bellingcat/auto-archiver
1
0
Fork 0
mirror of https://github.com/bellingcat/auto-archiver.git synced 2026-06-12 05:08:28 +03:00
Code Issues Packages Projects Releases Wiki Activity

Updates tests to use pytest-mock.

Browse Source
This commit is contained in:
erinhmclark
2025-02-18 23:32:03 +00:00
parent 657fbd357d
commit f0fd9bf445
19 changed files with 584 additions and 600 deletions
Download Patch File Download Diff File Expand all files Collapse all files

52
tests/databases/test_api_db.py
View File

@@ -1,5 +1,3 @@
from unittest.mock import patch
import pytest
from auto_archiver.core import Metadata
@@ -35,35 +33,35 @@ def test_fetch_no_cache(api_db, metadata):
assert api_db.fetch(metadata) is None
def test_fetch_fail_status(api_db, metadata):
def test_fetch_fail_status(api_db, metadata, mocker):
# Test response fail in fetch method
with patch("auto_archiver.modules.api_db.api_db.requests.get") as mock_get:
mock_get.return_value.status_code = 400
mock_get.return_value.json.return_value = {}
with patch("loguru.logger.error") as mock_error:
assert api_db.fetch(metadata) is False
mock_error.assert_called_once_with("AA API FAIL (400): {}")
mock_get = mocker.patch("auto_archiver.modules.api_db.api_db.requests.get")
mock_get.return_value.status_code = 400
mock_get.return_value.json.return_value = {}
mock_error = mocker.patch("loguru.logger.error")
assert api_db.fetch(metadata) is False
mock_error.assert_called_once_with("AA API FAIL (400): {}")
def test_fetch(api_db, metadata):
def test_fetch(api_db, metadata, mocker):
# Test successful fetch method
with patch("auto_archiver.modules.api_db.api_db.requests.get") as mock_get,\
patch("auto_archiver.core.metadata.datetime.datetime") as mock_datetime:
mock_datetime.now.return_value = "2021-01-01T00:00:00"
mock_get.return_value.status_code = 200
mock_get.return_value.json.return_value = [{"result": {}}, {"result":
{'media': [], 'metadata': {'_processed_at': '2021-01-01T00:00:00', 'url': 'https://example.com'},
'status': 'no archiver'}}]
assert api_db.fetch(metadata) == metadata
mock_get = mocker.patch("auto_archiver.modules.api_db.api_db.requests.get")
mock_datetime = mocker.patch("auto_archiver.core.metadata.datetime.datetime")
mock_datetime.now.return_value = "2021-01-01T00:00:00"
mock_get.return_value.status_code = 200
mock_get.return_value.json.return_value = [{"result": {}}, {"result":
{'media': [], 'metadata': {'_processed_at': '2021-01-01T00:00:00', 'url': 'https://example.com'},
'status': 'no archiver'}}]
assert api_db.fetch(metadata) == metadata
def test_done_success(api_db, metadata):
with patch("auto_archiver.modules.api_db.api_db.requests.post") as mock_post:
mock_post.return_value.status_code = 201
api_db.done(metadata)
mock_post.assert_called_once()
mock_post.assert_called_once_with("https://api.example.com/interop/submit-archive",
json={'author_id': 'Someone', 'url': 'https://example.com',
'public': False, 'group_id': '123', 'tags': ['[', ']'], 'result': '{"status": "no archiver", "metadata": {"_processed_at": "2021-01-01T00:00:00", "url": "https://example.com"}, "media": []}'},
headers={'Authorization': 'Bearer test-token'})
def test_done_success(api_db, metadata, mocker):
mock_post = mocker.patch("auto_archiver.modules.api_db.api_db.requests.post")
mock_post.return_value.status_code = 201
api_db.done(metadata)
mock_post.assert_called_once()
mock_post.assert_called_once_with("https://api.example.com/interop/submit-archive",
json={'author_id': 'Someone', 'url': 'https://example.com',
'public': False, 'group_id': '123', 'tags': ['[', ']'], 'result': '{"status": "no archiver", "metadata": {"_processed_at": "2021-01-01T00:00:00", "url": "https://example.com"}, "media": []}'},
headers={'Authorization': 'Bearer test-token'})

