Differential D5081 Diff 18232 swh/model/tests/test_identifiers.py

Changeset View

Standalone View

swh/model/tests/test_identifiers.py

# Copyright (C) 2015-2018 The Software Heritage developers		# Copyright (C) 2015-2021 The Software Heritage developers
# See the AUTHORS file at the top-level directory of this distribution		# See the AUTHORS file at the top-level directory of this distribution
# License: GNU General Public License version 3, or any later version		# License: GNU General Public License version 3, or any later version
# See top-level LICENSE file for more information		# See top-level LICENSE file for more information

import binascii		import binascii
import datetime		import datetime
from typing import Dict		from typing import Dict
import unittest		import unittest

import pytest		import pytest

from swh.model import hashutil, identifiers		from swh.model import hashutil, identifiers
from swh.model.exceptions import ValidationError		from swh.model.exceptions import ValidationError
from swh.model.hashutil import hash_to_bytes as _x		from swh.model.hashutil import hash_to_bytes as _x
from swh.model.identifiers import (		from swh.model.identifiers import (
CONTENT,		CONTENT,
DIRECTORY,		DIRECTORY,
RELEASE,		RELEASE,
REVISION,		REVISION,
SNAPSHOT,		SNAPSHOT,
SWHID,		SWHID,
		ObjectType,
		QualifiedSWHID,
normalize_timestamp,		normalize_timestamp,
)		)


def remove_id(d: Dict) -> Dict:		def remove_id(d: Dict) -> Dict:
"""Returns a (shallow) copy of a dict with the 'id' key removed."""		"""Returns a (shallow) copy of a dict with the 'id' key removed."""
d = d.copy()		d = d.copy()
if "id" in d:		if "id" in d:
▲ Show 20 Lines • Show All 847 Lines • ▼ Show 20 Lines

@pytest.mark.parametrize("dict_input", TS_DICTS_INVALID_TIMESTAMP)		@pytest.mark.parametrize("dict_input", TS_DICTS_INVALID_TIMESTAMP)
def test_normalize_timestamp_dict_invalid_timestamp(dict_input):		def test_normalize_timestamp_dict_invalid_timestamp(dict_input):
with pytest.raises(ValueError, match="non-integer timestamp"):		with pytest.raises(ValueError, match="non-integer timestamp"):
normalize_timestamp(dict_input)		normalize_timestamp(dict_input)


class TestSwhid(unittest.TestCase):		class TestSwhid(unittest.TestCase):
		@pytest.mark.filterwarnings("ignore:.SWHID.:DeprecationWarning")
def test_swhid(self):		def test_swhid(self):
_snapshot_id = _x("c7c108084bc0bf3d81436bf980b46e98bd338453")		_snapshot_id = _x("c7c108084bc0bf3d81436bf980b46e98bd338453")
_release_id = "22ece559cc7cc2364edc5e5593d63ae8bd229f9f"		_release_id = "22ece559cc7cc2364edc5e5593d63ae8bd229f9f"
_revision_id = "309cf2674ee7a0749978cf8265ab91a60aea0f7d"		_revision_id = "309cf2674ee7a0749978cf8265ab91a60aea0f7d"
_directory_id = "d198bc9d7a6bcf6db04f476d29314f157507d505"		_directory_id = "d198bc9d7a6bcf6db04f476d29314f157507d505"
_content_id = "94a9ed024d3859793618152ea559a168bbcbb5e2"		_content_id = "94a9ed024d3859793618152ea559a168bbcbb5e2"
_snapshot = {"id": _snapshot_id}		_snapshot = {"id": _snapshot_id}
_release = {"id": _release_id}		_release = {"id": _release_id}
▲ Show 20 Lines • Show All 96 Lines • ▼ Show 20 Lines	def test_swhid_wrong_input(self):
for _type, _hash in [		for _type, _hash in [
(SNAPSHOT, _snapshot_id),		(SNAPSHOT, _snapshot_id),
(SNAPSHOT, _snapshot),		(SNAPSHOT, _snapshot),
("lines", "42"),		("lines", "42"),
]:		]:
with self.assertRaises(ValidationError):		with self.assertRaises(ValidationError):
identifiers.swhid(_type, _hash)		identifiers.swhid(_type, _hash)

