Assigned
Status Update
No update yet.
Comments
at...@google.com
I have one question related to Google translate API. but it is not related to this issue.
When i am trying to access Google Translate API using javascript when the query string size is around 13000 characters i am getting server response error code 400. Can you please let me know what is the request size limit if any exist.
When i am trying to access Google Translate API using javascript when the query string size is around 13000 characters i am getting server response error code 400. Can you please let me know what is the request size limit if any exist.
da...@gear4music.com
@vrames
according to Google API FAQ
The maximum size of each text to be translated is 5000 characters, not including any HTML tags.
according to Google API FAQ
The maximum size of each text to be translated is 5000 characters, not including any HTML tags.
al...@gmail.com
N/A
em...@tomas-novotny.eu
I also have a same problem some html tags who are divided in few tags , and it's problematic .. there is some news about it ??
zo...@gmail.com
I met with the problem just now, it seems the bug nof fixed yet.
ro...@gmail.com
Update: I found the bug triggered not only by anchor tag, but also others. It seems that the problem is the context, and when I replace the tag with special symbol, 【】 for example, it worked.
In some context:
<a id="text">bla bla...</a> -> <a id="text">alb</a>alb<a id="text">...</a>
after change anchor tag to 【】,it worked:
【bla bla ...】 -> 【alb alb ...】
In some context:
<a id="text">bla bla...</a> -> <a id="text">alb</a>alb<a id="text">...</a>
after change anchor tag to 【】,it worked:
【bla bla ...】 -> 【alb alb ...】
mo...@gmail.com
This same issue exists for <span> tags. If I send the translate API a document consisting of text in span tags, each with a unique id, the result comes back with many of the span tags mixed up, nested and duplicated. A partial example from a test document, translated from German to English:
Input: <span id='1'>Per </span><span id='2'>Web-ERV</span><span id='3'>Handelsgericht Wien</span><span id='4'>Marxergasse 1a</span><span id='5'>1030 Wien</span><span id='6'>Firmenbuchsache</span>
Output: <span id='1'>Via</span> <span id='2'>web ERV</span> <span id='3'>Vienna Commercial Court</span> <span id='4'>Marxergasse 1a</span> <span id='5'>1030 <span id='6'>Wien</span></span> <span id='6'>Commercial thing</span>
Notice above in the output from the API, there is now an extra <span id="6"> now nested in <span id="5">, around partial content of <span id="5">.. This is completely unacceptable for a paid service, html translation is completely unusable with this bug.
Input: <span id='1'>Per </span><span id='2'>Web-ERV</span><span id='3'>Handelsgericht Wien</span><span id='4'>Marxergasse 1a</span><span id='5'>1030 Wien</span><span id='6'>Firmenbuchsache</span>
Output: <span id='1'>Via</span> <span id='2'>web ERV</span> <span id='3'>Vienna Commercial Court</span> <span id='4'>Marxergasse 1a</span> <span id='5'>1030 <span id='6'>Wien</span></span> <span id='6'>Commercial thing</span>
Notice above in the output from the API, there is now an extra <span id="6"> now nested in <span id="5">, around partial content of <span id="5">.. This is completely unacceptable for a paid service, html translation is completely unusable with this bug.
ma...@gmail.com
Hi there,
any update on this one?
any update on this one?
sa...@gmail.com
Hi,
This issue is also affecting a user in case number 17353456.
Is there any update?
This issue is also affecting a user in case number 17353456.
Is there any update?
zo...@gmail.com
Hi, any updates on this?
to...@gmail.com
I guess it's not an easy one.
I kind of "solved" this by replacing my placeholders with some utf8 encoded string.
I was having trouble with strings like: "{s} Lorem Ipsum is simply dummy text of the printing and typesetting industry." but if I encode it like "{123_115_125} Lorem Ipsum is simply dummy text of the printing and typesetting industry." it seems to work fine.
Of course it's kind of cheating but it worked so far.
I kind of "solved" this by replacing my placeholders with some utf8 encoded string.
I was having trouble with strings like: "{s} Lorem Ipsum is simply dummy text of the printing and typesetting industry." but if I encode it like "{123_115_125} Lorem Ipsum is simply dummy text of the printing and typesetting industry." it seems to work fine.
