Back

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

Yomichan 1.1.12 has been released, fixing various cloze related bugs.
Reply
(2017-04-08, 8:10 pm)FooSoft Wrote: Yomichan 1.1.12 has been released, fixing various cloze related bugs.

How can I update my chrome Yomichan?
Reply
The update happens automatically.
Reply
JapanesePod101
Yomichan-Import now has a GUI for Windows, Linux, and Mac OS X

[Image: import.png]
Edited: 2017-04-14, 11:06 pm
Reply
May I have a request?

I converted ゼロの使い魔 light novel into HTML using the cb's JNovelFormatter app.

Currently, I am using old version of Firefox with Rikaisama to help me with reading it, but I would like to switch to Chrome with Yomichan.

Yomichan almost work - it's HTML, after all, with one exception. In the text, there is some furigana generated in ruby elements and some if it does not work. It seems that if the entire word is in ruby element, it works, but if word is splitted, it does not.

Here is the example with the word 覗き込む.

Code:
<ruby>
 <rb>覗</rb>
 <rp>《</rp>
 <rt>のぞき</rt>
 <rp>》</rp>
</ruby>
き込んでいる女の子が言った

Here is the snippet I prepared for testing purposes.
Reply
(2017-04-14, 11:06 pm)FooSoft Wrote: Yomichan-Import now has a GUI for Windows, Linux, and Mac OS X

[Image: import.png]

Yay!! Thank you for this.
Reply
(2017-04-15, 8:14 am)Robik Wrote: May I have a request?

I converted ゼロの使い魔 light novel into HTML using the cb's JNovelFormatter app.

Currently, I am using old version of Firefox with Rikaisama to help me with reading it, but I would like to switch to Chrome with Yomichan.

Yomichan almost work - it's HTML, after all, with one exception. In the text, there is some furigana generated in ruby elements and some if it does not work. It seems that if the entire word is in ruby element, it works, but if word is splitted, it does not.

Here is the example with the word 覗き込む.

Code:
<ruby>
 <rb>覗</rb>
 <rp>《</rp>
 <rt>のぞき</rt>
 <rp>》</rp>
</ruby>
き込んでいる女の子が言った

Here is the snippet I prepared for testing purposes.

You can use JNovelFormatter's option to strip Ruby - it's not really necessary if you're using Yomichan anyway.
Reply
(2017-04-15, 8:14 am)Robik Wrote: May I have a request?

I converted ゼロの使い魔 light novel into HTML using the cb's JNovelFormatter app.

Currently, I am using old version of Firefox with Rikaisama to help me with reading it, but I would like to switch to Chrome with Yomichan.

Yomichan almost work - it's HTML, after all, with one exception. In the text, there is some furigana generated in ruby elements and some if it does not work. It seems that if the entire word is in ruby element, it works, but if word is splitted, it does not.

Here is the example with the word 覗き込む.

Code:
<ruby>
 <rb>覗</rb>
 <rp>《</rp>
 <rt>のぞき</rt>
 <rp>》</rp>
</ruby>
き込んでいる女の子が言った

Here is the snippet I prepared for testing purposes.


This is something I’m planning to do, but it is not trivial. The problem is that Yomichan just reads text as if there were no tags present. So when you mouse-over “覗” you actually get “覗《のぞき》”, which obviously will not get you meaningful results. Maybe I’ll do it over a rainy weekend (lots of those here in Seattle).
Reply
(2017-04-15, 6:04 pm)gaiaslastlaugh Wrote:
(2017-04-15, 8:14 am)Robik Wrote: ...

You can use JNovelFormatter's option to strip Ruby - it's not really necessary if you're using Yomichan anyway.

Yeah, I would do that, main problem are names. Even thought in this particular novel, 99% will be in katakana anyway...
Reply
Anki connect stopped working today. Getting `Error: TypeError: Cannot read property 'sort' of null` in the extensions options error section.

EDIT: Oops, Issue was having two instances of anki open?...
Edited: 2017-04-17, 10:50 pm
Reply
Yeah I guess that's just a thing that doesn't work, will take a later at what actually causes that error message. I'm surprised you were able to launch two instances of Anki without AnkiConnect complaining about not being able to bind to the network port. You are probably running an old version that fails silently.
Reply
Is there a way to export to a delineated text file (like Rikaisama, which is now officially unsupported)? Since I've already got everything set up to import that way, I'd prefer to continue doing so instead of messing with direct Anki integration.

If not, that's fine too, but I prefer being lazy and not running Anki when it's unnecessary.
Reply
(2017-04-18, 8:07 pm)sholum Wrote: Is there a way to export to a delineated text file (like Rikaisama, which is now officially unsupported)? Since I've already got everything set up to import that way, I'd prefer to continue doing so instead of messing with direct Anki integration.

If not, that's fine too, but I prefer being lazy and not running Anki when it's unnecessary.