		@pytest.mark.filterwarnings("ignore:.SWHID.:DeprecationWarning")
def test_parse_swhid(self):		def test_parse_swhid(self):
for swhid, _type, _version, _hash in [		for swhid, _type, _version, _hash in [
(		(
"swh:1:cnt:94a9ed024d3859793618152ea559a168bbcbb5e2",		"swh:1:cnt:94a9ed024d3859793618152ea559a168bbcbb5e2",
CONTENT,		CONTENT,
1,		1,
"94a9ed024d3859793618152ea559a168bbcbb5e2",		"94a9ed024d3859793618152ea559a168bbcbb5e2",
),		),
Show All 17 Lines	def test_parse_swhid(self):
),		),
(		(
"swh:1:snp:c7c108084bc0bf3d81436bf980b46e98bd338453",		"swh:1:snp:c7c108084bc0bf3d81436bf980b46e98bd338453",
SNAPSHOT,		SNAPSHOT,
1,		1,
"c7c108084bc0bf3d81436bf980b46e98bd338453",		"c7c108084bc0bf3d81436bf980b46e98bd338453",
),		),
]:		]:
		with pytest.warns(DeprecationWarning):
expected_result = SWHID(		expected_result = SWHID(
namespace="swh",		namespace="swh",
scheme_version=_version,		scheme_version=_version,
object_type=_type,		object_type=_type,
object_id=_hash,		object_id=_hash,
metadata={},		metadata={},
)		)
actual_result = identifiers.parse_swhid(swhid)		actual_result = identifiers.parse_swhid(swhid)
self.assertEqual(actual_result, expected_result)		self.assertEqual(actual_result, expected_result)

for swhid, _type, _version, _hash, _metadata in [		for swhid, _type, _version, _hash, _metadata in [
(		(
"swh:1:cnt:9c95815d9e9d91b8dae8e05d8bbc696fe19f796b;lines=1-18;origin=https://github.com/python/cpython", # noqa		"swh:1:cnt:9c95815d9e9d91b8dae8e05d8bbc696fe19f796b;lines=1-18;origin=https://github.com/python/cpython", # noqa
CONTENT,		CONTENT,
1,		1,
"9c95815d9e9d91b8dae8e05d8bbc696fe19f796b",		"9c95815d9e9d91b8dae8e05d8bbc696fe19f796b",
{"lines": "1-18", "origin": "https://github.com/python/cpython"},		{"lines": "1-18", "origin": "https://github.com/python/cpython"},
),		),
(		(
"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;origin=deb://Debian/packages/linuxdoc-tools", # noqa		"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;origin=deb://Debian/packages/linuxdoc-tools", # noqa
DIRECTORY,		DIRECTORY,
1,		1,
"0b6959356d30f1a4e9b7f6bca59b9a336464c03d",		"0b6959356d30f1a4e9b7f6bca59b9a336464c03d",
{"origin": "deb://Debian/packages/linuxdoc-tools"},		{"origin": "deb://Debian/packages/linuxdoc-tools"},
),		),
]:		]:
		with pytest.warns(DeprecationWarning):
expected_result = SWHID(		expected_result = SWHID(
namespace="swh",		namespace="swh",
scheme_version=_version,		scheme_version=_version,
object_type=_type,		object_type=_type,
object_id=_hash,		object_id=_hash,
metadata=_metadata,		metadata=_metadata,
)		)
actual_result = identifiers.parse_swhid(swhid)		actual_result = identifiers.parse_swhid(swhid)
self.assertEqual(actual_result, expected_result)		self.assertEqual(actual_result, expected_result)
self.assertEqual(		self.assertEqual(
expected_result.to_dict(),		expected_result.to_dict(),
{		{
"namespace": "swh",		"namespace": "swh",
"scheme_version": _version,		"scheme_version": _version,
"object_type": _type,		"object_type": _type,
"object_id": _hash,		"object_id": _hash,
▲ Show 20 Lines • Show All 44 Lines • ▼ Show 20 Lines	[
"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;lines=12\v",		"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;lines=12\v",
],		],
)		)
def test_parse_swhid_parsing_error(invalid_swhid):		def test_parse_swhid_parsing_error(invalid_swhid):
with pytest.raises(ValidationError):		with pytest.raises(ValidationError):
identifiers.parse_swhid(invalid_swhid)		identifiers.parse_swhid(invalid_swhid)


