Closed treyhunner closed 1 year ago
This may be related to #214
Looks like the slugging algorithms for the website and the search indexer. Both are to some degree using the title stored within the video JSON files but the unicode handling is completely different. I'll check if I can find a different approach here where the unicode-handling is closer to the one for the static pages
I'm now deploying a fix for the first issue. Need to rebuild the index so there's going to be a short downtime.
@zerok the site generator is using unidecode to handle this translation. Would it be possible to use a similar go library to avoid the need to make these sorts of one-off substitutions? I have not used https://github.com/mozillazg/go-unidecode, but it claims to be inspired by the Python library, so I imagine that it would be at least largely compatible in its behavior.
Yes, that would definitely an option. I'd suggest we wait with that change, though, until another concrete issue appears.
You can find a bunch of similar issues by searching e.g. "PyCon JP" or "Moscow Python Conf", where much of the talk titles are being removed from the search result URL completely (vs. the links that are displayed from the event pages for these conferences).
Ah, good to know :) Then I will collect a handful of testcases and proceed from there :)
Thanks!
JP and Moscow links look better now. Slugging still doesn't work in exactly the same way but I think we're getting closer. If you stumble upon a broken link in the index, please create a ticket for that and I will tune the function some more :)
The first result here links to a URL which doesn't resolve: https://pyvideo.org/search.html?q=unicode+python+2+better+python+3