ikegami-yukino / jaconv Goto Github PK

Nice to get started.
Thank you for providing a great library.
In ver.0.3, when I pass "えっっ" to a function as follows, the function freezes.
jaconv.kana2alphabet("えっっ")

Hi!
Pretty lucky to get to use this convenient tool for handling Japanese script.
Keyword argument ignore comes handy in some circumstances where you don't want some of the characters to stay unchanged. I just found ignore was not used in jaconv.normalize. It'd be an easy fix though.
ありがとうございます〜

※日本語で申し訳ないですが、これを英語でというのも若干罰ゲーム感漂い、だったらスルーしちゃおうって流れになるところからのIssue作成なのでどうかご容赦を。

以下の部分の日本語が間違っていないでしょうか？

# ASCII以外の半角文字 to 全角文字
jaconv.h2z(u'abc', ascii=True)
# => u'ａｂｃ'

# 数字以外の半角文字 to 全角文字
jaconv.h2z(u'123', digit=True)
# => u'１２３'

なお、使う側からすると、混ざっている文字列に対して操作をするケースがあることを考えるとASCII以外→ASCIIのみという言い換えも良くないと思います。

というのは、kanaはデフォルトTrueで、いまのロジックだと以下のようになり、カナは常に変換されます。

>>> import jaconv
>>> jaconv.h2z(u'123abcﾃｨﾛ･ﾌｨﾅｰﾚ', digit=True)
'１２３abcティロ・フィナーレ'
>>> jaconv.h2z(u'123abcﾃｨﾛ･ﾌｨﾅｰﾚ', ascii=True)
'123ａｂｃティロ・フィナーレ'
>>>

これに関しては、英語のコメントを修正するのか、ロジックの修正なのか判断つきませんが、合わせてご判断頂ければと思います。

OS: Arch Linux
jaconv version: 0.3.4

• When installing jaconv as a system package, README.rst and CHANGES.rst get installed to /usr/README.rst and /usr/CHANGES.rst.
• When installing locally for the current user, the files get installed to $HOME/.local/README.rst and $HOME/.local/CHANGES.rst.

This happens with both python setup.py install and pip install.

The issue seems to be the data_files argument to setup in setup.py. This keyword is deprecated according to the documentation and removing it fixes the problem.

1. su is converted before tsu leading to words like 'atsui' being converted to 「あっすい」
2. Trailing oh is converted too soon leaving words like 'itoh' as 「いっおお」
3. Remaining oh replacements should be done after other replacements, current placement causes words like toho to become 「っおおお」
4. singular m remains unconverted and becomes 「っ」, e.g., namba => なっば
5. dzu is not converted, e.g. tsudsuku => っすっずく

>>> from jaconv import alphabet2kana as a2k
>>> a2k('tsudzuku')
'っすっずく'
>>> a2k('namba')
'なっば'
>>> a2k('itoh')
'いっおお'
>>> a2k('toho')
'っおおお'

As it says on the title, it would be nice to be able to automatically change ゑ to え, even を to お (as in をとこ). As well as be able to automatically transform けふ to きょう, くわ to か, げう to ぎょう etc.

I'm sure you know this very well, but there is also a convenient list here if it is helpful: https://www.bunka.go.jp/kokugo_nihongo/sisaku/joho/joho/kijun/naikaku/gendaikana/index.html

This is mostly useful for OCR purposes, but being able to change っ to つ, ぃ to い, etc. would be helpful when standardizing texts for search.

I found a bug in your code when I tried to convert a sentence that ends with 'っ' . (example : 'あっ' , 'ぐっ')
IMO, Cause is clear, list variable 'text' must be cast to str

In [2]: jaconv.kana2alphabet('あっ')

---

TypeError Traceback (most recent call last)
in ()
----> 1 jaconv.kana2alphabet('あっ')

C:\ProgramData\Anaconda3\lib\site-packages\jaconv\jaconv.py in kana2alphabet(text)
211 tsu_pos = text.index('っ')
212 if len(text) <= tsu_pos + 1:
--> 213 return text[:-1] + 'xtsu'
214 text[tsu_pos] = text[tsu_pos + 1]
215 text = ''.join(text)

TypeError: can only concatenate list (not "str") to list

Python3.4以下もサポートされているようですが、typingモジュールはPython3.5で追加されたようです。
https://docs.python.org/ja/3/library/typing.html

Python3.4の環境で検証ができなかったのですが、とりあえず型ヒントのある記法はPython2.7ではinvalid syntaxになるようです…（Wandboxで確認）

enlargesmallkana()にある型ヒントは<=3.4の環境で邪魔にならないのでしょうか？

いまいち確証が持てていなくて申し訳ないです。よろしくお願いします。

Hi, i want to know that is there any api to convert romaji or alpha to kanji?

It would be nice for standardizing text inputs to be able to switch a string from shin to kyuu (or vice-versa).

Could you please tag the source again? This allows distributions to get the complete source from GitHub if they want.

This was done in the past but not for 0.3.3.

Thanks

The git tag for version v0.4.0 is missing. Could it be added please?

See also #27.