i can't work out how to download the sounds for JDIC (jim breen dictionary) http://www.csse.monash.edu.au/~jwb/cgi- … dic.cgi?1C
... anyone know how to do it?

Took a bit of figuring out in the source of the page but, you can do this:

Code:

http://assets.languagepod101.com/dictionary/japanese/audiomp3.php?kana=どくしょ&kanji=読書

To get the audio =) Quite cool, you can put just about any word in.

file->save as ?

You can use firefox, just do file->save page. When I tried it saved as audiomp3.php.mpga (mpeg audio). Plays fine in mplayer.

bombpersons Wrote:Took a bit of figuring out in the source of the page but, you can do this:

Code:

http://assets.languagepod101.com/dictionary/japanese/audiomp3.php?kana=どくしょ&kanji=読書

To get the audio =) Quite cool, you can put just about any word in.

Nice find. This is going to be mad useful. Much easier than ripping the audio through Total Recorder.

How would I go about automating this for batches of audio from a list of words? I'm sure it's something really simple to input the words from a list into that format, but my brain's not working. ;p

Try this python script:

Code:

import urllib, codecs

URL = u"http://assets.languagepod101.com/dictionary/japanese/audiomp3.php?kana=%(reading)s&kanji=%(kanji)s"

class Word:
    def __init__(self, reading=u"", kanji=u""):
        self.reading = reading
        self.kanji = kanji

class Ripper:
    def __init__(self):
        self.words = []
        
    def load(self, filename): # Load a list from file
        file = codecs.open(filename, "rb", "utf-8")
        for line in file:
            word = line.split("\t")
            self.words.append(Word(word[0], word[1]))
        
        print("Loaded list at: ", filename)
    
    def download(self):
        for word in self.words:
            # Try and download this word
            webFile = urllib.urlopen((URL % {u'reading': word.reading, u'kanji': word.kanji}).encode("utf-8"))
            file = open(word.kanji + u".mpga", "wb")
            file.write(webFile.read())
            file.close()
            webFile.close()
            
            print "Downloaded " + word.kanji

if __name__ == "__main__":
    ripper = Ripper()
    ripper.load("words.txt")
    ripper.download()

Just put a file called words.txt with the reading and then the kanji with tab inbetween on each line.

Though, for some reason when I download the audio and I try to play it, the file is silent... Even when I use a browser...

It's probably me who's messed up some sort of codec, so you might be able to use this.

Thanks for the effort but I've never been able to get python stuff working, and my mind cringes at the thought of trying again. ;p

Here is a shell script (UNIX only?) for batch downloading. If a clip for an entry does not exist, it is logged in log.txt:

Code:

#!/bin/bash

missing=7e2c2f954ef6051373ba916f000168dc # Hash of "no clip exist" message

if [ $# -ne 0 ]
then
  echo "Usage: `basename $0` < input "
  echo "input: Field 1: かな, Field 2: 漢字"
  exit 65 # Bad arguments
fi

while read kana kanji
do
  wget -q "http://assets.languagepod101.com/dictionary/japanese/audiomp3.php?kana=$kana&kanji=$kanji" \
         -O "$kana-$kanji.mp3"

  clip_hash=`md5sum $kana-$kanji.mp3 | cut -d ' ' -f1 `

  if [ "$clip_hash" = "$missing" ]
  then
    echo "Clip for $kana $kanji did not exist" >> log.txt
    rm $kana-$kanji.mp3
  fi
done

Edit: Fixed a mistake with the logging.
Edit2: Fixed bug with renaming of audio file

Thanks, maybe cygwin or something will be easier to make work than python? ;p Geez, ought to be something for n00bs to easily use, though.

Maybe the following: format a list of words into two columns (?), then script runs and attaches link template around the first and second column, and then a download manager can add them as a batch or something? Or is that actually what the above posts already do, hehe.

nest0r Wrote:Maybe the following: format a list of words into two columns (?), then script runs and attaches link template around the first and second column, and then a download manager can add them as a batch or something? Or is that actually what the above posts already do, hehe.