		@pytest.mark.filterwarnings("ignore:.SWHID.:DeprecationWarning")
@pytest.mark.parametrize(		@pytest.mark.parametrize(
"ns,version,type,id",		"ns,version,type,id",
[		[
("foo", 1, CONTENT, "abc8bc9d7a6bcf6db04f476d29314f157507d505",),		("foo", 1, CONTENT, "abc8bc9d7a6bcf6db04f476d29314f157507d505",),
("swh", 2, DIRECTORY, "def8bc9d7a6bcf6db04f476d29314f157507d505",),		("swh", 2, DIRECTORY, "def8bc9d7a6bcf6db04f476d29314f157507d505",),
("swh", 1, "foo", "fed8bc9d7a6bcf6db04f476d29314f157507d505",),		("swh", 1, "foo", "fed8bc9d7a6bcf6db04f476d29314f157507d505",),
("swh", 1, SNAPSHOT, "gh6959356d30f1a4e9b7f6bca59b9a336464c03d",),		("swh", 1, SNAPSHOT, "gh6959356d30f1a4e9b7f6bca59b9a336464c03d",),
],		],
)		)
def test_SWHID_class_validation_error(ns, version, type, id):		def test_SWHID_class_validation_error(ns, version, type, id):
with pytest.raises(ValidationError):		with pytest.raises(ValidationError):
SWHID(		SWHID(
namespace=ns, scheme_version=version, object_type=type, object_id=id,		namespace=ns, scheme_version=version, object_type=type, object_id=id,
)		)


def test_swhid_hash():		@pytest.mark.filterwarnings("ignore:.SWHID.:DeprecationWarning")
		def test_SWHID_hash():
object_id = "94a9ed024d3859793618152ea559a168bbcbb5e2"		object_id = "94a9ed024d3859793618152ea559a168bbcbb5e2"

assert hash(SWHID(object_type="directory", object_id=object_id)) == hash(		assert hash(SWHID(object_type="directory", object_id=object_id)) == hash(
SWHID(object_type="directory", object_id=object_id)		SWHID(object_type="directory", object_id=object_id)
)		)

assert hash(		assert hash(
SWHID(object_type="directory", object_id=object_id, metadata=dummy_qualifiers,)		SWHID(object_type="directory", object_id=object_id, metadata=dummy_qualifiers,)
Show All 13 Lines	) == hash(
SWHID(		SWHID(
object_type="directory",		object_type="directory",
object_id=object_id,		object_id=object_id,
metadata={"lines": "42", "origin": "https://example.com"},		metadata={"lines": "42", "origin": "https://example.com"},
)		)
)		)


def test_swhid_eq():		@pytest.mark.filterwarnings("ignore:.SWHID.:DeprecationWarning")
		def test_SWHID_eq():
object_id = "94a9ed024d3859793618152ea559a168bbcbb5e2"		object_id = "94a9ed024d3859793618152ea559a168bbcbb5e2"

assert SWHID(object_type="directory", object_id=object_id) == SWHID(		assert SWHID(object_type="directory", object_id=object_id) == SWHID(
object_type="directory", object_id=object_id		object_type="directory", object_id=object_id
)		)

assert SWHID(		assert SWHID(
object_type="directory", object_id=object_id, metadata=dummy_qualifiers,		object_type="directory", object_id=object_id, metadata=dummy_qualifiers,
) == SWHID(object_type="directory", object_id=object_id, metadata=dummy_qualifiers,)		) == SWHID(object_type="directory", object_id=object_id, metadata=dummy_qualifiers,)

assert SWHID(		assert SWHID(
object_type="directory", object_id=object_id, metadata=dummy_qualifiers,		object_type="directory", object_id=object_id, metadata=dummy_qualifiers,
) == SWHID(object_type="directory", object_id=object_id, metadata=dummy_qualifiers,)		) == SWHID(object_type="directory", object_id=object_id, metadata=dummy_qualifiers,)


