I wrote a little program that compiles flashcards from Japanese (the iPhone dictionary) vocabulary lists. I've been using goo to get the definition and pitch accent, but they've gone and changed the website for the third time this year and the pitch accent seems to have disappeared completely.

I'd been getting annoyed with goo, and now it's just a liability. But what are the alternatives? I'd really rather an offline source, but I imagine that would be difficult to work with.

I have the slight issue of typing being a major problem. I wrote the program to avoid as much manual work as possible when making flashcards. I'm also no programming expert, so I could well have hit a wall here...

Does anyone else automate their flashcard building process? Any ideas?

Depending on how much time you're willing to invest, you could download the Daijirin EPWING and parse it for the information, it's offline and will obviously never change. Otherwise, I'd say dic.yahoo.co.jp is your best bet since it has both daijisen and daijirin.

I don't mind buying a dictionary as long as it has the pitch accent code. I didn't think it was as easy to extract the data from offline sources though.

Javizy Wrote:I don't mind buying a dictionary as long as it has the pitch accent code. I didn't think it was as easy to extract the data from offline sources though.

It might not be, you have to learn the internals of an EPWING data file, I guess those should be available as documentation online though, some guy on this forum created an EPWING viewer, he might be able to help you out.

EPWING isn't trivial to work with. It's an archaic file format- it's not practical to parse yourself. You have to use it via libeb, or some wrapper utility that uses libeb.

I've created a Python library that makes it much easier, but you still need to understand Python fairly well to use it.

My epwing library is located here: http://github.com/aehlke/epywing
You also need my updated EBmodule, which is a Python wrapper for libeb: http://github.com/aehlke/ebmodule

If you're on Windows, you're going to have a much harder time with this though.

One more option that's probably more reasonable is to use "lookup" via its command-line interface. This is what the anki epwing plugin does. It's not as fine-grained control over EPWING, but it's probably sufficient. You can see that anki plugin here: http://github.com/dae/ankiplugins/blob/master/epwing.py