fix: handle curly quotes in Twitter Jina fallback detection
Jina Reader returns smart/curly quotes (U+2019) instead of ASCII apostrophes. Use apostrophe-free substrings for detection.
This commit is contained in:
@@ -73,10 +73,10 @@ class TwitterChannel(Channel):
|
|||||||
|
|
||||||
# Detect unusable Jina responses for X/Twitter (JS-required pages)
|
# Detect unusable Jina responses for X/Twitter (JS-required pages)
|
||||||
unusable_indicators = [
|
unusable_indicators = [
|
||||||
"this page doesn't exist",
|
"page doesn", # "this page doesn't exist" (handles both ' and ')
|
||||||
"Don't miss what's happening",
|
"miss what", # "Don't miss what's happening"
|
||||||
"Something went wrong. Try reloading",
|
"Something went wrong. Try reloading",
|
||||||
"Log in</a>",
|
"Log in](", # Markdown link: [Log in](...)
|
||||||
]
|
]
|
||||||
if any(indicator in text for indicator in unusable_indicators):
|
if any(indicator in text for indicator in unusable_indicators):
|
||||||
return ReadResult(
|
return ReadResult(
|
||||||
|
|||||||
Reference in New Issue
Block a user