diff --git a/swh/model/identifiers.py b/swh/model/identifiers.py index 63d028d..6c61317 100644 --- a/swh/model/identifiers.py +++ b/swh/model/identifiers.py @@ -1,860 +1,817 @@ # Copyright (C) 2015-2019 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information import binascii import datetime from functools import lru_cache import hashlib from typing import Any, Dict, Union import attr -from deprecated import deprecated from .collections import ImmutableDict from .exceptions import ValidationError from .fields.hashes import validate_sha1 from .hashutil import MultiHash, hash_git_data, hash_to_hex ORIGIN = "origin" SNAPSHOT = "snapshot" REVISION = "revision" RELEASE = "release" DIRECTORY = "directory" CONTENT = "content" SWHID_NAMESPACE = "swh" SWHID_VERSION = 1 SWHID_TYPES = ["ori", "snp", "rel", "rev", "dir", "cnt"] SWHID_SEP = ":" SWHID_CTXT_SEP = ";" -# deprecated variables -PID_NAMESPACE = SWHID_NAMESPACE -PID_VERSION = SWHID_VERSION -PID_TYPES = SWHID_TYPES -PID_SEP = SWHID_SEP -PID_CTXT_SEP = SWHID_CTXT_SEP - @lru_cache() def identifier_to_bytes(identifier): """Convert a text identifier to bytes. Args: identifier: an identifier, either a 40-char hexadecimal string or a bytes object of length 20 Returns: The length 20 bytestring corresponding to the given identifier Raises: ValueError: if the identifier is of an unexpected type or length. """ if isinstance(identifier, bytes): if len(identifier) != 20: raise ValueError( "Wrong length for bytes identifier %s, expected 20" % len(identifier) ) return identifier if isinstance(identifier, str): if len(identifier) != 40: raise ValueError( "Wrong length for str identifier %s, expected 40" % len(identifier) ) return bytes.fromhex(identifier) raise ValueError( "Wrong type for identifier %s, expected bytes or str" % identifier.__class__.__name__ ) @lru_cache() def identifier_to_str(identifier): """Convert an identifier to an hexadecimal string. Args: identifier: an identifier, either a 40-char hexadecimal string or a bytes object of length 20 Returns: The length 40 string corresponding to the given identifier, hex encoded Raises: ValueError: if the identifier is of an unexpected type or length. """ if isinstance(identifier, str): if len(identifier) != 40: raise ValueError( "Wrong length for str identifier %s, expected 40" % len(identifier) ) return identifier if isinstance(identifier, bytes): if len(identifier) != 20: raise ValueError( "Wrong length for bytes identifier %s, expected 20" % len(identifier) ) return binascii.hexlify(identifier).decode() raise ValueError( "Wrong type for identifier %s, expected bytes or str" % identifier.__class__.__name__ ) def content_identifier(content): """Return the intrinsic identifier for a content. A content's identifier is the sha1, sha1_git and sha256 checksums of its data. Args: content: a content conforming to the Software Heritage schema Returns: A dictionary with all the hashes for the data Raises: KeyError: if the content doesn't have a data member. """ return MultiHash.from_data(content["data"]).digest() def directory_entry_sort_key(entry): """The sorting key for tree entries""" if entry["type"] == "dir": return entry["name"] + b"/" else: return entry["name"] @lru_cache() def _perms_to_bytes(perms): """Convert the perms value to its bytes representation""" oc = oct(perms)[2:] return oc.encode("ascii") def escape_newlines(snippet): """Escape the newlines present in snippet according to git rules. New lines in git manifests are escaped by indenting the next line by one space. """ if b"\n" in snippet: return b"\n ".join(snippet.split(b"\n")) else: return snippet def directory_identifier(directory): """Return the intrinsic identifier for a directory. A directory's identifier is the tree sha1 à la git of a directory listing, using the following algorithm, which is equivalent to the git algorithm for trees: 1. Entries of the directory are sorted using the name (or the name with '/' appended for directory entries) as key, in bytes order. 2. For each entry of the directory, the following bytes are output: - the octal representation of the permissions for the entry (stored in the 'perms' member), which is a representation of the entry type: - b'100644' (int 33188) for files - b'100755' (int 33261) for executable files - b'120000' (int 40960) for symbolic links - b'40000' (int 16384) for directories - b'160000' (int 57344) for references to revisions - an ascii space (b'\x20') - the entry's name (as raw bytes), stored in the 'name' member - a null byte (b'\x00') - the 20 byte long identifier of the object pointed at by the entry, stored in the 'target' member: - for files or executable files: their blob sha1_git - for symbolic links: the blob sha1_git of a file containing the link destination - for directories: their intrinsic identifier - for revisions: their intrinsic identifier (Note that there is no separator between entries) """ components = [] for entry in sorted(directory["entries"], key=directory_entry_sort_key): components.extend( [ _perms_to_bytes(entry["perms"]), b"\x20", entry["name"], b"\x00", identifier_to_bytes(entry["target"]), ] ) return identifier_to_str(hash_git_data(b"".join(components), "tree")) def format_date(date): """Convert a date object into an UTC timestamp encoded as ascii bytes. Git stores timestamps as an integer number of seconds since the UNIX epoch. However, Software Heritage stores timestamps as an integer number of microseconds (postgres type "datetime with timezone"). Therefore, we print timestamps with no microseconds as integers, and timestamps with microseconds as floating point values. We elide the trailing zeroes from microsecond values, to "future-proof" our representation if we ever need more precision in timestamps. """ if not isinstance(date, dict): raise ValueError("format_date only supports dicts, %r received" % date) seconds = date.get("seconds", 0) microseconds = date.get("microseconds", 0) if not microseconds: return str(seconds).encode() else: float_value = "%d.