bellingcat/vk-url-scraper
1
0
Fork 0
mirror of https://github.com/bellingcat/vk-url-scraper.git synced 2026-06-07 19:08:38 +03:00
Code Issues Packages Projects Releases Wiki Activity

fix: private videos

Browse Source
This commit is contained in:
msramalho
2023-08-18 21:15:34 +01:00
parent 12a5d22f64
commit a1c098335c
3 changed files with 28 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
tests/scraper_test.py
View File

@@ -150,3 +150,21 @@ def test_scrape_video_only2():
vks.download_media(res, tempdir)
found_files = set(os.listdir(tempdir))
assert "video-17546758_456239898_0.mp4" in found_files
def test_scrape_private_video():
"""
> Some videos are kept private and cannot be accessed without a passkey . In this case, send the ID in {owner_id}_{video_id}_{access_key}.
From https://dev.vk.com/ru/method/video.get
"""
res = vks.scrape("https://vk.com/wall-127774884_178565")
with tempfile.TemporaryDirectory(dir="./") as tempdir:
vks.download_media(res, tempdir)
expect_files = {
"wall-127774884_178565_0.mp4",
"wall-127774884_178565_1.mp4",
"wall-127774884_178565_2.mp4",
}
found_files = set(os.listdir(tempdir))
assert len(expect_files) == len(expect_files & found_files)

4
vk_url_scraper/__main__.py
View File

@@ -19,7 +19,7 @@ def get_argument_parser():
action="store",
dest="username",
required=True,
help="username for a valid vk.com account",
help="username for a valid vk.com account (pass empty if using --token)",
)
parser.add_argument(
"-p",
@@ -27,7 +27,7 @@ def get_argument_parser():
action="store",
dest="password",
required=True,
help="password for the valid vk.com account",
help="password for the valid vk.com account (pass empty if using --token)",
)
parser.add_argument(
"-t",

14
vk_url_scraper/scraper.py
View File

@@ -37,7 +37,7 @@ class VkScraper:
WALL_PATTERN = re.compile(r"(wall.{0,1}\d+_\d+)")
PHOTO_PATTERN = re.compile(r"(photo.{0,1}\d+_\d+)")
VIDEO_PATTERN = re.compile(r"(video.{0,1}\d+_\d+)")
VIDEO_PATTERN = re.compile(r"(video.{0,1}\d+_\d+(?:_\w+)?)")
def __init__(
self,
@@ -144,10 +144,11 @@ class VkScraper:
first_type = a["type"]
attachment = a[first_type]
if first_type == "video":
video_path = f'video{attachment["owner_id"]}_{attachment["id"]}'
if "access_key" in attachment:
video_path += f"_{attachment['access_key']}"
attachments["video"].extend(
self.scrape_videos(f'video{attachment["owner_id"]}_{attachment["id"]}')[
0
]
self.scrape_videos(video_path)[0]
.get("attachments", {})
.get("video", [""])
)
@@ -352,9 +353,10 @@ class VkScraper:
info = ydl.extract_info(url, download=True)
filename = ydl.prepare_filename(info)
if "unknown_video" in filename:
old_filename = filename
filename = shutil.copy(
filename, filename.replace("unknown_video", "mkv")
filename, filename.replace("unknown_video", "mp4")
)
os.remove(filename)
os.remove(old_filename)
downloaded.append(filename)
return downloaded