Fixes: #219 typo in Japanese post rules

[massimolipari]

• What kind of change does this PR introduce? (Language addition, bug fix, feature, docs update, ...)

Bug fix, fixes: #219. Changed u to ɯ in one rule in both data/post/jpn-Hira.txt and data/post/jpn-Kana.txt.

• What is the current behavior? (You can also link to an open issue here)

See #219: words like とうきょう/トウキョウ are incorrectly transcribed as /toukʲou/ rather than /toːkʲoː/.

• What is the new behavior (if this is a feature change)?

See above.

• Does this PR introduce a breaking change? (What changes might users need to make in their application due to this PR?)

This will have a substantial effect on output transcriptions.

[juice500ml]

Hi, thanks a lot for your efforts! If it's possible, can you add the aforementioned Tokyo example as a test case? Thanks!

[massimolipari]

It doesn't seem like there's a test file for Japanese yet, is there any documentation on how to set one up? (Especially since there's three different scripts)

[juice500ml]

Thanks!! There's no documentation yet, but a good example will be https://github.com/dmort27/epitran/blob/master/epitran/test/test_slovene.py or https://github.com/dmort27/epitran/blob/master/epitran/test/test_finnish.py

The test file's name should always be "test_something.py", in this case, "test_japanese.py" at that folder.

[juice500ml]

FYI, you pushed bunch of Turkish in this PR!

[massimolipari]

Whoops, sorry, that commit was meant for my fork!

[massimolipari]

I've tried adding a test file, but I'm not sure if I've done it right. I couldn't tell from the example files how to go about adding tests for multiple scripts (should they be in separate files?), so I've only done the Hiragana case for now.

[juice500ml]

You can do something similar like: https://github.com/dmort27/epitran/blob/master/epitran/test/test_slovene.py#L25-L39
It doesn't have to be that many, but there should be adequate number of sample cases that can check the sanity.

[dmort27]

Actually, these look good.