add support for Japanese (JA)

[siikamiika]

Changes proposed in this pull request:

• Add support for JA

Status

• READY
• HOLD
• WIP (Work-In-Progress)

How to verify this change

import num2words
num2words.num2words(123456, lang='ja')
num2words.num2words(123456, lang='ja', reading=True)
num2words.num2words(-5, lang='ja')
num2words.num2words(0, lang='ja', reading=True)
num2words.num2words(74, lang='ja', reading=True)
num2words.num2words(74, lang='ja', reading=True, prefer=['し', 'しち'])
num2words.num2words(123456, lang='ja', reading=True, to="ordinal")
num2words.num2words(123456, lang='ja', reading=True, to="ordinal_num")

Additional notes

[coveralls]

Coverage increased (+0.2%) to 89.805% when pulling d96b828 on siikamiika:japanese-language-support into b492530 on savoirfairelinux:master.

[erozqba]

Hi @siikamiika
Thanks a lot for taking the time to make this PR! It's really appreciated.
I don't know any Japanese so we really trust the contributors that make the PR.
I can see this is a work in progress, as a suggestion, don't forget to update the README.rst and to implement as many automatic tests as you can.

Thanks a lot again!

[ventilooo]

hey, @erozqba ! I speak a bit of Japanese.
I'll try to handle the review as much as I can 😉

@siikamiika ping me once the WIP is over !
Thanks again for your contribution 👍

[siikamiika]

Seems like the code is having some issues with Python 2. Probably Unicode related. The tests should be fine for Python 3 now, but I haven't updated README yet.

[siikamiika]

@erozqba @ventilooo Everything looks fine to me now. This sure was an interesting language to implement. 😅 I'm also just learning Japanese so I did this as an exercise. I did some research on the preferred readings of some number words in Japanese, but there could be errors somewhere.

[erozqba]

Thanks a lot, @siikamiika for this contribution!
@ventilooo thanks in advance for your code review.

[erozqba]

Why there are some commented lines as this one in this list?

[siikamiika]

It's because there have been multiple emperors/empresses during the same year and one has to be chosen since num2words doesn't distinguish between months and days. The last one is the one who has reigned until the next emperor/empress. I actually had them the wrong way around first (used the first one) but fixed them in a commit

…

On Thu, Jun 14, 2018, 19:01 Ernesto Rodriguez Ortiz < ***@***.***> wrote: ***@***.**** commented on this pull request. Thanks a lot, siikamiika for this contribution! ------------------------------ In num2words/lang_JA.py <#171 (comment)> : > + +# Source: https://www.sljfaq.org/afaq/era-list.html +# if there are multiple eras for the same year, use the last one +ERA_START = [ + (645, ("大化", "たいか")), + (650, ("白雉", "はくち")), + (686, ("朱鳥", "しゅちょう")), + (701, ("大宝", "たいほう")), + (704, ("慶雲", "けいうん")), + (708, ("和銅", "わどう")), + (715, ("霊亀", "れいき")), + (717, ("養老", "ようろう")), + (724, ("神亀", "じんき")), + (729, ("天平", "てんぴょう")), + # (749, ("天平感宝", "てんぴょうかんぽう")), + (749, ("天平勝宝", "てんぴょうしょうほう")), Why there are some commented lines as this one in this list? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#171 (review)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFV8s1gsB792hxY4yW9hFsp7rXB6j45_ks5t8oi8gaJpZM4UeNjH> .

[ventilooo]

@siikamiika Instead of putting some emperor/empress in comment, could you solve this with a method ?

Like with ⬇️

self.assertEqual(n2j(0, prefer=["〇"]), "〇")

In order to easily give the choice to the user.

Also I don't know how useful this could be, but having the option to choose the the output format between

• Romaji
• Kanji
• Hiragana

If I want to write an app for kid, they don't necessarily know all the kanji associated with numbers.

Or if I want to write an app for foreigner, to learn Japanese. Having an romaji output would help.

[siikamiika]

@ventilooo

Instead of putting some emperor/empress in comment, could you solve this with a method ?

Like with ⬇️

self.assertEqual(n2j(0, prefer=["〇"]), "〇")

In order to easily give the choice to the user.

Sounds like a good idea, I'll try to do that.

Also I don't know how useful this could be, but having the option to choose the the output format between

• Romaji
• Kanji
• Hiragana

If I want to write an app for kid, they don't necessarily know all the kanji associated with numbers.
Or if I want to write an app for foreigner, to learn Japanese. Having an romaji output would help.

Hiragana output is currently supported with the reading flag, but I'm not sure if that's the best way to choose it if you want to choose output between multiple different writing systems. There's also katakana, which would be pretty easy to implement if I remember the Unicode code points correctly.

Romaji is a bit complicated, though. First, there are multiple ways to romanize Japanese text, as seen from your link. An optional dependency to something like python-romkan could be added, but even then there would be the problem of separating the number words with spaces only for romanized Japanese but not with kana. Anyway, I agree that there could be some use for that feature. I guess I could look into implementing romaji at some point.

[siikamiika]

I just realized that the era thing wasn't as simple as having some ruler for less than a year. The prefer suddenly makes a lot more sense but I might replace the year in the tuples with another tuple containing both the start and the end of the era.

Some reading:
https://en.wikipedia.org/wiki/Eiwa
https://en.wikipedia.org/wiki/Tenju

[ventilooo]

That's perfect !
Once again @siikamiika , thanks for your contribution !

If you want to discuss further on future improvement for the Japanese support.

Just open an issue with the RFC tag 😉 🤞