Nope, there is no way to output data to files in the WebExtensions plugin model that is used by Chrome (and future versions of Firefox). The idea of Google Docs integration was brought up before but I haven't had the time to implement it.

You're better off using the Anki integration anyway since it helps you keep track of what definitions you have already added to your deck.
Reply
(2017-04-18, 11:43 pm)FooSoft Wrote:
(2017-04-18, 8:07 pm)sholum Wrote: Is there a way to export to a delineated text file (like Rikaisama, which is now officially unsupported)? Since I've already got everything set up to import that way, I'd prefer to continue doing so instead of messing with direct Anki integration.

If not, that's fine too, but I prefer being lazy and not running Anki when it's unnecessary.

Nope, there is no way to output data to files in the WebExtensions plugin model that is used by Chrome (and future versions of Firefox). The idea of Google Docs integration was brought up before but I haven't had the time to implement it.

You're better off using the Anki integration anyway since it helps you keep track of what definitions you have already added to your deck.
You were initially open for CSV files, weren't you?
Reply
(2017-04-19, 4:11 am)Matthias Wrote:
(2017-04-18, 11:43 pm)FooSoft Wrote:
(2017-04-18, 8:07 pm)sholum Wrote: Is there a way to export to a delineated text file (like Rikaisama, which is now officially unsupported)? Since I've already got everything set up to import that way, I'd prefer to continue doing so instead of messing with direct Anki integration.

If not, that's fine too, but I prefer being lazy and not running Anki when it's unnecessary.

Nope, there is no way to output data to files in the WebExtensions plugin model that is used by Chrome (and future versions of Firefox). The idea of Google Docs integration was brought up before but I haven't had the time to implement it.

You're better off using the Anki integration anyway since it helps you keep track of what definitions you have already added to your deck.
You were initially open for CSV files, weren't you?

I'm open to the idea if somebody goes and implements it in a nice and clean way Smile

Since extensions have no file system access there would need to be a separate process running to create the CSV file, or it would need to hook up with Google Docs
Edited: 2017-04-19, 1:40 pm
Reply
I don't know why they (Chrome and Firefox teams) had to go and break everything... been having tons of trouble with things recently because of it.

If a separate process would be required to output to a file, I guess I might as well use Anki integration.
I do most of my reading (and thus my card adding) on my S3, which is why I was wondering about exporting to a file (I had to take the file from my phone to my computer and import it from the desktop version of Anki).

Is it possible to send card data from Yomichan to AnkiDroid in any way? I don't plan on updating Firefox on my phone, but just in case, it'd be nice to have a backup plan (and taking card data from my phone to my computer is a pain anyway).

Thanks for answering questions so quickly!
Reply
Chrome is actually pretty good about maintaining compatibility, it's just Firefox that is breaking everything to copy Chrome's extension model. Honestly, I have no idea how Firefox extension developers put up with all the breakage and several-month waits to get their extensions submissions / updates approved.

AnkiDroid is a completely separate application from Anki and it has no plugin system to enable you to use something like AnkiConnect to send flashcard data across. You can however just run Anki on a different computer than what you are using and send data to it that way if it makes your workflow easier...
Reply
The Typhon reading app claims it can create flashcards in AnkiDroid (though I've not looked into this at all).

... of course, that's an app, so it does have file system access.
Edited: 2017-04-19, 3:45 pm
Reply
Tried the import feature today for the first time. Took me a while to figure out that it didn't like Japanese folder names at all (I'm using Windows 10 with English as my display language, but my dictionary was in a folder named 三省堂大辞林) On the gui it just gave me exit code 1 as output. I thought I'd try it from console and when I tried to specify the input path and the kanji came up as gibberish (with tabbing) I kind of figured that might be the problem. So after renaming the folders I tried again from the gui and that was seemingly more successful. I just thought I might mention it in case somebody else runs into the problem as well.

So for the gui import.. when "2017/04/20 22:39:43 starting dictionary server on port 9876..." showed up nothing happened for a while... and I figured I'd just go into the addon in firefox because I remembered there was the option to import "Local Dictionary" (and that had the same default port shown) so I started that as well and it seems like everything worked and the dictionary got imported just fine. But the last output in the gui is still "2017/04/20 22:39:43 starting dictionary server on port 9876..." So uh, I'm just wondering, is it supposed to be like that or did I do something wrong?
Reply
(2017-04-20, 3:59 pm)sumsum Wrote: Tried the import feature today for the first time. Took me a while to figure out that it didn't like Japanese folder names at all (I'm using Windows 10 with English as my display language, but my dictionary was in a folder named 三省堂大辞林)  On the gui it just gave me exit code 1 as output. I thought I'd try it from console and when I tried to specify the input path and the kanji came up as gibberish (with tabbing) I kind of figured that might be the problem. So after renaming the folders I tried again from the gui and that was seemingly more successful. I just thought I might mention it in case somebody else runs into the problem as well.

