opendatakosovo / cyrillic-transliteration Goto Github PK

input a .txt and execute it through a command "CyrTranslit -i text.txt"

As you know, it is rather common in Slavic texts to use accent marks to disambiguate between homographs, or to indicate intonation or vowel pronunciation.

https://en.wikipedia.org/wiki/Ye_with_grave and https://en.wikipedia.org/wiki/I_with_grave_(Cyrillic) are rather common in Macedonian.

When I try to transliterate any word written in Latin, which contain letter q, q letter kept as is.

Example:

In [1]: import cyrtranslit

In [2]: cyrtranslit.to_cyrillic('Question', lang_code='ru')
Out[2]: 'Qуестион'

Letter q doesn't exist in Russian alphabet, and in most cases q should be replaced with к.

I received different outputs for the same inputs based on code version:

Version: 1.0

>>> import cyrtranslit
>>> cyrtranslit.to_latin("Я часто пью водку", "ru")
"JA chasto p'ju vodku"

Version: 1.1

>>> import cyrtranslit
>>> cyrtranslit.to_latin("Я часто пью водку", "ru")
"YA chasto p'yu vodku"

Sorry, but you should know that "Republika Kosovo" doesn't exist. And most of all THERE IS NO "Serbian phrase 'Republika kosovo'". Please, don't spread the confusion.

Couldn't you also make this concession?

string_to_transliterate = string_to_transliterate.decode('utf-8') does not work in Python 3, returns the error

AttributeError: 'str' object has no attribute 'decode'

Similarly, return latinized_str.encode('utf-8') means a byte-string is returned in Python 3.

b'my latin string'

Removing the encode and decode of strings in init.py makes the package work with python 3

Related to issue #6

installed cyrtranslit-1.1.1

echo 'Її' | cyrtranslit -l UA
output:
Ïï
(i.e. not changed)

Thank to develop and share this great project. I wonder what is the reference or historical background of the Cyrillic-Latin mappings in CyrTranslit. Is it based on the Soviet project? Or something similar exsiting transliteration scheme?

Mongolian language does not support casing variations for 2 letters in latin, e.g.:

>>> import cyrtranslit
>>> cyrtranslit.to_cyrillic("Kh", "mn")
'Х'
>>> cyrtranslit.to_cyrillic("KH", "mn")
'КH'
>>> cyrtranslit.to_cyrillic("kh", "mn")
'х'
>>> cyrtranslit.to_cyrillic("kH", "mn")
'кH'
>>> cyrtranslit.to_cyrillic("Sh", "mn")
'Ш'
>>> cyrtranslit.to_cyrillic("SH", "mn")
'СH'
>>> cyrtranslit.to_cyrillic("TS", "mn")
'ТС'
>>> cyrtranslit.to_cyrillic("Ts", "mn")
'Ц'

This was raised in #12 by @097115. To reproduce this issue:

python cyrtranslit.py -l RU -i tests/ru.txt

python 3.5.2
cyrtranslit-0.3

Traceback:

>>> import cyrtranslit
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "[...]/lib/python3.5/site-packages/cyrtranslit/__init__.py", line 2, in <module>
    from mapping import TRANSLIT_DICT
ImportError: No module named 'mapping'

It seems to me that the safest code would simply combine all 3 - sr, me and mk.

The extra letters in me and mk do not conflict with any sr letters.

That way, a client can pass any Western South Slavic Cyrillic text, and be guaranteed an output.

(Likewise, in production code I would want Russian or Bulgarian letters to be handled in some way, in case they occur in text my code must process.)

Was fine before commit 26a63d6