		def test_parse_qualified_swhid():
		for swhid, _type, _version, _hash in [
		(
		"swh:1:cnt:94a9ed024d3859793618152ea559a168bbcbb5e2",
		ObjectType.CONTENT,
		1,
		ardumontUnsubmitted Not Done Inline Actions what's the `_x` ? (it's not immediatly apparent to me). ardumont: what's the `_x` ? (it's not immediatly apparent to me).
		vlorentzAuthorUnsubmitted Done Inline Actions hash_to_bytes. it's used everywhere in this module so it makes sense to have a short alias vlorentz: hash_to_bytes. it's used everywhere in this module so it makes sense to have a short alias
		ardumontUnsubmitted Not Done Inline Actions lol, i noticed it after writing the comment and removed the comment (or so i thought)... sorry for the noise. ardumont: lol, i noticed it after writing the comment and removed the comment (or so i thought)... sorry…
		_x("94a9ed024d3859793618152ea559a168bbcbb5e2"),
		),
		(
		"swh:1:dir:d198bc9d7a6bcf6db04f476d29314f157507d505",
		ObjectType.DIRECTORY,
		1,
		_x("d198bc9d7a6bcf6db04f476d29314f157507d505"),
		),
		(
		"swh:1:rev:309cf2674ee7a0749978cf8265ab91a60aea0f7d",
		ObjectType.REVISION,
		1,
		_x("309cf2674ee7a0749978cf8265ab91a60aea0f7d"),
		),
		(
		"swh:1:rel:22ece559cc7cc2364edc5e5593d63ae8bd229f9f",
		ObjectType.RELEASE,
		1,
		_x("22ece559cc7cc2364edc5e5593d63ae8bd229f9f"),
		),
		(
		"swh:1:snp:c7c108084bc0bf3d81436bf980b46e98bd338453",
		ObjectType.SNAPSHOT,
		1,
		_x("c7c108084bc0bf3d81436bf980b46e98bd338453"),
		),
		]:
		expected_result = QualifiedSWHID(
		namespace="swh",
		scheme_version=_version,
		object_type=_type,
		object_id=_hash,
		qualifiers={},
		)
		actual_result = QualifiedSWHID.from_string(swhid)
		assert actual_result == expected_result

		for swhid, _type, _version, _hash, _qualifiers in [
		(
		"swh:1:cnt:9c95815d9e9d91b8dae8e05d8bbc696fe19f796b;lines=1-18;origin=https://github.com/python/cpython", # noqa
		ObjectType.CONTENT,
		1,
		_x("9c95815d9e9d91b8dae8e05d8bbc696fe19f796b"),
		{"lines": "1-18", "origin": "https://github.com/python/cpython"},
		),
		(
		"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;origin=deb://Debian/packages/linuxdoc-tools", # noqa
		ObjectType.DIRECTORY,
		1,
		_x("0b6959356d30f1a4e9b7f6bca59b9a336464c03d"),
		{"origin": "deb://Debian/packages/linuxdoc-tools"},
		),
		]:
		expected_result = QualifiedSWHID(
		namespace="swh",
		scheme_version=_version,
		object_type=_type,
		object_id=_hash,
		qualifiers=_qualifiers,
		)
		actual_result = QualifiedSWHID.from_string(swhid)
		assert actual_result == expected_result
		assert expected_result.to_dict() == {
		"namespace": "swh",
		"scheme_version": _version,
		"object_type": _type,
		"object_id": _hash,
		"qualifiers": _qualifiers,
		}


