bellingcat/auto-archiver
1
0
Fork 0
mirror of https://github.com/bellingcat/auto-archiver.git synced 2026-06-12 13:18:28 +03:00
Code Issues Packages Projects Releases Wiki Activity

Fix the extractor args for new list structure.

Browse Source
This commit is contained in:
erinhmclark
2025-03-17 14:17:31 +00:00
parent f5bbfe5d1c
commit 5daeae994a
2 changed files with 53 additions and 37 deletions
Download Patch File Download Diff File Expand all files Collapse all files

37
src/auto_archiver/modules/generic_extractor/generic_extractor.py
View File

@@ -404,12 +404,20 @@ class GenericExtractor(Extractor):
"--write-subs" if self.subtitles else "--no-write-subs",
"--write-auto-subs" if self.subtitles else "--no-write-auto-subs",
"--live-from-start" if self.live_from_start else "--no-live-from-start",
"--proxy",
self.proxy if self.proxy else "",
f"--max-downloads {self.max_downloads}" if self.max_downloads != "inf" else "",
f"--playlist-end {self.max_downloads}" if self.max_downloads != "inf" else "",
# TODO: Move this to documentation
# Note: add the --verbose flag for debugging
"--verbose",
]
# proxy handling
if self.proxy:
ydl_options.extend(["--proxy", self.proxy])
# max_downloads handling
if self.max_downloads != "inf":
ydl_options.extend(["--max-downloads", str(self.max_downloads)])
ydl_options.extend(["--playlist-end", str(self.max_downloads)])
# set up auth
auth = self.auth_for_site(url, extract_cookies=False)
# order of importance: username/password -> api_key -> cookie -> cookies_from_browser -> cookies_file
@@ -428,22 +436,15 @@ class GenericExtractor(Extractor):
logger.debug(f"Using cookies from file {auth['cookies_file']} for {url}")
ydl_options.extend(("--cookies", auth["cookies_file"]))
# Applying user-defined extractor_args
if self.extractor_args:
logger.info(f"Applying user-defined extractor_args")
ydl_options.setdefault('extractor_args', {})
for key, args in self.extractor_args.items():
logger.debug(f"Setting extractor_args: {key}")
if isinstance(args, dict):
# Site specific arguments (e.g., youtube: somekey=value)
ydl_options['extractor_args'].setdefault(key, {}).update(args)
else:
# General extractor_args (e.g., somekey=value)
ydl_options['extractor_args'][key] = args
for key, args in self.extractor_args.items():
logger.debug(f"Setting extractor_args: {key}")
if isinstance(args, dict):
arg_str = ";".join(f"{k}={v}" for k, v in args.items())
else:
arg_str = str(args)
ydl_options.extend(["--extractor-args", f"{key}:{arg_str}"])
if self.ytdlp_args:
logger.debug("Adding additional ytdlp arguments: {self.ytdlp_args}")