39
tests/databases/test_gsheet_db.py
View File

@@ -1,6 +1,4 @@
from datetime import datetime, timezone
from unittest.mock import MagicMock, patch
import pytest
from auto_archiver.core import Metadata, Media
@@ -9,8 +7,8 @@ from auto_archiver.modules.gsheet_feeder import GWorksheet
@pytest.fixture
def mock_gworksheet():
mock_gworksheet = MagicMock(spec=GWorksheet)
def mock_gworksheet(mocker):
mock_gworksheet = mocker.MagicMock(spec=GWorksheet)
mock_gworksheet.col_exists.return_value = True
mock_gworksheet.get_cell.return_value = ""
mock_gworksheet.get_row.return_value = {}
@@ -18,14 +16,14 @@ def mock_gworksheet():
@pytest.fixture
def mock_metadata():
metadata: Metadata = MagicMock(spec=Metadata)
def mock_metadata(mocker):
metadata: Metadata = mocker.MagicMock(spec=Metadata)
metadata.get_url.return_value = "http://example.com"
metadata.status = "done"
metadata.get_title.return_value = "Example Title"
metadata.get.return_value = "Example Content"
metadata.get_timestamp.return_value = "2025-01-01T00:00:00"
metadata.get_final_media.return_value = MagicMock(spec=Media)
metadata.get_final_media.return_value = mocker.MagicMock(spec=Media)
metadata.get_all_media.return_value = []
metadata.get_media_by_id.return_value = None
metadata.get_first_image.return_value = None
@@ -47,21 +45,21 @@ def metadata():
@pytest.fixture
def mock_media():
def mock_media(mocker):
"""Fixture for a mock Media object."""
mock_media = MagicMock(spec=Media)
mock_media = mocker.MagicMock(spec=Media)
mock_media.urls = ["http://example.com/media"]
mock_media.get.return_value = "not-calculated"
return mock_media
@pytest.fixture
def gsheets_db(mock_gworksheet, setup_module):
def gsheets_db(mock_gworksheet, setup_module, mocker):
db = setup_module("gsheet_db", {
"allow_worksheets": "set()",
"block_worksheets": "set()",
"use_sheet_names_in_stored_paths": "True",
})
db._retrieve_gsheet = MagicMock(return_value=(mock_gworksheet, 1))
db._retrieve_gsheet = mocker.MagicMock(return_value=(mock_gworksheet, 1))
return db
@@ -109,27 +107,26 @@ def test_aborted(gsheets_db, mock_metadata, mock_gworksheet):
mock_gworksheet.set_cell.assert_called_once_with(1, 'status', '')
def test_done(gsheets_db, metadata, mock_gworksheet, expected_calls):
with patch("auto_archiver.modules.gsheet_db.gsheet_db.get_current_timestamp", return_value='2025-02-01T00:00:00+00:00'):
gsheets_db.done(metadata)
def test_done(gsheets_db, metadata, mock_gworksheet, expected_calls, mocker):
mocker.patch("auto_archiver.modules.gsheet_db.gsheet_db.get_current_timestamp", return_value='2025-02-01T00:00:00+00:00')
gsheets_db.done(metadata)
mock_gworksheet.batch_set_cell.assert_called_once_with(expected_calls)
def test_done_cached(gsheets_db, metadata, mock_gworksheet):
with patch("auto_archiver.modules.gsheet_db.gsheet_db.get_current_timestamp", return_value='2025-02-01T00:00:00+00:00'):
gsheets_db.done(metadata, cached=True)
def test_done_cached(gsheets_db, metadata, mock_gworksheet, mocker):
mocker.patch("auto_archiver.modules.gsheet_db.gsheet_db.get_current_timestamp", return_value='2025-02-01T00:00:00+00:00')
gsheets_db.done(metadata, cached=True)
# Verify the status message includes "[cached]"
call_args = mock_gworksheet.batch_set_cell.call_args[0][0]
assert any(call[2].startswith("[cached]") for call in call_args)
def test_done_missing_media(gsheets_db, metadata, mock_gworksheet):
def test_done_missing_media(gsheets_db, metadata, mock_gworksheet, mocker):
# clear media from metadata
metadata.media = []
with patch("auto_archiver.modules.gsheet_db.gsheet_db.get_current_timestamp",
return_value='2025-02-01T00:00:00+00:00'):
gsheets_db.done(metadata)
mocker.patch("auto_archiver.modules.gsheet_db.gsheet_db.get_current_timestamp", return_value='2025-02-01T00:00:00+00:00')
gsheets_db.done(metadata)
# Verify nothing media-related gets updated
call_args = mock_gworksheet.batch_set_cell.call_args[0][0]
media_fields = {'archive', 'screenshot', 'thumbnail', 'wacz', 'replaywebpage'}