Git-Mirrors/brozzler
1
0
Fork 0
mirror of https://github.com/internetarchive/brozzler.git synced 2025-02-23 08:09:48 -05:00
Code Issues Actions Packages Projects Releases Wiki Activity

Merge pull request #312 from internetarchive/adam/patch-yt-dlp-infinite-loop-bug

Browse Source 
feat: override yt-dlp generic extractor to add redirect loop detectio…
This commit is contained in:
Adam Miller 2025-01-23 15:30:56 -08:00 committed by GitHub
commit 1e30b4f478
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 28 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

29
brozzler/ydl.py
View File

@ -18,7 +18,7 @@ limitations under the License.
import logging
import yt_dlp
from yt_dlp.utils import match_filter_func
from yt_dlp.utils import match_filter_func, ExtractorError
import brozzler
import urllib.request
import tempfile
@ -37,6 +37,7 @@ thread_local = threading.local()
YTDLP_PROXY = ""
PROXY_ATTEMPTS = 4
YTDLP_WAIT = 10
YTDLP_MAX_REDIRECTS = 5
def should_ytdlp(site, page, page_status, skip_av_seeds):
@ -113,6 +114,28 @@ def _build_youtube_dl(worker, destdir, site, page):
class _YoutubeDL(yt_dlp.YoutubeDL):
logger = logging.getLogger(__module__ + "." + __qualname__)
def process_ie_result(self, ie_result, download=True, extra_info=None):
if extra_info is None:
extra_info = {}
result_type = ie_result.get("_type", "video")
if result_type in ("url", "url_transparent"):
if "extraction_depth" in extra_info:
self.logger.info(
f"Following redirect URL: {ie_result['url']} extraction_depth: {extra_info['extraction_depth']}"
)
extra_info["extraction_depth"] = 1 + extra_info.get(
"extraction_depth", 0
)
else:
extra_info["extraction_depth"] = 0
if extra_info["extraction_depth"] >= YTDLP_MAX_REDIRECTS:
raise ExtractorError(
f"Too many hops for URL: {ie_result['url']}",
expected=True,
)
return super().process_ie_result(ie_result, download, extra_info)
def add_default_extra_info(self, ie_result, ie, url):
# hook in some logging
super().add_default_extra_info(ie_result, ie, url)
@ -361,6 +384,10 @@ def _try_youtube_dl(worker, ydl, site, page):
and e.exc_info[1].code == 420
):
raise brozzler.ReachedLimit(e.exc_info[1])
elif isinstance(e, yt_dlp.utils.DownloadError) and (
"Redirect loop detected" in e.msg or "Too many redirects" in e.msg
):
raise brozzler.VideoExtractorError(e.msg)
else:
# todo: other errors to handle separately?
# OSError('Tunnel connection failed: 464 Host Not Allowed') (caused by ProxyError...)