Of course it's kind of cheating but it worked so far.
jo...@kognity.com
We're seeing these issues as well with html tags being swapped around breaking links. Example output:
<p class="tile-land-image">
<a href="../tutorials/1_overview.htm">
<img class="tile-image" src="../Resources/images/thumbnails/landing_environs.png"></a> </img>
</p>
<p class="tile-land-image">
<a href="../tutorials/1_overview.htm">
<img class="tile-image" src="../Resources/images/thumbnails/landing_environs.png"></a> </img>
</p>
sh...@duve.com
Hi, I am trying to fix it, could you provide some examples, including input/output sentences?
ja...@massivemedia.eu
Below is an example where the anchor tag goes missing after translating:
Part of html Sent for translation:
<a class="d-flex align-items-center text-decoration-none" href="/customers">
<div class="ml-4">
<h6 class="font-weight-bold text-uppercase text-primary mb-0">Customer</h6>
</div>
</a>
Part of html received after translation:
<div class="ml-4">
<h6 class="font-weight-bold text-uppercase text-primary mb-0"> Client </h6>
</div>
This can be reproduced following the code from the Python guide[1]. The library version used is google-cloud-translate==2.0.0
[1]https://cloud.google.com/translate/docs/advanced/translating-text-v3#translating_input_strings
Part of html Sent for translation:
<a class="d-flex align-items-center text-decoration-none" href="/customers">
<div class="ml-4">
<h6 class="font-weight-bold text-uppercase text-primary mb-0">Customer</h6>
</div>
</a>
Part of html received after translation:
<div class="ml-4">
<h6 class="font-weight-bold text-uppercase text-primary mb-0"> Client </h6>
</div>
This can be reproduced following the code from the Python guide[1]. The library version used is google-cloud-translate==2.0.0
[1]
nw...@gmail.com
The product engineering team is aware of this issue and are investigating solutions. There is no ETA at this time for a fix, but all further updates should occur here.
cz...@google.com
se...@google.com
This issue hasn't been fixed after more than 6 years now?
th...@gmail.com
c'mon Google - 7 years and still nothing??
ym...@gmail.com
My theory from day one: the added spaces are markers, so Google's crawlbot can identify auto translated content on the web more easily. In other words: a way to weed out most auto-translated pages. Not sure if it is an effective strategy given there are other great translator APIs or weird work-arounds to remove the spaces.
pe...@psmits.com
same with me why is this problem not solved after more then two years?
It appears that if the verb includes the % it will be broken into "% 1 $ s"
How can I prevent this from happening ?
It appears that if the verb includes the % it will be broken into "% 1 $ s"
How can I prevent this from happening ?
yi...@google.com
Hi all,
From the thread, there are two types of space issue, one is introduced between html tags, another is between some special characters. I made some improvements on the html one, the special characters one will be in prod later.
Now the html space will not have more than one space between two plain words,
Before: <a>Text</a> <p> Hallo <a>Text</a> </p>
Now: <a>Text</a> <p>Hallo <a>Text</a></p>
It may not respect to the source sentence, but it can guarantee the browser can show correct sentences.
From the thread, there are two types of space issue, one is introduced between html tags, another is between some special characters. I made some improvements on the html one, the special characters one will be in prod later.
Now the html space will not have more than one space between two plain words,
Before: <a>Text</a> <p> Hallo <a>Text</a> </p>
Now: <a>Text</a> <p>Hallo <a>Text</a></p>
It may not respect to the source sentence, but it can guarantee the browser can show correct sentences.
te...@synology.com
The same issue appears here with extra spaces randomly added to html tags.