%06d" % (seconds, microseconds) return float_value.rstrip("0").encode() @lru_cache() def format_offset(offset, negative_utc=None): """Convert an integer number of minutes into an offset representation. The offset representation is [+-]hhmm where: - hh is the number of hours; - mm is the number of minutes. A null offset is represented as +0000. """ if offset < 0 or offset == 0 and negative_utc: sign = "-" else: sign = "+" hours = abs(offset) // 60 minutes = abs(offset) % 60 t = "%s%02d%02d" % (sign, hours, minutes) return t.encode() def normalize_timestamp(time_representation): """Normalize a time representation for processing by Software Heritage This function supports a numeric timestamp (representing a number of seconds since the UNIX epoch, 1970-01-01 at 00:00 UTC), a :obj:`datetime.datetime` object (with timezone information), or a normalized Software Heritage time representation (idempotency). Args: time_representation: the representation of a timestamp Returns: dict: a normalized dictionary with three keys: - timestamp: a dict with two optional keys: - seconds: the integral number of seconds since the UNIX epoch - microseconds: the integral number of microseconds - offset: the timezone offset as a number of minutes relative to UTC - negative_utc: a boolean representing whether the offset is -0000 when offset = 0. """ if time_representation is None: return None negative_utc = False if isinstance(time_representation, dict): ts = time_representation["timestamp"] if isinstance(ts, dict): seconds = ts.get("seconds", 0) microseconds = ts.get("microseconds", 0) elif isinstance(ts, int): seconds = ts microseconds = 0 else: raise ValueError( "normalize_timestamp received non-integer timestamp member:" " %r" % ts ) offset = time_representation["offset"] if "negative_utc" in time_representation: negative_utc = time_representation["negative_utc"] if negative_utc is None: negative_utc = False elif isinstance(time_representation, datetime.datetime): seconds = int(time_representation.timestamp()) microseconds = time_representation.microsecond utcoffset = time_representation.utcoffset() if utcoffset is None: raise ValueError( "normalize_timestamp received datetime without timezone: %s" % time_representation ) # utcoffset is an integer number of minutes seconds_offset = utcoffset.total_seconds() offset = int(seconds_offset) // 60 elif isinstance(time_representation, int): seconds = time_representation microseconds = 0 offset = 0 else: raise ValueError( "normalize_timestamp received non-integer timestamp:" " %r" % time_representation ) return { "timestamp": {"seconds": seconds, "microseconds": microseconds,}, "offset": offset, "negative_utc": negative_utc, } def format_author(author): """Format the specification of an author. An author is either a byte string (passed unchanged), or a dict with three keys, fullname, name and email. If the fullname exists, return it; if it doesn't, we construct a fullname using the following heuristics: if the name value is None, we return the email in angle brackets, else, we return the name, a space, and the email in angle brackets. """ if isinstance(author, bytes) or author is None: return author if "fullname" in author: return author["fullname"] ret = [] if author["name"] is not None: ret.append(author["name"]) if author["email"] is not None: ret.append(b"".join([b"<", author["email"], b">"])) return b" ".join(ret) def format_author_line(header, author, date_offset): """Format a an author line according to git standards. An author line has three components: - a header, describing the type of author (author, committer, tagger) - a name and email, which is an arbitrary bytestring - optionally, a timestamp with UTC offset specification The author line is formatted thus:: `header` `name and email`[ `timestamp` `utc_offset`] The timestamp is encoded as a (decimal) number of seconds since the UNIX epoch (1970-01-01 at 00:00 UTC). As an extension to the git format, we support fractional timestamps, using a dot as the separator for the decimal part. The utc offset is a number of minutes encoded as '[+-]HHMM'. Note some tools can pass a negative offset corresponding to the UTC timezone ('-0000'), which is valid and is encoded as such. For convenience, this function returns the whole line with its trailing newline. Args: header: the header of the author line (one of 'author', 'committer', 'tagger') author: an author specification (dict with two bytes values: name and email, or byte value) date_offset: a normalized date/time representation as returned by :func:`normalize_timestamp`. Returns: the newline-terminated byte string containing the author line """ ret = [header.encode(), b" ", escape_newlines(format_author(author))] date_offset = normalize_timestamp(date_offset) if date_offset is not None: date_f = format_date(date_offset["timestamp"]) offset_f = format_offset(date_offset["offset"], date_offset["negative_utc"]) ret.extend([b" ", date_f, b" ", offset_f]) ret.append(b"\n") return b"".join(ret) def revision_identifier(revision): """Return the intrinsic identifier for a revision. The fields used for the revision identifier computation are: - directory - parents - author - author_date - committer - committer_date - extra_headers or metadata -> extra_headers - message A revision's identifier is the 'git'-checksum of a commit manifest constructed as follows (newlines are a single ASCII newline character):: tree [for each parent in parents] parent [end for each parents] author committer [for each key, value in extra_headers] [end for each extra_headers] The directory identifier is the ascii representation of its hexadecimal encoding. Author and committer are formatted with the :func:`format_author` function. Dates are formatted with the :func:`format_offset` function. Extra headers are an ordered list of [key, value] pairs. Keys are strings and get encoded to utf-8 for identifier computation. Values are either byte strings, unicode strings (that get encoded to utf-8), or integers (that get encoded to their utf-8 decimal representation). Multiline extra header values are escaped by indenting the continuation lines with one ascii space. If the message is None, the manifest ends with the last header. Else, the message is appended to the headers after an empty line. The checksum of the full manifest is computed using the 'commit' git object type. """ components = [ b"tree ", identifier_to_str(revision["directory"]).encode(), b"\n", ] for parent in revision["parents"]: if parent: components.extend( [b"parent ", identifier_to_str(parent).encode(), b"\n",] ) components.extend( [ format_author_line("author", revision["author"], revision["date"]), format_author_line( "committer", revision["committer"], revision["committer_date"] ), ] ) # Handle extra headers