.""" return [] def build_extrinsic_origin_metadata(self) -> List[RawExtrinsicMetadata]: """Builds a list of full RawExtrinsicMetadata objects, using metadata returned by get_extrinsic_origin_metadata.""" metadata_items = self.get_extrinsic_origin_metadata() if not metadata_items: # If this package loader doesn't write metadata, no need to require # an implementation for get_metadata_authority. return [] authority = self.get_metadata_authority() fetcher = self.get_metadata_fetcher() metadata_objects = [] for item in metadata_items: metadata_objects.append( RawExtrinsicMetadata( target=self.origin.swhid(), discovery_date=item.discovery_date or self.visit_date, authority=authority, fetcher=fetcher, format=item.format, metadata=item.metadata, ) ) return metadata_objects def get_extrinsic_snapshot_metadata(self) -> List[RawExtrinsicMetadataCore]: """Returns metadata items, used by build_extrinsic_snapshot_metadata.""" return [] def build_extrinsic_snapshot_metadata( self, snapshot_id: Sha1Git ) -> List[RawExtrinsicMetadata]: """Builds a list of full RawExtrinsicMetadata objects, using metadata returned by get_extrinsic_snapshot_metadata.""" metadata_items = self.get_extrinsic_snapshot_metadata() if not metadata_items: # If this package loader doesn't write metadata, no need to require # an implementation for get_metadata_authority. return [] authority = self.get_metadata_authority() fetcher = self.get_metadata_fetcher() metadata_objects = [] for item in metadata_items: metadata_objects.append( RawExtrinsicMetadata( target=ExtendedSWHID( object_type=ExtendedObjectType.SNAPSHOT, object_id=snapshot_id ), discovery_date=item.discovery_date or self.visit_date, authority=authority, fetcher=fetcher, format=item.format, metadata=item.metadata, origin=self.origin.url, ) ) return metadata_objects def build_extrinsic_directory_metadata( self, p_info: TPackageInfo, release_id: Sha1Git, directory_id: Sha1Git, ) -> List[RawExtrinsicMetadata]: if not p_info.directory_extrinsic_metadata: # If this package loader doesn't write metadata, no need to require # an implementation for get_metadata_authority. return [] authority = self.get_metadata_authority() fetcher = self.get_metadata_fetcher() metadata_objects = [] for item in p_info.directory_extrinsic_metadata: metadata_objects.append( RawExtrinsicMetadata( target=ExtendedSWHID( object_type=ExtendedObjectType.DIRECTORY, object_id=directory_id ), discovery_date=item.discovery_date or self.visit_date, authority=authority, fetcher=fetcher, format=item.format, metadata=item.metadata, origin=self.origin.url, release=CoreSWHID( object_type=ObjectType.RELEASE, object_id=release_id ), ) ) return metadata_objects def _load_extrinsic_directory_metadata( self, p_info: TPackageInfo, release_id: Sha1Git, directory_id: Sha1Git, ) -> None: metadata_objects = self.build_extrinsic_directory_metadata( p_info, release_id, directory_id ) self.load_metadata_objects(metadata_objects) def _load_extids(self, extids: Set[ExtID]) -> None: if not extids: return try: self.storage.extid_add(list(extids)) except Exception as e: logger.exception("Failed to load new ExtIDs for %s", self.origin.url) sentry_sdk.capture_exception(e) # No big deal, it just means the next visit will load the same versions # again. def rev2rel(rev: Revision, version: str) -> Release: """Converts a revision to a release.""" message = rev.message if message and not message.endswith(b"\n"): message += b"\n" return Release( name=version.encode(), message=message, target=rev.directory, target_type=ModelObjectType.DIRECTORY, synthetic=rev.synthetic, author=rev.author, date=rev.date, )