diff --git a/swh/model/identifiers.py b/swh/model/identifiers.py
index e040d068c79e7ea0e9b09b62eb6c2e46dbc7e66c..ab65904b900bcb18e3fa8206eb7149eec147305e 100644
--- a/swh/model/identifiers.py
+++ b/swh/model/identifiers.py
@@ -11,7 +11,18 @@ import enum
from functools import lru_cache
import hashlib
import re
-from typing import Any, Dict, Iterable, List, Optional, Tuple, Union
+from typing import (
+ Any,
+ Dict,
+ Generic,
+ Iterable,
+ List,
+ Optional,
+ Tuple,
+ Type,
+ TypeVar,
+ Union,
+)
import warnings
import attr
@@ -735,35 +746,21 @@ _swhid_type_map = {
}
-@attr.s(frozen=True, kw_only=True)
-class CoreSWHID:
- """
- Dataclass holding the relevant info associated to a SoftWare Heritage
- persistent IDentifier (SWHID).
-
- Unlike `QualifiedSWHID`, it is restricted to core SWHIDs, ie. SWHIDs
- with no qualifiers.
-
- Raises:
- swh.model.exceptions.ValidationError: In case of invalid object type or id
+# type of the "object_type" attribute of the SWHID class; either
+# ObjectType or ExtendedObjectType
+_TObjectType = TypeVar("_TObjectType", ObjectType, ExtendedObjectType)
- To get the raw SWHID string from an instance of this class,
- use the :func:`str` function:
+# the SWHID class itself (this is used so that X.from_string() can return X
+# for all X subclass of _BaseSWHID)
+_TSWHID = TypeVar("_TSWHID", bound="_BaseSWHID")
- >>> swhid = CoreSWHID(
- ... object_type=ObjectType.CONTENT,
- ... object_id=bytes.fromhex('8ff44f081d43176474b267de5451f2c2e88089d0'),
- ... )
- >>> str(swhid)
- 'swh:1:cnt:8ff44f081d43176474b267de5451f2c2e88089d0'
- And vice-versa with :meth:`CoreSWHID.from_string`:
+@attr.s(frozen=True, kw_only=True)
+class _BaseSWHID(Generic[_TObjectType]):
+ """Common base class for CoreSWHID, QualifiedSWHID, and ExtendedSWHID.
- >>> swhid == CoreSWHID.from_string(
- ... "swh:1:cnt:8ff44f081d43176474b267de5451f2c2e88089d0"
- ... )
- True
- """
+ This is an "abstract" class and should not be instantiated directly;
+ it only exists to deduplicate code between these three SWHID classes."""
namespace = attr.ib(type=str, default=SWHID_NAMESPACE)
"""the namespace of the identifier, defaults to ``swh``"""
@@ -771,9 +768,8 @@ class CoreSWHID:
scheme_version = attr.ib(type=int, default=SWHID_VERSION)
"""the scheme version of the identifier, defaults to 1"""
- object_type = attr.ib(
- type=ObjectType, validator=type_validator(), converter=ObjectType
- )
+ # overridden by subclasses
+ object_type: _TObjectType
"""the type of object the identifier points to"""
object_id = attr.ib(type=bytes, validator=type_validator())
@@ -813,16 +809,52 @@ class CoreSWHID:
)
@classmethod
- def from_string(cls, s: str) -> CoreSWHID:
+ def from_string(cls: Type[_TSWHID], s: str) -> _TSWHID:
parts = _parse_swhid(s)
if parts.pop("qualifiers"):
- raise ValidationError("CoreSWHID does not support qualifiers.")
+ raise ValidationError(f"{cls.__name__} does not support qualifiers.")
try:
- return CoreSWHID(**parts)
+ return cls(**parts)
except ValueError as e:
raise ValidationError(*e.args) from None
+@attr.s(frozen=True, kw_only=True)
+class CoreSWHID(_BaseSWHID[ObjectType]):
+ """
+ Dataclass holding the relevant info associated to a SoftWare Heritage
+ persistent IDentifier (SWHID).
+
+ Unlike `QualifiedSWHID`, it is restricted to core SWHIDs, ie. SWHIDs
+ with no qualifiers.
+
+ Raises:
+ swh.model.exceptions.ValidationError: In case of invalid object type or id
+
+ To get the raw SWHID string from an instance of this class,
+ use the :func:`str` function:
+
+ >>> swhid = CoreSWHID(
+ ... object_type=ObjectType.CONTENT,
+ ... object_id=bytes.fromhex('8ff44f081d43176474b267de5451f2c2e88089d0'),
+ ... )
+ >>> str(swhid)
+ 'swh:1:cnt:8ff44f081d43176474b267de5451f2c2e88089d0'
+
+ And vice-versa with :meth:`CoreSWHID.from_string`:
+
+ >>> swhid == CoreSWHID.from_string(
+ ... "swh:1:cnt:8ff44f081d43176474b267de5451f2c2e88089d0"
+ ... )
+ True
+ """
+
+ object_type = attr.ib(
+ type=ObjectType, validator=type_validator(), converter=ObjectType
+ )
+ """the type of object the identifier points to"""
+
+
def _parse_core_swhid(swhid: Union[str, CoreSWHID, None]) -> Optional[CoreSWHID]:
"""Alias of CoreSWHID.from_string to make mypy happy......