metadata = revision.get("metadata") or {} extra_headers = revision.get("extra_headers", ()) if not extra_headers and "extra_headers" in metadata: extra_headers = metadata["extra_headers"] for key, value in extra_headers: components.extend([key, b" ", escape_newlines(value), b"\n"]) if revision["message"] is not None: components.extend([b"\n", revision["message"]]) commit_raw = b"".join(components) return identifier_to_str(hash_git_data(commit_raw, "commit")) def target_type_to_git(target_type): """Convert a software heritage target type to a git object type""" return { "content": b"blob", "directory": b"tree", "revision": b"commit", "release": b"tag", "snapshot": b"refs", }[target_type] def release_identifier(release): """Return the intrinsic identifier for a release.""" components = [ b"object ", identifier_to_str(release["target"]).encode(), b"\n", b"type ", target_type_to_git(release["target_type"]), b"\n", b"tag ", release["name"], b"\n", ] if "author" in release and release["author"]: components.append( format_author_line("tagger", release["author"], release["date"]) ) if release["message"] is not None: components.extend([b"\n", release["message"]]) return identifier_to_str(hash_git_data(b"".join(components), "tag")) def snapshot_identifier(snapshot, *, ignore_unresolved=False): """Return the intrinsic identifier for a snapshot. Snapshots are a set of named branches, which are pointers to objects at any level of the Software Heritage DAG. As well as pointing to other objects in the Software Heritage DAG, branches can also be *alias*es, in which case their target is the name of another branch in the same snapshot, or *dangling*, in which case the target is unknown (and represented by the ``None`` value). A snapshot identifier is a salted sha1 (using the git hashing algorithm with the ``snapshot`` object type) of a manifest following the algorithm: 1. Branches are sorted using the name as key, in bytes order. 2. For each branch, the following bytes are output: - the type of the branch target: - ``content``, ``directory``, ``revision``, ``release`` or ``snapshot`` for the corresponding entries in the DAG; - ``alias`` for branches referencing another branch; - ``dangling`` for dangling branches - an ascii space (``\\x20``) - the branch name (as raw bytes) - a null byte (``\\x00``) - the length of the target identifier, as an ascii-encoded decimal number (``20`` for current intrinsic identifiers, ``0`` for dangling branches, the length of the target branch name for branch aliases) - a colon (``:``) - the identifier of the target object pointed at by the branch, stored in the 'target' member: - for contents: their *sha1_git* - for directories, revisions, releases or snapshots: their intrinsic identifier - for branch aliases, the name of the target branch (as raw bytes) - for dangling branches, the empty string Note that, akin to directory manifests, there is no separator between entries. Because of symbolic branches, identifiers are of arbitrary length but are length-encoded to avoid ambiguity. Args: snapshot (dict): the snapshot of which to compute the identifier. A single entry is needed, ``'branches'``, which is itself a :class:`dict` mapping each branch to its target ignore_unresolved (bool): if `True`, ignore unresolved branch aliases. Returns: str: the intrinsic identifier for `snapshot` """ unresolved = [] lines = [] for name, target in sorted(snapshot["branches"].items()): if not target: target_type = b"dangling" target_id = b"" elif target["target_type"] == "alias": target_type = b"alias" target_id = target["target"] if target_id not in snapshot["branches"] or target_id == name: unresolved.append((name, target_id)) else: target_type = target["target_type"].encode() target_id = identifier_to_bytes(target["target"]) lines.extend( [ target_type, b"\x20", name, b"\x00", ("%d:" % len(target_id)).encode(), target_id, ] ) if unresolved and not ignore_unresolved: raise ValueError( "Branch aliases unresolved: %s" % ", ".join("%s -> %s" % x for x in unresolved), unresolved, ) return identifier_to_str(hash_git_data(b"".join(lines), "snapshot")) def origin_identifier(origin): """Return the intrinsic identifier for an origin. An origin's identifier is the sha1 checksum of the entire origin URL """ return hashlib.sha1(origin["url"].encode("utf-8")).hexdigest() _object_type_map = { ORIGIN: {"short_name": "ori", "key_id": "id"}, SNAPSHOT: {"short_name": "snp", "key_id": "id"}, RELEASE: {"short_name": "rel", "key_id": "id"}, REVISION: {"short_name": "rev", "key_id": "id"}, DIRECTORY: {"short_name": "dir", "key_id": "id"}, CONTENT: {"short_name": "cnt", "key_id": "sha1_git"}, } @attr.s(frozen=True) class SWHID: """ Named tuple holding the relevant info associated to a SoftWare Heritage persistent IDentifier (SWHID) Args: namespace (str): the namespace of the identifier, defaults to ``swh`` scheme_version (int): the scheme version of the identifier, defaults to 1 object_type (str): the type of object the identifier points to, either ``content``, ``directory``, ``release``, ``revision`` or ``snapshot`` object_id (str): object's identifier metadata (dict): optional dict filled with metadata related to pointed object Raises: swh.model.exceptions.ValidationError: In case of invalid object type or id Once created, it contains the following attributes: Attributes: namespace (str): the namespace of the identifier scheme_version (int): the scheme version of the identifier object_type (str): the type of object the identifier points to object_id (str): hexadecimal representation of the object hash metadata (dict): metadata related to the pointed object To get the raw SWHID string from an instance of this named tuple, use the :func:`str` function:: swhid = SWHID( object_type='content', object_id='8ff44f081d43176474b267de5451f2c2e88089d0' ) swhid_str = str(swhid) # 'swh:1:cnt:8ff44f081d43176474b267de5451f2c2e88089d0' """ namespace = attr.ib(type=str, default="swh") scheme_version = attr.ib(type=int, default=1) object_type = attr.ib(type=str, default="") object_id = attr.ib(type=str, converter=hash_to_hex, default="") # type: ignore metadata = attr.ib( type=ImmutableDict[str, Any], converter=ImmutableDict, default=ImmutableDict() ) @namespace.validator def check_namespace(self, attribute, value): if value != SWHID_NAMESPACE: raise ValidationError( "Wrong format: only supported namespace is '%s'" % SWHID_NAMESPACE ) @scheme_version.validator def check_scheme_version(self, attribute, value): if value != SWHID_VERSION: raise ValidationError( "Wrong format: only supported version is %d" % SWHID_VERSION ) @object_type.validator def check_object_type(self, attribute, value): if value not in _object_type_map: raise ValidationError( "Wrong input: Supported types are %s" % (list(_object_type_map.keys())) ) @object_id.validator def check_object_id(self, attribute, value): validate_sha1(value) # can raise if invalid hash def to_dict(self) -> Dict[str, Any]: return attr.asdict(self) def __str__(self) -> str: o = _object_type_map.get(self.object_type) assert o swhid = SWHID_SEP.join( [self.namespace, str(self.scheme_version), o["short_name"], self.object_id] ) if self.metadata: for k, v in self.metadata.items(): swhid += "%s%s=%s" % (SWHID_CTXT_SEP, k, v) return swhid -@deprecated("Use swh.model.identifiers.SWHID instead") -class PersistentId(SWHID): - """ - Named tuple holding the relevant info associated to a SoftWare Heritage - persistent IDentifier. - - .. deprecated:: 0.3.8 - Use :class:`swh.model.identifiers.SWHID` instead - - """ - - pass - - def swhid( object_type: str, object_id: Union[str, Dict[str, Any]], scheme_version: int = 1, metadata: Union[ImmutableDict[str, Any], Dict[str, Any]] = ImmutableDict(), ) -> str: """Compute :ref:`persistent-identifiers` Args: object_type: object's type, either ``content``, ``directory``, ``release``, ``revision`` or ``snapshot`` object_id: object's identifier scheme_version: SWHID scheme version, defaults to 1 metadata: metadata related to the pointed object Raises: swh.model.exceptions.ValidationError: In case of invalid object type or id Returns: the SWHID of the object """ if isinstance(object_id, dict): o = _object_type_map[object_type] object_id = object_id[o["key_id"]] swhid = SWHID( scheme_version=scheme_version, object_type=object_type, object_id=object_id, metadata=metadata, # type: ignore # mypy can't properly unify types ) return str(swhid) -@deprecated("Use swh.model.identifiers.swhid instead") -def persistent_identifier(*args, **kwargs) -> str: - """Compute :ref:`persistent-identifiers` - - .. deprecated:: 0.3.8 - Use :func:`swh.model.identifiers.swhid` instead - - """ - return swhid(*args, **kwargs) - - def parse_swhid(swhid: str) -> SWHID: """Parse :ref:`persistent-identifiers`. Args: swhid (str): A persistent identifier Raises: swh.model.exceptions.ValidationError: in case of: * missing mandatory values (4) * invalid namespace supplied * invalid version supplied * invalid type supplied * missing hash * invalid hash identifier supplied Returns: a named tuple holding the parsing result """ # ; swhid_parts = swhid.split(SWHID_CTXT_SEP) swhid_data = swhid_parts.pop(0).split(":") if len(swhid_data) != 4: raise ValidationError("Wrong format: There should be 4 mandatory values") # Checking for parsing errors _ns, _version, _type, _id = swhid_data for otype, data in _object_type_map.items(): if _type == data["short_name"]: _type = otype break if not _id: raise ValidationError("Wrong format: Identifier should be present") _metadata = {} for part in swhid_parts: try: key, val = part.split("=") _metadata[key] = val except Exception: msg = "Contextual data is badly formatted, form key=val expected" raise ValidationError(msg) return SWHID( _ns, int(_version), _type, _id, _metadata, # type: ignore # mypy can't properly unify types ) - - -@deprecated("Use swh.model.identifiers.parse_swhid instead") -def parse_persistent_identifier(persistent_id: str) -> PersistentId: - """Parse :ref:`persistent-identifiers`. - - .. deprecated:: 0.3.8 - Use :func:`swh.model.identifiers.parse_swhid` instead - """ - return PersistentId(**parse_swhid(persistent_id).to_dict()) diff --git a/swh/model/model.py b/swh/model/model.py index d719966..ca3a8c7 100644 --- a/swh/model/model.py +++ b/swh/model/model.py @@ -1,1026 +1,1026 @@ # Copyright (C) 2018-2020 The Software Heritage developers # See the AUTHORS file at the top-level directory of this distribution # License: GNU General Public License version 3, or any later version # See top-level LICENSE file for more information from abc import ABCMeta, abstractmethod import datetime from enum import Enum from hashlib import sha256 from typing import Any, Dict, Iterable, Optional, Tuple, TypeVar, Union import attr from attrs_strict import type_validator import dateutil.parser import iso8601 from typing_extensions import Final from .collections import ImmutableDict from .hashutil import DEFAULT_ALGORITHMS, MultiHash, hash_to_bytes from .identifiers import ( SWHID, directory_identifier, normalize_timestamp, parse_swhid, release_identifier, revision_identifier, snapshot_identifier, ) class MissingData(Exception): """Raised by `Content.with_data` when it has no way of fetching the data (but not when fetching the data fails).""" pass SHA1_SIZE = 20 # TODO: Limit this to 20 bytes Sha1Git = bytes Sha1 = bytes KT = TypeVar("KT") VT = TypeVar("VT") def freeze_optional_dict( d: Union[None, Dict[KT, VT], ImmutableDict[KT, VT]] # type: ignore ) -> Optional[ImmutableDict[KT, VT]]: if isinstance(d, dict): return ImmutableDict(d) else: return d def dictify(value): "Helper function used by BaseModel.to_dict()" if isinstance(value, BaseModel): return value.to_dict() elif isinstance(value, SWHID): return str(value) elif isinstance(value, Enum): return value.value elif isinstance(value, (dict, ImmutableDict)): return {k: dictify(v) for k, v in value.items()} elif isinstance(value, tuple): return tuple(dictify(v) for v in value) else: return value ModelType = TypeVar("ModelType", bound="BaseModel") class BaseModel: """Base class for SWH model classes. Provides serialization/deserialization to/from Python dictionaries, that are suitable for JSON/msgpack-like formats.""" def to_dict(self): """Wrapper of `attr.asdict` that can be overridden by subclasses that have special handling of some of the fields.""" return dictify(attr.asdict(self, recurse=False)) @classmethod def from_dict(cls, d): """Takes a dictionary representing a tree of SWH objects, and recursively builds the corresponding objects.""" return cls(**d) def anonymize(self: ModelType) -> Optional[ModelType]: """Returns an anonymized version of the object, if needed. If the object model does not need/support anonymization, returns None. """ return None class HashableObject(metaclass=ABCMeta): """Mixin to automatically compute object identifier hash when the associated model is instantiated.""" @staticmethod @abstractmethod def compute_hash(object_dict): """Derived model classes must implement this to compute the object hash from its dict representation.""" pass def __attrs_post_init__(self): if not self.id: obj_id = hash_to_bytes(self.compute_hash(self.to_dict())) object.__setattr__(self, "id", obj_id) @attr.s(frozen=True) class Person(BaseModel): """Represents the author/committer of a revision or release.""" object_type: Final = "person" fullname = attr.ib(type=bytes, validator=type_validator()) name = attr.ib(type=Optional[bytes], validator=type_validator()) email = attr.ib(type=Optional[bytes], validator=type_validator()) @classmethod def from_fullname(cls, fullname: bytes): """Returns a Person object, by guessing the name and email from the fullname, in the `name ` format. The fullname is left unchanged.""" if fullname is None: raise TypeError("fullname is None.") name: Optional[bytes] email: Optional[bytes] try: open_bracket = fullname.index(b"<") except ValueError: name = fullname email = None else: raw_name = fullname[:open_bracket] raw_email = fullname[open_bracket + 1 :] if not raw_name: name = None else: name = raw_name.strip() try: close_bracket = raw_email.rindex(b">") except ValueError: email = raw_email else: email = raw_email[:close_bracket] return Person(name=name or None, email=email or None, fullname=fullname,) def anonymize(self) -> "Person": """Returns an anonymized version of the Person object. Anonymization is simply a Person which fullname is the hashed, with unset name or email. """ return Person(fullname=sha256(self.fullname).digest(), name=None, email=None,) @attr.s(frozen=True) class Timestamp(BaseModel): """Represents a naive timestamp from a VCS.""" object_type: Final = "timestamp" seconds = attr.ib(type=int, validator=type_validator()) microseconds = attr.ib(type=int, validator=type_validator()) @seconds.validator def check_seconds(self, attribute, value): """Check that seconds fit in a 64-bits signed integer.""" if not (-(2 ** 63) <= value < 2 ** 63): raise ValueError("Seconds must be a signed 64-bits integer.") @microseconds.validator def check_microseconds(self, attribute, value): """Checks that microseconds are positive and < 1000000.""" if not (0 <= value < 10 ** 6): raise ValueError("Microseconds must be in [0, 1000000[.") @attr.s(frozen=True) class TimestampWithTimezone(BaseModel): """Represents a TZ-aware timestamp from a VCS.""" object_type: Final = "timestamp_with_timezone" timestamp = attr.ib(type=Timestamp, validator=type_validator()) offset = attr.ib(type=int, validator=type_validator()) negative_utc = attr.ib(type=bool, validator=type_validator()) @offset.validator def check_offset(self, attribute, value): """Checks the offset is a 16-bits signed integer (in theory, it should always be between -14 and +14 hours).""" if not (-(2 ** 15) <= value < 2 ** 15): # max 14 hours offset in theory, but you never know what # you'll find in the wild... raise ValueError("offset too large: %d minutes" % value) @negative_utc.validator def check_negative_utc(self, attribute, value): if self.offset and value: raise ValueError("negative_utc can only be True is offset=0") @classmethod def from_dict(cls, obj: Union[Dict, datetime.datetime, int]): """Builds a TimestampWithTimezone from any of the formats accepted by :func:`swh.model.normalize_timestamp`.""" # TODO: this accept way more types than just dicts; find a better # name d = normalize_timestamp(obj) return cls( timestamp=Timestamp.from_dict(d["timestamp"]), offset=d["offset"], negative_utc=d["negative_utc"], ) @classmethod def from_datetime(cls, dt: datetime.datetime): return cls.from_dict(dt) @classmethod def from_iso8601(cls, s): """Builds a TimestampWithTimezone from an ISO8601-formatted string. """ dt = iso8601.parse_date(s) tstz = cls.from_datetime(dt) if dt.tzname() == "-00:00": tstz = attr.evolve(tstz, negative_utc=True) return tstz @attr.s(frozen=True) class Origin(BaseModel): """Represents a software source: a VCS and an URL.""" object_type: Final = "origin" url = attr.ib(type=str, validator=type_validator()) @attr.s(frozen=True) class OriginVisit(BaseModel): """Represents an origin visit with a given type at a given point in time, by a SWH loader.""" object_type: Final = "origin_visit" origin = attr.ib(type=str, validator=type_validator()) date = attr.ib(type=datetime.datetime, validator=type_validator()) type = attr.ib(type=str, validator=type_validator()) """Should not be set before calling 'origin_visit_add()'.""" visit = attr.ib(type=Optional[int], validator=type_validator(), default=None) @date.validator def check_date(self, attribute, value): """Checks the date has a timezone.""" if value is not None and value.tzinfo is None: raise ValueError("date must be a timezone-aware datetime.") def to_dict(self): """Serializes the date as a string and omits the visit id if it is `None`.""" ov = super().to_dict() if ov["visit"] is None: del ov["visit"] return ov @attr.s(frozen=True) class OriginVisitStatus(BaseModel): """Represents a visit update of an origin at a given point in time. """ object_type: Final = "origin_visit_status" origin = attr.ib(type=str, validator=type_validator()) visit = attr.ib(type=int, validator=type_validator()) date = attr.ib(type=datetime.datetime, validator=type_validator()) status = attr.ib( type=str, validator=attr.validators.in_(["created", "ongoing", "full", "partial"]), ) snapshot = attr.ib(type=Optional[Sha1Git], validator=type_validator()) metadata = attr.ib( type=Optional[ImmutableDict[str, object]], validator=type_validator(), converter=freeze_optional_dict, default=None, ) @date.validator def check_date(self, attribute, value): """Checks the date has a timezone.""" if value is not None and value.tzinfo is None: raise ValueError("date must be a timezone-aware datetime.") class TargetType(Enum): """The type of content pointed to by a snapshot branch. Usually a revision or an alias.""" CONTENT = "content" DIRECTORY = "directory" REVISION = "revision" RELEASE = "release" SNAPSHOT = "snapshot" ALIAS = "alias" class ObjectType(Enum): """The type of content pointed to by a release. Usually a revision""" CONTENT = "content" DIRECTORY = "directory" REVISION = "revision" RELEASE = "release" SNAPSHOT = "snapshot" @attr.s(frozen=True) class