bellingcat/auto-archiver
1
0
Fork 0
mirror of https://github.com/bellingcat/auto-archiver.git synced 2026-06-08 03:18:28 +03:00
Code Issues Packages Projects Releases Wiki Activity

adds clips extraction to VK, though generic_extractor should still be run for those

Browse Source
This commit is contained in:
msramalho
2025-06-08 14:36:55 +01:00
parent 1f2d637928
commit 6f02493ff1
2 changed files with 30 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
tests/extractors/test_antibot_dropin_vk.py
View File

@@ -5,6 +5,15 @@ from auto_archiver.modules.antibot_extractor_enricher.dropins.vk import VkDropin
@pytest.mark.parametrize(
"input_url,expected",
[
# Unrelated URL, should return unchanged
(
"https://vk.com/id123456",
"https://vk.com/id123456",
),
(
"https://example.com/",
"https://example.com/",
),
# Wall post modal URL
(
"https://vk.com/somepage?w=wall-123456_7890",
@@ -53,14 +62,25 @@ from auto_archiver.modules.antibot_extractor_enricher.dropins.vk import VkDropin
"https://vk.com/video-111222_3334",
"https://vk.com/video-111222_3334",
),
# Unrelated URL, should return unchanged
# Clip modal URL
(
"https://vk.com/id123456",
"https://vk.com/id123456",
"https://vk.com/somepage?w=clip-555666_7778",
"https://vk.com/clip-555666_7778",
),
# Clip modal URL with no dash
(
"https://example.com/",
"https://example.com/",
"https://vk.com/somepage?w=clip555666_7778",
"https://vk.com/clip555666_7778",
),
# Clip modal URL with extra part
(
"https://vk.com/somepage?w=clip-555666_7778_ABC",
"https://vk.com/clip-555666_7778",
),
# No modal, should return unchanged (clip)
(
"https://vk.com/clip-555666_7778",
"https://vk.com/clip-555666_7778",
),
# Modal with multiple params, should still work with right priority
(