Automatically download dictionary updates

[batterseapower]

A bit tricky because there is no stable URL for e.g. CFDICT

[Nick3C]

CEDICT has one though (and as it is doing the pinyin lookups it remains the most important to keep up to date :)

http://www.mdbg.net/chindict/export/cedict/cedict_1_0_ts_utf-8_mdbg.zip

[Nick3C]

Probably not too bad. We can just pull the url: http://www.chinaboard.de/chinesisch_deutsch.php?mode=dl and use regex to return the text matching: http://www.chinaboard.de/handedict/handedict-[8 numbers].zip

Then we have a URL for it. CFEDICT is similar I think.

Once we have the file we just:

• copy the current dict to [name].[date].bak (copy so that we don't get left with no dictionary if anki crashes)
• unzip to a temporary directory
• copy the new file into the directories dir, overwriting the old dictionary

[cburgmer]

You might want to have a look at class DictionaryDownloader from http://code.google.com/p/eclectus/source/browse/trunk/eclectusqt/update.py

You need to exchange the KDE io backend with the pure python ones though.

[Nick3C]

ah, that's fantastic.

[batterseapower]

I've added a small script to do the download to the repo, in pinyin/dictionaries/downloader.py.

It still needs to:

• Hook up to the user interface
• Be imbued with proper progress and error reporting
• Do checking of the available version vs. the last detected one, so we don't needlessly download an update

So quite a lot still to do.

[cburgmer]

Sorry, I forgot about this topic. I actually extracted the base functions mysqelf some weeks ago. I have put that into http://code.google.com/p/cjklib/source/browse/trunk/test/download.py.

Have you seen the table UpdateVersion that Eclectus writes? It might be helpful to make a separtate library out of all that.