Deepl does not seem to work about translate HOT 16 CLOSED

@Marocco2

Hey and thank you for helping me!
What kind of request do I need to make? And how should I use the response from it?
Thank you in advance

This is the first call you should make to DeepL
https://gist.github.com/Marocco2/5268909dba2c450c214c240e90defb27

The response is a JSON with your IP and the id you've created, meaning it's ready to translate.

DeepL server can check which gateway is coming the request, meaning that deploying the py module in a server could be blocked even with proxies, VPN and TORs

Keep in mind that this code is outdated. Once has worked, but it's too spaghetti to be recovered. I made a quick glance to your code and you did a much better job keeping it organized. You should fix it in no time with proper edits

I added your implementation of DeepL's RPC API in bb433ca !

from translate.

Hey and thank you for helping me!

What kind of request do I need to make? And how should I use the response from it?

Thank you in advance

This is the first call you should make to DeepL
https://gist.github.com/Marocco2/5268909dba2c450c214c240e90defb27

The response is a JSON with your IP and the id you've created, meaning it's ready to translate.

DeepL server can check which gateway is coming the request, meaning that deploying the py module in a server could be blocked even with proxies, VPN and TORs

Keep in mind that this code is outdated. Once has worked, but it's too spaghetti to be recovered. I made a quick glance to your code and you did a much better job keeping it organized. You should fix it in no time with proper edits

from translate.

Are you sure you are correctly connected to the internet?

Yes

(maybe a firewall blocking http connections?)

I checked the firewall rules - nothing like that

Thanks for the help, I found the main problem culprit - it turned out to be the old version of urllib3

Solution: https://stackoverflow.com/questions/61076294/how-to-handle-requests-exceptions-invalidurl-failed-to-parse-in-python

from translate.

Also remember to update translatepy to v1.6

Thank you, I have already updated, although I installed it from the source code. I am preparing a PR for you so that Yandex Translator can start working correctly

from translate.

The problem with DeepL is that they seem to integrate some bot detection algorithm which I need to understand for it to work well.

The problem with pyppeteer is obviously the Chromium instance running with it which is way too heavy for something like translatepy (already experimented it with my other, old translation module pyGoogleTranslate)

For now, DeepL seems to work from my location but I guess that any kind of suspicious activity (even calling the API a few times will get me rate limited with the same message "Too many requests")

Also, try to use the Translator class from translatepy.Translator as some things are optimised with it (tip: you can use or not a service over another with the use_service arguments)

Translator(use_google=True, use_yandex=True, use_bing=True, use_reverso=True, use_deepl=True, yandex_sid_refresh=False)

from translate.

I'll close this issue since it didn't get any response for 10 days

from translate.

@Animenosekai Maybe I can help.
First, you need to send a getClientState json to DeepL to accept your "ticket" called id

From now on, your id is gonna increase by one every time you want to reach DeepL. This is the first step to make your request more "genuine"

from translate.

Hey and thank you for helping me!

What kind of request do I need to make? And how should I use the response from it?

Thank you in advance

from translate.

Some serious problem seems like:

from translate.

Some serious problem seems like:

Are you sure you are correctly connected to the internet? (maybe a firewall blocking http connections?)
What makes me say this is that this behaviour occurs when translatepy skipped all of the translators (and I just fixed it to just return None when there is no result, which is normally the expected value)

Also weird exception because I don't get any problem parsing the same URL with the same function:

But seeing this exception doesn't mean that it stopped there (hence the returned value) as translatepy should not throw any exception from the Translator class

from translate.

@ZhymabekRoman Yea thought so (it was very weird that urllib3 gave this exception while there were no mistake in the URL)

Also remember to update translatepy to v1.6 (pip install --upgrade translatepy) because I fixed some things (should just give you None now instead of a TranslationResult with None in it in your case)

from translate.

@ZhymabekRoman Thank you!
Also, look at #4 to have an understanding of what we already tried to do with Yandex Translate 〜

from translate.

I will now close this issue as everything has been fixed in translatepy v1.6.1

from translate.

from time import sleep
from translatepy.translators.deepl import DeepL

dl = DeepL()

for elm in ['test this and that'] * 10:
    print(dl.translate(elm, 'de'))
    sleep(1)

# (None, None)
# (None, None)
# (None, None)
# (None, None)
# (None, None)
# (None, None)
# (None, None)
# (None, None)
# (None, None)
# (None, None)

Thanks for the hardwork. Sadly it still does not seem to work. Admittedly it works sometimes. Oftentimes it does not.

from translate.

from time import sleep
from translatepy.translators.deepl import DeepL

dl = DeepL()

for elm in ['test this and that'] * 10:
    print(dl.translate(elm, 'de'))
    sleep(1)

# (None, None)
# (None, None)
# (None, None)
# (None, None)
# (None, None)
# (None, None)
# (None, None)
# (None, None)
# (None, None)
# (None, None)

Thanks for the hardwork. Sadly it still does not seem to work. Admittedly it works sometimes. Oftentimes it does not.

Hmmmm I'll look into but won't be able to make too much changes since it might conflict with what @ZhymabekRoman is doing in the v2

Also, try using the Translator class from translatepy.Translator() because it provides fixes on the language name, the words, caches and fallbacks (if you just want to use DeepL, try disabling the other ones with use_bing=False, etc.)

from translate.

Admittedly it works sometimes. Oftentimes it does not.

Unfortunately seems this problem cannot be solved. I looked at all the implementations of the DeepL translator on the vastness of Github - no one works more or less stably. https://stackoverflow.com/questions/56114463/how-to-request-a-translation-from-deepl-at-the-moment

I analyzed the Windows version of the DeepL application and found out that it just opens the web address in a separate Chrome process - https://www.deepl.com/en/windows/translator?windows_app_version=2.3.0%2b6072bbdc (why then develop a separate application at all if it just opens a separate web page?). I opened this web page in a browser and tried to make a lot of translation requests, and it seems like I did not catch the error "Too many requests".

The only, as it seems to me, normal option is to use the official Deepl API Free. But it also has limitations-a maximum of 500,000 characters per month. I tried to register to get a free API key - it turns out that I need to link a bank card, but it does not support the countries of the post-Soviet states

from translate.