@@ -845,7 +877,7 @@ def _parse_lines_qualifier(
@attr.s(frozen=True, kw_only=True)
-class QualifiedSWHID:
+class QualifiedSWHID(_BaseSWHID[ObjectType]):
"""
Dataclass holding the relevant info associated to a SoftWare Heritage
persistent IDentifier (SWHID)
@@ -872,20 +904,11 @@ class QualifiedSWHID:
True
"""
- namespace = attr.ib(type=str, default=SWHID_NAMESPACE)
- """the namespace of the identifier, defaults to ``swh``"""
-
- scheme_version = attr.ib(type=int, default=SWHID_VERSION)
- """the scheme version of the identifier, defaults to 1"""
-
object_type = attr.ib(
type=ObjectType, validator=type_validator(), converter=ObjectType
)
"""the type of object the identifier points to"""
- object_id = attr.ib(type=bytes, validator=type_validator())
- """object's identifier"""
-
# qualifiers:
origin = attr.ib(type=Optional[str], default=None, validator=type_validator())
@@ -921,29 +944,6 @@ class QualifiedSWHID:
)
"""lines: line number(s) of interest, usually within a content object"""
- @namespace.validator
- def check_namespace(self, attribute, value):
- if value != SWHID_NAMESPACE:
- raise ValidationError(
- "Invalid SWHID: invalid namespace: %(namespace)s",
- params={"namespace": value},
- )
-
- @scheme_version.validator
- def check_scheme_version(self, attribute, value):
- if value != SWHID_VERSION:
- raise ValidationError(
- "Invalid SWHID: invalid version: %(version)s", params={"version": value}
- )
-
- @object_id.validator
- def check_object_id(self, attribute, value):
- if len(value) != 20:
- raise ValidationError(
- "Invalid SWHID: invalid checksum: %(object_id)s",
- params={"object_id": hash_to_hex(value)},
- )
-
@visit.validator
def check_visit(self, attribute, value):
if value and value.object_type != ObjectType.SNAPSHOT:
@@ -1006,7 +1006,7 @@ class QualifiedSWHID:
@attr.s(frozen=True, kw_only=True)
-class ExtendedSWHID:
+class ExtendedSWHID(_BaseSWHID[ExtendedObjectType]):
"""
Dataclass holding the relevant info associated to a SoftWare Heritage
persistent IDentifier (SWHID).
@@ -1035,12 +1035,6 @@ class ExtendedSWHID:
True
"""
- namespace = attr.ib(type=str, default=SWHID_NAMESPACE)
- """the namespace of the identifier, defaults to ``swh``"""
-
- scheme_version = attr.ib(type=int, default=SWHID_VERSION)
- """the scheme version of the identifier, defaults to 1"""
-
object_type = attr.ib(
type=ExtendedObjectType,
validator=type_validator(),
@@ -1048,49 +1042,6 @@ class ExtendedSWHID:
)
"""the type of object the identifier points to"""
- object_id = attr.ib(type=bytes, validator=type_validator())
- """object's identifier"""
-
- @namespace.validator
- def check_namespace(self, attribute, value):
- if value != SWHID_NAMESPACE:
- raise ValidationError(
- "Invalid SWHID: invalid namespace: %(namespace)s",
- params={"namespace": value},
- )
-
- @scheme_version.validator
- def check_scheme_version(self, attribute, value):
- if value != SWHID_VERSION:
- raise ValidationError(
- "Invalid SWHID: invalid version: %(version)s", params={"version": value}
- )
-
- @object_id.validator
- def check_object_id(self, attribute, value):
- if len(value) != 20:
- raise ValidationError(
- "Invalid SWHID: invalid checksum: %(object_id)s",
- params={"object_id": hash_to_hex(value)},
- )
-
- def __str__(self) -> str:
- return SWHID_SEP.join(
- [
- self.namespace,
- str(self.scheme_version),
- self.object_type.value,
- hash_to_hex(self.object_id),
- ]
- )
-
- @classmethod
- def from_string(cls, s: str) -> ExtendedSWHID:
- parts = _parse_swhid(s)
- if parts.pop("qualifiers"):
- raise ValidationError("ExtendedSWHID does not support qualifiers.")
- return ExtendedSWHID(**parts)
-
@attr.s(frozen=True)
class SWHID: