Back

Yomichan (JP dictionary for Chrome/Firefox + EPWING + auto flashcards)

Congratulations!

I have to say that the lack of Sanseido compatibility limits Yomichan's value to me. I take the points you made on GitHub but it remains the case that Sanseido function of Rikaisama is much more useful than an Epwing J-J dictionary to me for a variety of reasons that I won't restate here,

Nevertheless I sincerely admire what you are doing and want to give you all the support and encouragement I can.

Thank you so much for your wonderful work. We all owe you a debt of gratitude.
Edited: 2017-05-17, 3:51 pm
Reply
New version 1.1.16 released; ruby tags are now ignored when scanning text Big Grin
Reply
(2017-05-20, 8:45 pm)FooSoft Wrote: New version 1.1.16 released; ruby tags are now ignored when scanning text Big Grin

Awesome thank you!!
Reply
Thanksgiving Sale: 30% OFF Basic, Premium & Premium PLUS Subscriptions! (Nov 13 - 22)
JapanesePod101
(2017-05-20, 9:28 pm)gaiaslastlaugh Wrote:
(2017-05-20, 8:45 pm)FooSoft Wrote: New version 1.1.16 released; ruby tags are now ignored when scanning text Big Grin

Awesome thank you!!

+1 Smile
Reply
1.1.17 released:
* Support for scanning text from pages with hidden elements (see https://github.com/FooSoft/yomichan/issues/23) Smile
* Fix form scanning which was broken in the last release Sad
Reply
(2017-05-23, 12:42 am)FooSoft Wrote: * Support for scanning text from pages with hidden elements (see https://github.com/FooSoft/yomichan/issues/23) Smile
Thanks!

I became more interested in Satori Reader since the time I mentioned it on the ticket (apparently I said it's not very important?) - very happy I can make flashcards from it now.
Edited: 2017-05-23, 1:10 pm
Reply
Thanks, FooSoft!

This enhances the experience a lot, but I've commented there (https://github.com/FooSoft/yomichan/issues/23) a little nice-to-have use case that was possible before, and I'm wondering if it'd be possible to keep it along this enhancement.

Anyway, your extension only keeps getting better and better ;-)
Reply
(2017-05-24, 1:43 am)faneca Wrote: Thanks, FooSoft!

This enhances the experience a lot, but I've commented there (https://github.com/FooSoft/yomichan/issues/23) a little nice-to-have use case that was possible before, and I'm wondering if it'd be possible to keep it along this enhancement.

Anyway, your extension only keeps getting better and better ;-)

See my comment:

https://github.com/FooSoft/yomichan/issu...-303635283

Basically you can't have your cake and eat it too  Smile
Reply
Would there be any way to remove the J-E entries in Daijirin and use it in Yomichan? I like the pitch accent but I'm avoiding English. Or is there a version of Daijirin without J-E?
Reply
You could hack Yomichan Import to get rid of English entries or post-process the generated JSON files if you want this. Anything related to EPWING parsing is a huge can of worms that I like to stay far far away from unless absolutely necessary.
Reply
I've only noticed this on Narou, so it might be a problem with that specific site, but something seems to be wrong with detecting sentence boundaries when making Anki cards with the {sentence} tag.

Here's an example where the effect is especially bad:  http://ncode.syosetu.com/n7975cr/491/
Try making an Anki card with the word 本日 right at the beginning. Instead of just "本日二話目" the created card gets this as the sentence:
"その2 本日二話目 D「ごきげんよう。やってまいりましたみんな大好きこのコーナー」 冥「前回の不評を全力でなかったことにしてますね」 D「み、ん、な、だ、い、す、き!」 冥「そんな必死にならなくていいですから……。早く進めましょう」 D「み、ん、な、だ「いい加減にしましょうね?」 D「コホン。それでは改めて、第二回教えてD先生のお時間です。今回は神々についてです」 冥「あー、来ちゃいましたか」 D「過去編"

So it grabs text before and after the sentence and then stops in the middle of a sentence (probably because it reached some kind of max length). This happens with every word in the line "本日二話目", but it doesn't happen with other words in the sentence it mistakenly put out, like the ごきげんよう or 前回


edit: Another example: http://ncode.syosetu.com/n7975cr/494/
Using any word in the first sentence "なぜそんなまどろっこしいことをする" works fine, but all words in the second sentence include everything before it (the first sentence, the title, the chapter number and the author name)
Edited: 2017-05-25, 6:41 am
Reply
This is happening because Yomichan is looking for actual sentences - expressions contained within quotes or bounded on both sides by unquoted punctuation marks. It looks like it also needs to look for line break characters to signal it to give up trying to find the start/end of the current sentence.

edit: fixed for next release
Edited: 2017-05-25, 11:19 am
Reply
1.1.18 released:
* can now set scan activate key to shift/alt/ctrl/none
* fixed incorrect extra deinflection of certain くverbs
* removed options that never make sense to turn off
* better sentence detection for card creation
* debug info for people who are into that sort of stuff
Reply
Hi, I previously reported about audio not found for kana-only words. That was fixed and I made a few kana-only cards since then.

Now, the audio for kana-only words plays in Chrome, but doesn't seem to be arriving in Anki. I'm trying to add ふわり at the moment. It's the same whether I use "JapanesePod101" or "JapanesePod101 (alternate)".
Reply
I think I see what is going on, will fix for next release.
Reply
Fixed in dev, lol...

https://github.com/FooSoft/yomichan/comm...aa1e77b3a2
Reply
Hey there! Would you be able to add support for Eijiro 英辞郎 EPWING dictionary? This is the dictionary that alc.co.jp uses. I would be forever grateful!!
Edited: 2017-06-26, 5:12 pm
Reply
(2017-01-09, 1:53 am)FooSoft Wrote: No words can describe my hate of EPWING and its formatting. One day once I have everything else working I'll probably make a pass to parse out the useful bits from all of the other garbage packed into every entry.

(sorry for the double post, rereading through this thread for all the updates I missed (holy shit how this add-on has improved!!! thank you!!!!) and and just had a thought)

What if you made an optional pop-out field for the user when they added a card? I don't know how possible this is, but I feel like working on a reg-expression parsing would cause a world of hell because:

1. You may wind up accidentally deleting information a user may actually want
2. The formatting isn't consistent within even the same dictionary
3. You would have to do this over and over for any additional dictionaries you plan to support

Meanwhile, a pop-up of some sort that displayed the definition information and allowed the user to add or delete things at will (including definitions they may like to see but not actually import) seems like it could be a flexible enough solution? Thoughts?

In any case, thank you again, FooSoft. This add-on is about to literally save my life  Big Grin

E D I T: One last thought after reading through the thread and messing around with this:

Has anyone figured out how to deal with words that have multiple readings? 泡沫 for example, can be ほうまつ or うたかた. Since I use {expression} as the first unique field, I can only add a card for one of these readings. Anyone know a work around?
Edited: 2017-06-26, 8:03 pm
Reply
New version, 1.2.0 released:

* Dictionaries can now be imported from ZIP files (download new version of yomichan-import to package EPWING dictionaries).
* Support for using glossaries in languages other than English.
* Fixed audio for creating Anki cards for terms that use no Kanji.

[Image: Screenshot_2017-06-26_20-21-49.png]

(2017-06-26, 5:07 pm)cerobi Wrote: Hey there! Would you be able to add support for Eijiro 英辞郎 EPWING dictionary? This is the dictionary that alc.co.jp uses. I would be forever grateful!!

I'll check out how the formatting on this dictionary looks this weekend. If it's not too bad, I can hook it up.
EDIT: It looks like I don't actually have this dictionary in my collection.

(2017-06-26, 7:25 pm)cerobi Wrote:
(2017-01-09, 1:53 am)FooSoft Wrote: No words can describe my hate of EPWING and its formatting. One day once I have everything else working I'll probably make a pass to parse out the useful bits from all of the other garbage packed into every entry.

(sorry for the double post, rereading through this thread for all the updates I missed (holy shit how this add-on has improved!!! thank you!!!!) and and just had a thought)

What if you made an optional pop-out field for the user when they added a card? I don't know how possible this is, but I feel like working on a reg-expression parsing would cause a world of hell because:

1. You may wind up accidentally deleting information a user may actually want
2. The formatting isn't consistent within even the same dictionary
3. You would have to do this over and over for any additional dictionaries you plan to support

Meanwhile, a pop-up of some sort that displayed the definition information and allowed the user to add or delete things at will (including definitions they may like to see but not actually import) seems like it could be a flexible enough solution? Thoughts?