		@pytest.mark.parametrize(
		"invalid_swhid",
		[
		"swh:1:cnt",
		"swh:1:",
		"swh:",
		"swh:1:cnt:",
		"foo:1:cnt:abc8bc9d7a6bcf6db04f476d29314f157507d505",
		"swh:2:dir:def8bc9d7a6bcf6db04f476d29314f157507d505",
		"swh:1:foo:fed8bc9d7a6bcf6db04f476d29314f157507d505",
		"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;invalid;malformed",
		"swh:1:snp:gh6959356d30f1a4e9b7f6bca59b9a336464c03d",
		"swh:1:snp:foo",
		# wrong qualifier: ori should be origin
		"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;ori=something;anchor=1;visit=1;path=/", # noqa
		# wrong qualifier: anc should be anchor
		"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;origin=something;anc=1;visit=1;path=/", # noqa
		# wrong qualifier: vis should be visit
		"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;origin=something;anchor=1;vis=1;path=/", # noqa
		# wrong qualifier: pa should be path
		"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;origin=something;anchor=1;visit=1;pa=/", # noqa
		# wrong qualifier: line should be lines
		"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;line=10;origin=something;anchor=1;visit=1;path=/", # noqa
		# wrong qualifier value: it contains space before of after
		"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;origin= https://some-url", # noqa
		"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;origin=something;anchor=some-anchor ", # noqa
		"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;origin=something;anchor=some-anchor ;visit=1", # noqa
		# invalid swhid: whitespaces
		"swh :1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;ori=something;anchor=1;visit=1;path=/", # noqa
		"swh: 1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;ori=something;anchor=1;visit=1;path=/", # noqa
		"swh: 1: dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;ori=something;anchor=1;visit=1;path=/", # noqa
		"swh:1: dir: 0b6959356d30f1a4e9b7f6bca59b9a336464c03d",
		"swh:1: dir: 0b6959356d30f1a4e9b7f6bca59b9a336464c03d; origin=blah",
		"swh:1: dir: 0b6959356d30f1a4e9b7f6bca59b9a336464c03d;lines=12",
		# other whitespaces
		"swh\t:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;lines=12",
		"swh:1\n:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;lines=12",
		"swh:1:\rdir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;lines=12",
		"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d\f;lines=12",
		"swh:1:dir:0b6959356d30f1a4e9b7f6bca59b9a336464c03d;lines=12\v",
		],
		)
		def test_parse_qualified_swhid_parsing_error(invalid_swhid):
		with pytest.raises(ValidationError):
		QualifiedSWHID.from_string(invalid_swhid)


		@pytest.mark.filterwarnings("ignore:.SWHID.:DeprecationWarning")
		@pytest.mark.parametrize(
		"ns,version,type,id",
		[
		("foo", 1, ObjectType.CONTENT, "abc8bc9d7a6bcf6db04f476d29314f157507d505",),
		("swh", 2, ObjectType.DIRECTORY, "def8bc9d7a6bcf6db04f476d29314f157507d505",),
		],
		)
		def test_QualifiedSWHID_validation_error(ns, version, type, id):
		with pytest.raises(ValidationError):
		QualifiedSWHID(
		namespace=ns, scheme_version=version, object_type=type, object_id=_x(id),
		)


		def test_QualifiedSWHID_hash():
		object_id = _x("94a9ed024d3859793618152ea559a168bbcbb5e2")

		assert hash(
		QualifiedSWHID(object_type=ObjectType.DIRECTORY, object_id=object_id)
		) == hash(QualifiedSWHID(object_type=ObjectType.DIRECTORY, object_id=object_id))

		assert hash(
		QualifiedSWHID(
		object_type=ObjectType.DIRECTORY,
		object_id=object_id,
		qualifiers=dummy_qualifiers,
		)
		) == hash(
		QualifiedSWHID(
		object_type=ObjectType.DIRECTORY,
		object_id=object_id,
		qualifiers=dummy_qualifiers,
		)
		)

		# Different order of the dictionary, so the underlying order of the tuple in
		# ImmutableDict is different.
		assert hash(
		QualifiedSWHID(
		object_type=ObjectType.DIRECTORY,
		object_id=object_id,
		qualifiers={"origin": "https://example.com", "lines": "42"},
		)
		) == hash(
		QualifiedSWHID(
		object_type=ObjectType.DIRECTORY,
		object_id=object_id,
		qualifiers={"lines": "42", "origin": "https://example.com"},
		)
		)


		def test_QualifiedSWHID_eq():
		object_id = _x("94a9ed024d3859793618152ea559a168bbcbb5e2")

		assert QualifiedSWHID(
		object_type=ObjectType.DIRECTORY, object_id=object_id
		) == QualifiedSWHID(object_type=ObjectType.DIRECTORY, object_id=object_id)

		assert QualifiedSWHID(
		object_type=ObjectType.DIRECTORY,
		object_id=object_id,
		qualifiers=dummy_qualifiers,
		) == QualifiedSWHID(
		object_type=ObjectType.DIRECTORY,
		object_id=object_id,
		qualifiers=dummy_qualifiers,
		)

		assert QualifiedSWHID(
		object_type=ObjectType.DIRECTORY,
		object_id=object_id,
		qualifiers=dummy_qualifiers,
		) == QualifiedSWHID(
		object_type=ObjectType.DIRECTORY,
		object_id=object_id,
		qualifiers=dummy_qualifiers,
		)