Yes, my script is using wget to download and save the audio files to current directory.

Usage
1) Create a file with two columns: kana kanji (use space or tab between them)
2) Run ./script.sh < file.txt
3) Check log.txt for failed downloads

Pauline Wrote:Usage
1) Create a file with two columns: kana kanji (use space or tab between them)
2) Run ./script.sh < file.txt
3) Check log.txt for failed downloads

Does your script with work Cygwin, do you know? If so, that'd be quite handy.

brandon7s Wrote:

Pauline Wrote:Usage
1) Create a file with two columns: kana kanji (use space or tab between them)
2) Run ./script.sh < file.txt
3) Check log.txt for failed downloads

Does your script with work Cygwin, do you know? If so, that'd be quite handy.

Not completely, I think. Seems like md5sum is not included by default in Cygwin, but that only affects the detection of missing audio (uses hashes to identify them). The important program is wget.

Here is the script without removal of missing audio:

Code:

#!/bin/bash

if [ $# -ne 0 ]
then
  echo "Usage: `basename $0` < input"
  echo "input: Field 1: かな, Field 2: 漢字"
  exit 65 # Bad arguments
fi

while read kana kanji
do
  wget -q "http://assets.languagepod101.com/dictionary/japanese/audiomp3.php?kana=$kana&kanji=$kanji" \
         -O "$kana-$kanji.mp3"
done

IceCream Wrote:if there's anyone out there that can turn this into a really simple program, i will love you forever

Second that! This thread is incomprehensible to a person like me...

I'm sure there's a way to do it with regex--err, whatever it's called. Maybe a batch file... ? I guess for now I'll just make sure to keep my lists organized, do some clever replace/finding of that url template, and batch download with DownloadThemAll or something.

Well, this was quite the adventure, but I've finally gotten a lot more good soundfiles for my vocab cards.

Had to get tobberoth nofurigana plugin: http://groups.google.com/group/ankisrs/b...237dda4813
(By the way, don't know how it was supposed to work, but didn't seem to do anything, so I just configured it to overwrite the reading field with the new nofurigana reading, I don't think I was so supposed to do that. But did get it to do what I wanted to do.)

Got myself a nice export list with kanji's and of course readings without kanji's, so they could be used for the urls.

In Excel I added an concatenate function, to generate all the urls.
like this: =CONCATENATE("http://assets.languagepod101.com/dictionary/japanese/audiomp3.php?kana=",B1,"&kanji=",A1,"")

Copied all the urls to another list.

Imported that list into a downloader, pressed automatic download.
Renamed the whole batch to something generic and logic.
Put the sound file names in my excel file.
Configured the overwrite plugin to do overwrite things.

And now finally almost all of my vocab files have got soundfiles!!

So if anyone else was still planning to get some vocab files, it's quite manageable to do so. And even more so if you already have a clean-kana-only list with the readings, that was the biggest challenge in getting these files. But with the nofurigana plugin also a piece of cake.

Would really love this if it were an Anki plugin that could show up whilst making a card. Say you have a word 遺体 in the question field and could click a button to download and add the sound file would be so awesome.

mezbup Wrote:Would really love this if it were an Anki plugin that could show up whilst making a card. Say you have a word 遺体 in the question field and could click a button to download and add the sound file would be so awesome.

I second that!

If anyone using Firefox with it's awesome smart keyword search function (if not, take a look here) then there is a solution to download audio clips from WWWJDIC from your browser's addressbar with the help of the link provided by sugarlevi in one of the previous comments.

If you check out this link there is a looong javascript code. Change the '###INSERT YOUR KEYWORD SEARCH URL HERE###' to 'http://assets.languagepod101.com/dictionary/japanese/audiomp3.php?kana=%s&kanji=%s' and copy the whole thing onto your clipboard. Create a new bookmark which name could be anything. The location should be the loooong javascript thingy (which you've edited). The keyword also could be anything, but make it descriptive and short (I use 'ja' for the keyword), then save it.

