Closed NekoB0x closed 3 months ago
That is true and I've also already noticed that it should be fixed. The first 80% of the name have to mostly match the name on anidb for the plugin to match it. Then the levenshtein distance is used to find the closest one of the matches.
With short names cutting of only 20% prevents matching in case there is additional data at the end (that's not on anidb) like the year.
In version 5 and earlier it was 50%. However there were much more wrong matches and the year would just be ignored.
The year could be removed from the name before trying to find matching names. Would that be okay? It could in the next step be used for finding the best match. But then the question is what to prioritize.
The year could be removed from the name before trying to find matching names. Would that be okay? It could in the next step be used for finding the best match. But then the question is what to prioritize.
Yes, strip the year in the query, then match it with priority to the original "unstripped" name, because there are titles with seasons that are just "Sametitle (year)".
https://anidb.net/anime/16804 https://anidb.net/anime/15484 https://anidb.net/anime/15743 https://anidb.net/anime/17072 https://anidb.net/anime/15428
When following the common naming scheme (Series (year)) the titles aren't searched.
Example: Canaan (2009) - returns empty result Canaan - works
The last version that didn't have this issue was 5.0.0.0.