Consider this a mini-PSA to people working on JMDICT-based dictionaries. Yomichan's dev is MIA and rikaichamp doesn't support audio (I think?), so I'll post this here.
Somehow, a couple of the jpod101 audio files are stored wrong on their servers, with names that have things like (ik) or (oK) in them. I noticed this when I was bouncing all of the /EntL.....X/ words in edict2 against their asset server (because I wanted a list instead of brute forcing it whenever the user tries to play audio) but screwed up and forgot to remove the metadata; some of them actually went through with metadata. I decided to try every combination of with-and-without metadata for every /...X/ word. Here's a list:
Left of the , is the kana= kanji= pair that'll find the audio file, on the right is either the reading;kanji pair or just the reading (if there's no kanji).
Consider this a mini-PSA to people working on JMDICT-based dictionaries. Yomichan's dev is MIA and rikaichamp doesn't support audio (I think?), so I'll post this here.
Somehow, a couple of the jpod101 audio files are stored wrong on their servers, with names that have things like (ik) or (oK) in them. I noticed this when I was bouncing all of the /EntL.....X/ words in edict2 against their asset server (because I wanted a list instead of brute forcing it whenever the user tries to play audio) but screwed up and forgot to remove the metadata; some of them actually went through with metadata. I decided to try every combination of with-and-without metadata for every /...X/ word. Here's a list:
Left of the , is the kana= kanji= pair that'll find the audio file, on the right is either the reading;kanji pair or just the reading (if there's no kanji).