SnapshotBranch(BaseModel): """Represents one of the branches of a snapshot.""" object_type: Final = "snapshot_branch" target = attr.ib(type=bytes, validator=type_validator()) target_type = attr.ib(type=TargetType, validator=type_validator()) @target.validator def check_target(self, attribute, value): """Checks the target type is not an alias, checks the target is a valid sha1_git.""" if self.target_type != TargetType.ALIAS and self.target is not None: if len(value) != 20: raise ValueError("Wrong length for bytes identifier: %d" % len(value)) @classmethod def from_dict(cls, d): return cls(target=d["target"], target_type=TargetType(d["target_type"])) @attr.s(frozen=True) class Snapshot(BaseModel, HashableObject): """Represents the full state of an origin at a given point in time.""" object_type: Final = "snapshot" branches = attr.ib( type=ImmutableDict[bytes, Optional[SnapshotBranch]], validator=type_validator(), converter=freeze_optional_dict, ) id = attr.ib(type=Sha1Git, validator=type_validator(), default=b"") @staticmethod def compute_hash(object_dict): return snapshot_identifier(object_dict) @classmethod def from_dict(cls, d): d = d.copy() return cls( branches=ImmutableDict( (name, SnapshotBranch.from_dict(branch) if branch else None) for (name, branch) in d.pop("branches").items() ), **d, ) @attr.s(frozen=True) class Release(BaseModel, HashableObject): object_type: Final = "release" name = attr.ib(type=bytes, validator=type_validator()) message = attr.ib(type=Optional[bytes], validator=type_validator()) target = attr.ib(type=Optional[Sha1Git], validator=type_validator()) target_type = attr.ib(type=ObjectType, validator=type_validator()) synthetic = attr.ib(type=bool, validator=type_validator()) author = attr.ib(type=Optional[Person], validator=type_validator(), default=None) date = attr.ib( type=Optional[TimestampWithTimezone], validator=type_validator(), default=None ) metadata = attr.ib( type=Optional[ImmutableDict[str, object]], validator=type_validator(), converter=freeze_optional_dict, default=None, ) id = attr.ib(type=Sha1Git, validator=type_validator(), default=b"") @staticmethod def compute_hash(object_dict): return release_identifier(object_dict) @author.validator def check_author(self, attribute, value): """If the author is `None`, checks the date is `None` too.""" if self.author is None and self.date is not None: raise ValueError("release date must be None if author is None.") def to_dict(self): rel = super().to_dict() if rel["metadata"] is None: del rel["metadata"] return rel @classmethod def from_dict(cls, d): d = d.copy() if d.get("author"): d["author"] = Person.from_dict(d["author"]) if d.get("date"): d["date"] = TimestampWithTimezone.from_dict(d["date"]) return cls(target_type=ObjectType(d.pop("target_type")), **d) def anonymize(self) -> "Release": """Returns an anonymized version of the Release object. Anonymization consists in replacing the author with an anonymized Person object. """ author = self.author and self.author.anonymize() return attr.evolve(self, author=author) class RevisionType(Enum): GIT = "git" TAR = "tar" DSC = "dsc" SUBVERSION = "svn" MERCURIAL = "hg" def tuplify_extra_headers(value: Iterable): return tuple((k, v) for k, v in value) @attr.s(frozen=True) class Revision(BaseModel, HashableObject): object_type: Final = "revision" message = attr.ib(type=Optional[bytes], validator=type_validator()) author = attr.ib(type=Person, validator=type_validator()) committer = attr.ib(type=Person, validator=type_validator()) date = attr.ib(type=Optional[TimestampWithTimezone], validator=type_validator()) committer_date = attr.ib( type=Optional[TimestampWithTimezone], validator=type_validator() ) type = attr.ib(type=RevisionType, validator=type_validator()) directory = attr.ib(type=Sha1Git, validator=type_validator()) synthetic = attr.ib(type=bool, validator=type_validator()) metadata = attr.ib( type=Optional[ImmutableDict[str, object]], validator=type_validator(), converter=freeze_optional_dict, default=None, ) parents = attr.ib(type=Tuple[Sha1Git, ...], validator=type_validator(), default=()) id = attr.ib(type=Sha1Git, validator=type_validator(), default=b"") extra_headers = attr.ib( type=Tuple[Tuple[bytes, bytes], ...], validator=type_validator(), converter=tuplify_extra_headers, default=(), ) def __attrs_post_init__(self): super().__attrs_post_init__() # ensure metadata is a deep copy of whatever was given, and if needed # extract extra_headers from there if self.metadata: metadata = self.metadata if not self.extra_headers and "extra_headers" in metadata: (extra_headers, metadata) = metadata.copy_pop("extra_headers") object.__setattr__( self, "extra_headers", tuplify_extra_headers(extra_headers), ) attr.validate(self) object.__setattr__(self, "metadata", metadata) @staticmethod def compute_hash(object_dict): return revision_identifier(object_dict) @classmethod def from_dict(cls, d): d = d.copy() date = d.pop("date") if date: date = TimestampWithTimezone.from_dict(date) committer_date = d.pop("committer_date") if committer_date: committer_date = TimestampWithTimezone.from_dict(committer_date) return cls( author=Person.from_dict(d.pop("author")), committer=Person.from_dict(d.pop("committer")), date=date, committer_date=committer_date, type=RevisionType(d.pop("type")), parents=tuple(d.pop("parents")), # for BW compat **d, ) def anonymize(self) -> "Revision": """Returns an anonymized version of the Revision object. Anonymization consists in replacing the author and committer with an anonymized Person object. """ return attr.evolve( self, author=self.author.anonymize(), committer=self.committer.anonymize() ) @attr.s(frozen=True) class DirectoryEntry(BaseModel): object_type: Final = "directory_entry" name = attr.ib(type=bytes, validator=type_validator()) type = attr.ib(type=str, validator=attr.validators.in_(["file", "dir", "rev"])) target = attr.ib(type=Sha1Git, validator=type_validator()) perms = attr.ib(type=int, validator=type_validator()) """Usually one of the values of `swh.model.from_disk.DentryPerms`.""" @attr.s(frozen=True) class Directory(BaseModel, HashableObject): object_type: Final = "directory" entries = attr.ib(type=Tuple[DirectoryEntry, ...], validator=type_validator()) id = attr.ib(type=Sha1Git, validator=type_validator(), default=b"") @staticmethod def compute_hash(object_dict): return directory_identifier(object_dict) @classmethod def from_dict(cls, d): d = d.copy() return cls( entries=tuple( DirectoryEntry.from_dict(entry) for entry in d.pop("entries") ), **d, ) @attr.s(frozen=True) class BaseContent(BaseModel): status = attr.ib( type=str, validator=attr.validators.in_(["visible", "hidden", "absent"]) ) @staticmethod def _hash_data(data: bytes): """Hash some data, returning most of the fields of a content object""" d = MultiHash.from_data(data).digest() d["data"] = data d["length"] = len(data) return d @classmethod def from_dict(cls, d, use_subclass=True): if use_subclass: # Chooses a subclass to instantiate instead. if d["status"] == "absent": return SkippedContent.from_dict(d) else: return Content.from_dict(d) else: return super().from_dict(d) def get_hash(self, hash_name): if hash_name not in DEFAULT_ALGORITHMS: raise ValueError("{} is not a valid hash name.".format(hash_name)) return getattr(self, hash_name) def hashes(self) -> Dict[str, bytes]: """Returns a dictionary {hash_name: hash_value}""" return {algo: getattr(self, algo) for algo in DEFAULT_ALGORITHMS} @attr.s(frozen=True) class Content(BaseContent): object_type: Final = "content" sha1 = attr.ib(type=bytes, validator=type_validator()) sha1_git = attr.ib(type=Sha1Git, validator=type_validator()) sha256 = attr.ib(type=bytes, validator=type_validator()) blake2s256 = attr.ib(type=bytes, validator=type_validator()) length = attr.ib(type=int, validator=type_validator()) status = attr.ib( type=str, validator=attr.validators.in_(["visible", "hidden"]), default="visible", ) data = attr.ib(type=Optional[bytes], validator=type_validator(), default=None) ctime = attr.ib( type=Optional[datetime.datetime], validator=type_validator(), default=None, eq=False, ) @length.validator def check_length(self, attribute, value): """Checks the length is positive.""" if value < 0: raise ValueError("Length must be positive.") @ctime.validator def check_ctime(self, attribute, value): """Checks the ctime has a timezone.""" if value is not None and value.tzinfo is None: raise ValueError("ctime must be a timezone-aware datetime.") def to_dict(self): content = super().to_dict() if content["data"] is None: del content["data"] if content["ctime"] is None: del content["ctime"] return content @classmethod def from_data(cls, data, status="visible", ctime=None) -> "Content": """Generate a Content from a given `data` byte string. This populates the Content with the hashes and length for the data passed as argument, as well as the data itself. """ d = cls._hash_data(data) d["status"] = status d["ctime"] = ctime return cls(**d) @classmethod def from_dict(cls, d): if isinstance(d.get("ctime"), str): d = d.copy() d["ctime"] = dateutil.parser.parse(d["ctime"]) return super().from_dict(d, use_subclass=False) def with_data(self) -> "Content": """Loads the `data` attribute; meaning that it is guaranteed not to be None after this call. This call is almost a no-op, but subclasses may overload this method to lazy-load data (eg. from disk or objstorage).""" if self.data is None: raise MissingData("Content data is None.") return self @attr.s(frozen=True) class SkippedContent(BaseContent): object_type: Final = "skipped_content" sha1 = attr.ib(type=Optional[bytes], validator=type_validator()) sha1_git = attr.ib(type=Optional[Sha1Git], validator=type_validator()) sha256 = attr.ib(type=Optional[bytes], validator=type_validator()) blake2s256 = attr.ib(type=Optional[bytes], validator=type_validator()) length = attr.ib(type=Optional[int], validator=type_validator()) status = attr.ib(type=str, validator=attr.validators.in_(["absent"])) reason = attr.ib(type=Optional[str], validator=type_validator(), default=None) origin = attr.ib(type=Optional[str], validator=type_validator(), default=None) ctime = attr.ib( type=Optional[datetime.datetime], validator=type_validator(), default=None, eq=False, ) @reason.validator def check_reason(self, attribute, value): """Checks the reason is full if status != absent.""" assert self.reason == value if value is None: raise ValueError("Must provide a reason if content is absent.") @length.validator def check_length(self, attribute, value): """Checks the length is positive or -1.""" if value < -1: raise ValueError("Length must be positive or -1.") @ctime.validator def check_ctime(self, attribute, value): """Checks the ctime has a timezone.""" if value is not None and value.tzinfo is None: raise ValueError("ctime must be a timezone-aware datetime.") def to_dict(self): content = super().to_dict() if content["origin"] is None: del content["origin"] if content["ctime"] is None: del content["ctime"] return content @classmethod def from_data( cls, data: bytes, reason: str, ctime: Optional[datetime.datetime] = None ) -> "SkippedContent": """Generate a SkippedContent from a given `data` byte string. This populates the SkippedContent with the hashes and length for the data passed as argument. You can use `attr.evolve` on such a generated content to nullify some of its attributes, e.g. for tests. """ d = cls._hash_data(data) del d["data"] d["status"] = "absent" d["reason"] = reason d["ctime"] = ctime return cls(**d) @classmethod def from_dict(cls, d): d2 = d.copy() if d2.pop("data", None) is not None: raise ValueError('SkippedContent has no "data" attribute %r' % d) return super().from_dict(d2, use_subclass=False) class MetadataAuthorityType(Enum): DEPOSIT_CLIENT = "deposit_client" FORGE = "forge" REGISTRY = "registry" @attr.s(frozen=True) class MetadataAuthority(BaseModel): """Represents an entity that provides metadata about an origin or software artifact.""" object_type: Final = "metadata_authority" type = attr.ib(type=MetadataAuthorityType, validator=type_validator()) url = attr.ib(type=str, validator=type_validator()) metadata = attr.ib( type=Optional[ImmutableDict[str, Any]], default=None, validator=type_validator(), converter=freeze_optional_dict, ) def to_dict(self): d = super().to_dict() if d["metadata"] is None: del d["metadata"] return d @classmethod def from_dict(cls, d): d["type"] = MetadataAuthorityType(d["type"]) return super().from_dict(d) @attr.s(frozen=True) class MetadataFetcher(BaseModel): """Represents a software component used to fetch metadata from a metadata authority, and ingest them into the Software Heritage archive.""" object_type: Final = "metadata_fetcher" name = attr.ib(type=str, validator=type_validator()) version = attr.ib(type=str, validator=type_validator()) metadata = attr.ib( type=Optional[ImmutableDict[str, Any]], default=None, validator=type_validator(), converter=freeze_optional_dict, ) def to_dict(self): d = super().to_dict() if d["metadata"] is None: del d["metadata"] return d class MetadataTargetType(Enum): """The type of object extrinsic metadata refer to.""" CONTENT = "content" DIRECTORY = "directory" REVISION = "revision" RELEASE = "release" SNAPSHOT = "snapshot" ORIGIN = "origin" @attr.s(frozen=True) class RawExtrinsicMetadata(BaseModel): object_type: Final = "raw_extrinsic_metadata" # target object type = attr.ib(type=MetadataTargetType, validator=type_validator()) id = attr.ib(type=Union[str, SWHID], validator=type_validator()) """URL if type=MetadataTargetType.ORIGIN, else core SWHID""" # source discovery_date = attr.ib(type=datetime.datetime, validator=type_validator()) authority = attr.ib(type=MetadataAuthority, validator=type_validator()) fetcher = attr.ib(type=MetadataFetcher, validator=type_validator()) # the metadata itself format = attr.ib(type=str, validator=type_validator()) metadata = attr.ib(type=bytes, validator=type_validator()) # context origin = attr.ib(type=Optional[str], default=None, validator=type_validator()) visit = attr.ib(type=Optional[int], default=None, validator=type_validator()) snapshot = attr.ib(type=Optional[SWHID], default=None, validator=type_validator()) release = attr.ib(type=Optional[SWHID], default=None, validator=type_validator()) revision = attr.ib(type=Optional[SWHID], default=None, validator=type_validator()) path = attr.ib(type=Optional[bytes], default=None, validator=type_validator()) directory = attr.ib(type=Optional[SWHID], default=None, validator=type_validator()) @id.validator def check_id(self, attribute, value): if self.type == MetadataTargetType.ORIGIN: if isinstance(value, SWHID) or value.startswith("swh:"): raise ValueError( "Got SWHID as id for origin metadata (expected an URL)." ) else: - self._check_pid(self.type.value, value) + self._check_swhid(self.type.value, value) @discovery_date.validator def check_discovery_date(self, attribute, value): """Checks the discovery_date has a timezone.""" if value is not None and value.tzinfo is None: raise ValueError("discovery_date must be a timezone-aware datetime.") @origin.validator def check_origin(self, attribute, value): if value is None: return if self.type not in ( MetadataTargetType.SNAPSHOT, MetadataTargetType.RELEASE, MetadataTargetType.REVISION, MetadataTargetType.DIRECTORY, MetadataTargetType.CONTENT, ): raise ValueError( f"Unexpected 'origin' context for {self.type.value} object: {value}" ) if value.startswith("swh:"): # Technically this is valid; but: # 1. SWHIDs are URIs, not URLs # 2. if a SWHID gets here, it's very likely to be a mistake # (and we can remove this check if it turns out there is a # legitimate use for it). raise ValueError(f"SWHID used as context origin URL: {value}") @visit.validator def check_visit(self, attribute, value): if value is None: return if self.type not in ( MetadataTargetType.SNAPSHOT, MetadataTargetType.RELEASE, MetadataTargetType.REVISION, MetadataTargetType.DIRECTORY, MetadataTargetType.CONTENT, ): raise ValueError( f"Unexpected 'visit' context for {self.type.value} object: {value}" ) if self.origin is None: raise ValueError("'origin' context must be set if 'visit' is.") if value <= 0: raise ValueError("Nonpositive visit id") @snapshot.validator def check_snapshot(self, attribute, value): if value is None: return if self.type not in ( MetadataTargetType.RELEASE, MetadataTargetType.REVISION, MetadataTargetType.DIRECTORY, MetadataTargetType.CONTENT, ): raise ValueError( f"Unexpected 'snapshot' context for {self.type.value} object: {value}" ) - self._check_pid("snapshot", value) + self._check_swhid("snapshot", value) @release.validator def check_release(self, attribute, value): if value is None: return if self.type not in ( MetadataTargetType.REVISION, MetadataTargetType.DIRECTORY, MetadataTargetType.CONTENT, ): raise ValueError( f"Unexpected 'release' context for {self.type.value} object: {value}" ) - self._check_pid("release", value) + self._check_swhid("release", value) @revision.validator def check_revision(self, attribute, value): if value is None: return if self.type not in (MetadataTargetType.DIRECTORY, MetadataTargetType.CONTENT,): raise ValueError( f"Unexpected 'revision' context for {self.type.value} object: {value}" ) - self._check_pid("revision", value) + self._check_swhid("revision", value) @path.validator def check_path(self, attribute, value): if value is None: return if self.type not in (MetadataTargetType.DIRECTORY, MetadataTargetType.CONTENT,): raise ValueError( f"Unexpected 'path' context for {self.type.value} object: {value}" ) @directory.validator def check_directory(self, attribute, value): if value is None: return if self.type not in (MetadataTargetType.CONTENT,): raise ValueError( f"Unexpected 'directory' context for {self.type.value} object: {value}" ) - self._check_pid("directory", value) + self._check_swhid("directory", value) - def _check_pid(self, expected_object_type, pid): - if isinstance(pid, str): - raise ValueError(f"Expected SWHID, got a string: {pid}") + def _check_swhid(self, expected_object_type, swhid): + if isinstance(swhid, str): + raise ValueError(f"Expected SWHID, got a string: {swhid}") - if pid.object_type != expected_object_type: + if swhid.object_type != expected_object_type: raise ValueError( f"Expected SWHID type '{expected_object_type}', " - f"got '{pid.object_type}' in {pid}" + f"got '{swhid.object_type}' in {swhid}" ) - if pid.metadata: - raise ValueError(f"Expected core SWHID, but got: {pid}") + if swhid.metadata: + raise ValueError(f"Expected core SWHID, but got: {swhid}") def to_dict(self): d = super().to_dict() context_keys = ( "origin", "visit", "snapshot", "release", "revision", "directory", "path", ) for context_key in context_keys: if d[context_key] is None: del d[context_key] return d @classmethod def from_dict(cls, d): d = { **d, "type": MetadataTargetType(d["type"]), "authority": MetadataAuthority.from_dict(d["authority"]), "fetcher": MetadataFetcher.from_dict(d["fetcher"]), } if d["type"] != MetadataTargetType.ORIGIN: d["id"] = parse_swhid(d["id"]) swhid_keys = ("snapshot", "release", "revision", "directory") for swhid_key in swhid_keys: if d.get(swhid_key): d[swhid_key] = parse_swhid(d[swhid_key]) return super().from_dict(d)