In any case, thank you again, FooSoft. This add-on is about to literally save my life  Big Grin

E D I T: One last thought after reading through the thread and messing around with this:

Has anyone figured out how to deal with words that have multiple readings? 泡沫 for example, can be ほうまつ or うたかた. Since I use {expression} as the first unique field, I can only add a card for one of these readings. Anyone know a work around?

A pop-out to edit entries you are going to add would work fine for EPWING (I imagine it would be more like you click on the definition and are able to edit it), but the difficulty comes from all the nice HTML formatting I add to Anki cards. I don't see how I can expose a sane WYSIWYG editor in the definition window.

Regarding EPWING regex parsing, there is simply no way around it. At the very minimum I have to at least extract that Kanji expression and the reading for every definition. If you want deinflection support, I also have to extract information about the type of verb/adjective/whatever everything is. Oh, and all of this information is, of course, also stored in a completely bizarre and inconsistent manner.
Edited: 2017-06-26, 10:33 pm
Reply
(2017-06-26, 10:19 pm)FooSoft Wrote: A pop-out to edit entries you are going to add would work fine for EPWING (I imagine it would be more like you click on the definition and are able to edit it), but the difficulty comes from all the nice HTML formatting I add to Anki cards. I don't see how I can expose a sane WYSIWYG editor in the definition window.

Regarding EPWING regex parsing, there is simply no way around it. At the very minimum I have to at least extract that Kanji expression and the reading for every definition. If you want deinflection support, I also have to extract information about the type of verb/adjective/whatever everything is. Oh, and all of this information is, of course, also stored in a completely bizarre and inconsistent manner.

Ah, without a WYSIWYG, I guess I'm stuck with JMdict then. I just added 50 words last night to my vocab deck, and going back to edit each definition in anki was an absolute nightmare >.< Ah well. Thank you!
Reply
(2017-06-27, 8:50 pm)cerobi Wrote:
(2017-06-26, 10:19 pm)FooSoft Wrote: A pop-out to edit entries you are going to add would work fine for EPWING (I imagine it would be more like you click on the definition and are able to edit it), but the difficulty comes from all the nice HTML formatting I add to Anki cards. I don't see how I can expose a sane WYSIWYG editor in the definition window.

Regarding EPWING regex parsing, there is simply no way around it. At the very minimum I have to at least extract that Kanji expression and the reading for every definition. If you want deinflection support, I also have to extract information about the type of verb/adjective/whatever everything is. Oh, and all of this information is, of course, also stored in a completely bizarre and inconsistent manner.

Ah, without a WYSIWYG, I guess I'm stuck with JMdict then. I just added 50 words last night to my vocab deck, and going back to edit each definition in anki was an absolute nightmare >.< Ah well. Thank you!

I just wouldn't worry about formatting that much. You can never make it look perfect Wink
Reply
(2017-06-27, 10:00 pm)FooSoft Wrote: I just wouldn't worry about formatting that much. You can never make it look perfect Wink

Ah, ahaha, the issue is actually that I don't want all of the information that kenkyuusha gives. I just want the definitions, not all the example sentences and other usages and stuff. Alas...
Reply
Hi, I might like to use deinflect.json - what would be proper credits for that? It looks like it was derived from Rikaichan, but doesn't have any file header.
Reply
(2017-07-02, 7:50 am)HelenF Wrote: Hi, I might like to use deinflect.json - what would be proper credits for that? It looks like it was derived from Rikaichan, but doesn't have any file header.

There is no header for the simple reason that this is a JSON file, and JSON files do not support comments. The JSON itself was extracted from non-human-readable Rikaichan data back in the Yomichan for Anki extension days and since then has had various fixes and improvements applied to it. 

It is licensed under GPL3 and as long as you the Rikaichan and Yomichan projects as the source of it, you should be good.
Reply
(2017-07-02, 2:01 pm)FooSoft Wrote: The JSON itself was extracted from non-human-readable Rikaichan data back in the Yomichan for Anki extension days and since then has had various fixes and improvements applied to it. 

It is licensed under GPL3 and as long as you the Rikaichan and Yomichan projects as the source of it, you should be good.
Great, thanks.

Quote:JSON files do not support comments
... mind is blown somehow.
Reply