After this you can download an audio file from JapanesePod101 by writing "yourkeyword reading expression" to your location/address/search/whateverbar:
e.g. "keyword がくせいせいかつ 学生生活"

After this something should popup in FF, like windows media player, or some other software which is associated with the filetype in firefox (not in your OS). For me, media player classic pops up, plays the audio (if it's available) and I can save the mp3 as I wish.

I hope this idea will come in handy for someone.

mygbmygb Wrote:

mezbup Wrote:Would really love this if it were an Anki plugin that could show up whilst making a card. Say you have a word 遺体 in the question field and could click a button to download and add the sound file would be so awesome.

I second that!

This would be awesome. Might try to do something like this myself if I find the time.

Psycho_Dad Wrote:If anyone using Firefox with it's awesome smart keyword search function (if not, take a look here) then there is a solution to download audio clips from WWWJDIC from your browser's addressbar with the help of the link provided by sugarlevi in one of the previous comments.

If you check out this link there is a looong javascript code. Change the '###INSERT YOUR KEYWORD SEARCH URL HERE###' to 'http://assets.languagepod101.com/dictionary/japanese/audiomp3.php?kana=%s&kanji=%s' and copy the whole thing onto your clipboard. Create a new bookmark which name could be anything. The location should be the loooong javascript thingy (which you've edited). The keyword also could be anything, but make it descriptive and short (I use 'ja' for the keyword), then save it.

After this you can download an audio file from JapanesePod101 by writing "yourkeyword reading expression" to your location/address/search/whateverbar:
e.g. "keyword がくせいせいかつ 学生生活"

After this something should popup in FF, like windows media player, or some other software which is associated with the filetype in firefox (not in your OS). For me, media player classic pops up, plays the audio (if it's available) and I can save the mp3 as I wish.

I hope this idea will come in handy for someone.

Thanks for this tip! I only tried this today, and it went swimmingly. Not something I'd want to do in bulk, but it worked for individual cards I was too lazy to add audio to whilst making them the first time.

Also, re: the above Firefox shortcut, be sure, if you're copy/pasting, that the encoding is Unicode. I kept getting an 'audio is not available' before I realized somehow my txt hadn't been converted. What clued in my n00b self was the percentage/letters format in the URL, rather than the usual kana/kanji that shows up after I hit type the keyword/text and hit Enter.

I knocked together a prog that will do the download for you.
You just need to feed it a text file in the format

kana [tab] Kanji

eg:

だんだん 段々
だんかい 段階
こうしょう 交渉
しょくじ 食事

it will save the audio named (Kanji) - (kana).mp3
eg:

旅館 - りょかん.mp3

Image here
http://img716.imageshack.us/img716/9800/...wnload.png

Download from here:

http://www.multiupload.com/UJDPIDRTSM

NOTE:
Please note that this performs no kind of validation, error checking, anything even resembling what you'd expect of a professional utility! But I realised I'd knocked together a similar one last week to download the files from a list of filenames in the Kore sheet and it was a simple thing to modify it to do this.
Also, it runs on Windows with the .NET framework.

Please don't do anything bad with this, and no fitness for purpose or stability is in any way implied!

This is awesome! Thanks. All hail mistamark!!! ;p

It worked great, except I had to hit 'continue' after some kind of index out of bounds error. Doesn't seem to actually hurt anything. Pretty sure my computer always smokes like that.

ruiner Wrote:This is awesome! Thanks. All hail mistamark!!! ;p

It worked great, except I had to hit 'continue' after some kind of index out of bounds error. Doesn't seem to actually hurt anything. Pretty sure my computer always smokes like that.

Glad that worked for you! Yeah, about the smoke, it happens on mine too!