So for the gui import.. when "2017/04/20 22:39:43 starting dictionary server on port 9876..." showed up nothing happened for a while... and I figured I'd just go into the addon in firefox because I remembered there was the option to import "Local Dictionary" (and that had the same default port shown) so I started that as well and it seems like everything worked and the dictionary got imported just fine. But the last output in the gui is still "2017/04/20 22:39:43 starting dictionary server on port 9876..." So uh, I'm just wondering, is it supposed to be like that or did I do something wrong?

Yomichan-Import uses Zero-EPWING for EPWING data extraction, which in turn uses libeb. It is possible that some step in the process does not properly support Unicode characters, I will have to check it out. What OS were you running by the way?

I can see how the last message can be a little confusing. The problem is that the API I am using does not provide a success notification for when the web server starts, but it does tell me if it had failed. So when I output "starting dictionary server", it really means "I'm about to try to start a sever, if something fails I will tell you about it". Maybe different wording would help.
Reply
(2017-04-20, 4:20 pm)FooSoft Wrote:
(2017-04-20, 3:59 pm)sumsum Wrote: I'm using Windows 10 with English as my display language

 What OS were you running by the way?

As mentioned: Windows 10, more specifically the Education edition (not sure if that matters)  Smile Display and OS language should be English

Quote:I can see how the last message can be a little confusing. The problem is that the API I am using does not provide a success notification for when the web server starts, but it does tell me if it had failed. So when I output "starting dictionary server", it really means "I'm about to try to start a sever, if something fails I will tell you about it". Maybe different wording would help.

I guess I kind of thought the tool would walk me through the whole import process (holding my hand till the end)! So when nothing happened and no explanation aka "and now start the import from within the addon as well!" was shown, I was a bit confused. I figured it out, but that the gui didn't give any feedback at all was a bit unexpected. Even after I imported the dictionary I was left wondering if I could just close the gui or if I could start importing another dictionary with the same one (didn't seem like it as the "Import Dictionary" button remained greyed out) I guess it would probably be hard to get some feedback back to the gui from the server, but it would be nice if it was at least documented somewhere that this is how it works Smile  Well, I got it to work (and it didn't take me so long) so it's not like it's a super big deal even for people like me... (in Germany we sometimes use the term DAU as in "dumbest assumable user"/"dümmster anzunehmender User" *cough*, not sure if that's a thing in English as well)
Edited: 2017-04-20, 4:51 pm
Reply
(2017-04-20, 4:43 pm)sumsum Wrote: in Germany we sometimes use  the term DAU as in "dumbest assumable user"/"dümmster anzunehmender User" *cough*, not sure if that's a thing in English as well)

"Lowest common denominator" and "Joe Sixpack" are very roughly equivalent. Although lowest common denominator doesn't necessarily connote dumbness of the user - rather blandness of the product as it must appeal to the most number of people and offend the least.
Reply
(2017-04-19, 1:40 pm)FooSoft Wrote: I'm open to the idea if somebody goes and implements it in a nice and clean way  :-)
I am too stupid to even use ANKI, so I can't help on that. :-(

It is quite embarrassing but I'll ask anyway: What about bringing the dataset to the clipboard. From there it is just 2 finger moves away from whatever file you want.
Reply
In another thread ( http://forum.koohii.com/thread-14508.html ) someone is just switching over to Yomichan and for the sanseido dictionary sometimes the 和英 entries pop up as well. Would there be some way to disable that? (Or is the only solution to use other dictionaries if you don't want the 和英 part?)

Posting this here in case FooSoft might have missed it.
Edited: 2017-04-21, 5:24 am
Reply
I'd like to invite all people (with github accounts) who support the idea of ignoring entirely alphanumeric (english / number) entries in yomichan-import to go add +1s to the opened issue here: https://github.com/FooSoft/yomichan/issues/47

Does anyone else find it annoying when the yomichan box pops up when you're typing in a text box and hit Shift while your cursor happens to be hovering over some English word? I sure do (then again I don't know how many people other than me actually using the EPWING support in yomichan). I don't really want a popup box to appear when my mouse hovers over the word "great" because Daijisen has an entry for it for instance, or even a random letter.

I'd just go implement it myself as it's not particularly difficult but it would require changing method signatures in a couple of places, which depending on preference could also change method signatures on non-EPWING extractors where ignoring alphanumeric entries is probably not relevant. Since I dunno what the stance is on changing the existing method signatures I'd rather leave that to FooSoft (and cough cough he's not on the top secret irc channel anymore so I can't just ask him quickly there wink wink).

Also I suppose you could also implement this directly in the yomichan extension itself but to me it seems cleaner to do this on the yomichan-import side. That's my preference though.

EDIT: Replaced the old link with a link to the newly opened issue on the yomichan github project issue tracker.
EDIT 2: The functionality is going to be in the next version!
Edited: 2017-04-22, 3:04 pm
Reply