sherl/ythdd
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: comment reply count extraction

Browse Source 
suspected a/b test, sorting comments with "top" option returns a new
ctoken which will return "Replies" text for comments without replies.
to get rid of this behavior it'd be best to create known good ctokens
for every request.
This commit is contained in:
sherl
2025-10-19 23:32:12 +02:00
parent 668e8c32aa
commit 158dcc3b7f
1 changed files with 7 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
ythdd_extractor.py
View File

@@ -518,6 +518,13 @@ def WEBgetVideoComments(ctoken: str) -> tuple:
reply_ctoken = safeTraverse(reply_renderer, ["commentThreadRenderer", "replies", "commentRepliesRenderer", "contents", 0, "continuationItemRenderer", "continuationEndpoint", "continuationCommand", "token"], default="")
reply_count = safeTraverse(reply_renderer, ["commentThreadRenderer", "replies", "commentRepliesRenderer", "viewReplies", "buttonRenderer", "text", "runs", 0, "text"], default="0 replies").split(" ")[0]
# suspected a/b test. can be worked arount with on-demand ctoken creation.
# workaround for yt not showing replies when sorting for "top" comments
try:
int(reply_count) # can be just "Replies"
except:
reply_count = "0"
for comment in actual_comments:
found_key = safeTraverse(comment, ["entityKey"], default="unknown-key")
# try to link a relevant ctoken if a comment has response