github/gallery-dl
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[job] implement 'follow' option (#8752)

Browse Source 
Follow and process URLs found in the given format string result.
This commit is contained in:
Mike Fährmann
2026-02-07 11:41:32 +01:00
parent c978fe18d4
commit 2d64e76223
4 changed files with 90 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

34
gallery_dl/job.py
View File

@@ -211,13 +211,31 @@ class Job():
msg = None
process = True
if follow := self.extractor.config("follow"):
follow = formatter.parse(follow, None, util.identity).format_map
follow_urls = follow_kwdict = None
else:
follow = follow_urls = None
for msg, url, kwdict in messages:
if msg == Message.Directory:
if follow_urls is not None:
for furl in follow_urls:
if self.metadata_url:
follow_kwdict[self.metadata_url] = furl
if self.pred_queue(furl, follow_kwdict):
self.handle_queue(furl, follow_kwdict)
follow_urls = None
self.update_kwdict(kwdict)
if self.pred_post(url, kwdict):
process = True
self.handle_directory(kwdict)
if follow is not None:
follow_urls = self._collect_urls(follow(kwdict))
if follow_urls is not None:
follow_kwdict = kwdict.copy()
else:
process = None
if FLAGS.POST is not None:
@@ -253,6 +271,13 @@ class Job():
if FLAGS.CHILD is not None:
FLAGS.process("CHILD")
if follow_urls is not None:
for furl in follow_urls:
if self.metadata_url:
follow_kwdict[self.metadata_url] = furl
if self.pred_queue(furl, follow_kwdict):
self.handle_queue(furl, follow_kwdict)
return msg
def handle_url(self, url, kwdict):
@@ -300,6 +325,15 @@ class Job():
if init and init != "lazy":
self.initialize()
def _collect_urls(self, source):
if not source:
return None
if isinstance(source, list):
return source
if isinstance(source, str):
if urls := text.extract_urls(source):
return urls
def _prepare_predicates(self, target, alt=None, skip=None):
predicates = []
extr = self.extractor