terryyin/translate-python

auto-detect doesn't work when translating from zh_CN to en

Allianzcortex opened this issue · 9 comments

Allianzcortex commented

Awesome Job ! This project really helps me a lot ! But maybe there is a small issue about language auto-detect when translating from zh_CN to en.

Based on the document , if you want to translate a sentence from en to zh_CN , you just need to use three lines:

In [1]: from translate import Translator
In [2]: translator= Translator(to_lang="zh")
In [3]: translation = translator.translate("This is a pen.")
Out [3]: 这是一支笔

So based on my understanding , if you want to implement the reversed action, you also need three lines:

In [1]: from translate import Translator
In [2]: translator= Translator(to_lang="en")
In [3]: translation = translator.translate("这是一支笔")
Out [3]: 这是一支笔 # Still the original sentence. translation doesn't work

But if you specify the input language,it will work:

In [1]: from translate import Translator
In [2]: translator= Translator(from_lang="zh_CN",to_lang="en")
In [3]: translation = translator.translate("这是一支笔")
Out [3]: This is a pen # Translate successfully

Below is the screenshot tested on python3.5:

image

dejongyeong commented

Same issue too. Please fix, Thanks.

Deft-pawN commented

Same issue ~ Please fix, Thanks.

rhenter commented

This problem is on https://mymemory.translated.net. We are investigating to fix this problem.

  • 👍1
terryyin commented

Anybody can confirm if this is recovered?

JakkuSakura commented

This is because there is a default value 'en' for from_lang. As a result, auto-detect does not work. Check out the following codes.

class Translator:
    def __init__(self, to_lang, from_lang='en', provider=None, secret_access_key=None, **kwargs):

How do you translate "学而时习之“ from English to English?

terrosdesigns commented

Quick fix:

from translate import Translator
translator= Translator(from_lang="zh", to_lang="en")
translation = translator.translate("这是一支笔")
'This is a pen' #Output

Hope it helps !

  • 👍1
JakkuSakura commented

Quick fix:

from translate import Translator
translator= Translator(from_lang="zh", to_lang="en")
translation = translator.translate("这是一支笔")
'This is a pen' #Output

Hope it helps !

In this case, however, auto-detect cannot work at all.

terrosdesigns commented

Quick fix:

from translate import Translator
translator= Translator(from_lang="zh", to_lang="en")
translation = translator.translate("这是一支笔")
'This is a pen' #Output

Hope it helps !

In this case, however, auto-detect cannot work at all.

There you have it !

from langdetect import detect
from translate import Translator

def translate(sentence):
    lang_detected = detect(sentence)
    print(lang_detected)
    translator= Translator(from_lang=lang_detected, to_lang="en")
    translation = translator.translate(sentence)
    return translation

translate('"这是一支笔"')
# zh-cn
# 'This is a pen'
  • 🎉2
stoufa commented

I was going to share the same solution as yours @terrosdesigns 😁
I hope they can add langdetect as a dependency to this project, this way, we can make MyMemory's Translation API do less work and focus on "Translation" rather than "Language Detection". 😊