API version:
"google/apiclient": "^2.0",
"google/cloud-automl": "^0.10.0",
"google/cloud-translate": "^1.7"
Example:
Source language:
<li>If you must access file services over the Internet, it is strongly recommended that you use a VPN to connect to your device (see <a href="website.com ">Tutorial</a>).</li>
target language:
<li>Si vous devez accéder aux services de fichiers sur Internet, il est fortement recommandé d'utiliser un VPN pour vous connecter à votre périphérique (voir <a href="website.com ">Didacticiel</a> ).</li>
What we expected to happen:
<li>Si vous devez accéder aux services de fichiers sur Internet, il est fortement recommandé d'utiliser un VPN pour vous connecter à votre périphérique (voir <a href="website.com ">Didacticiel</a>).</li>
API version:
"google/apiclient": "^2.0",
"google/cloud-automl": "^0.10.0",
"google/cloud-translate": "^1.7"
Example:
Source language:
<li>If you must access file services over the Internet, it is strongly recommended that you use a VPN to connect to your device (see <a href="
target language:
<li>Si vous devez accéder aux services de fichiers sur Internet, il est fortement recommandé d'utiliser un VPN pour vous connecter à votre périphérique (voir <a href="
What we expected to happen:
<li>Si vous devez accéder aux services de fichiers sur Internet, il est fortement recommandé d'utiliser un VPN pour vous connecter à votre périphérique (voir <a href="
ap...@gmail.com
please fix this
br...@gmail.com
Workaround I came up with This is for text translation. I don't know how google behaves for html input. Could be similar...
Before passing the string to translator, run through this regex /(?=[^a-zA-Z _,.])/gm which matches all special characters and replace the matches with a special character that's not present in the string. For me, I used * character. Then pass it on to the translator and run the translated string through another simple regex /( * )|( *)|(* )/gm and replace the matches with empty string. Then you get the desired output. Other languages where input string contain non alphabet chars can filter out those chars too in the first regex if needed but it's not necessary. The first regex can further be modified(if speed is a concern) to match only those special characters which doesn't have a preceeding space character. Google won't add space to any special character which already has a preceeding space..
g3...@gmail.com
Any news about it? The whole Wikipedia Translation service is suffering.
se...@google.com
Hello,
Thank you for your patience.
I have forwarded your concerns to the product and they are working on it. Once I have an update it will be shared in this thread.
Kind Regards
se...@google.com
al...@gmail.com
Comment has been deleted.
su...@google.com
ar...@gmail.com
Any update on this? Content Translation on Wikipedia is still showing this bug.
si...@google.com
si...@google.com
ty...@gmail.com
Comment has been deleted.
ty...@gmail.com
Hey There,
Can we get an update on this issue since it has been a year since the last communication on this ticket or get its priority upped to a P1 since this issue has been around for four years and is the second most user-impacted cloud translation bug?
Thanks and best regards!
Can we get an update on this issue since it has been a year since the last communication on this ticket or get its priority upped to a P1 since this issue has been around for four years and is the second most user-impacted cloud translation bug?
Thanks and best regards!
ba...@mellink.net
Hi Google
This is annoying (also because the API is a paid service). Any update would be appreciated. Problem also happens with special HTML characters such as & for '&' and it seems to be depending on the fact if the words around the characters are in upper or lower case. Examples (translating from EN to FR):
The word "Crypto" is in uppercase: Extra space added in the middle of the & char:
The Crypto Fear & Greed index --> La crypto peur & amp; Indice de cupidité
The word "crypto" is in lowercase: No extra space added in the middle:
The crypto Fear & Greed index -> La crypto Peur & Indice de cupidité
The location of the space is inconsistent. Sometimes & will translate to & amp; and other times to & ;
This is annoying (also because the API is a paid service). Any update would be appreciated. Problem also happens with special HTML characters such as & for '&' and it seems to be depending on the fact if the words around the characters are in upper or lower case. Examples (translating from EN to FR):
The word "Crypto" is in uppercase: Extra space added in the middle of the & char:
The Crypto Fear & Greed index --> La crypto peur & amp; Indice de cupidité
The word "crypto" is in lowercase: No extra space added in the middle:
The crypto Fear & Greed index -> La crypto Peur & Indice de cupidité
The location of the space is inconsistent. Sometimes & will translate to & amp; and other times to & ;
da...@ex-situ.com
Hello, any update on this issue? still seeing it in 2023
pe...@foundry.com
Hey There,
I am wondering if there is any update on this, I'm still getting this issue when using the google.cloud.translate_v2 API?
Cheers
I am wondering if there is any update on this, I'm still getting this issue when using the google.cloud.translate_v2 API?
Cheers
ku...@google.com
pu...@google.com
27...@student.pwr.edu.pl
issue is still present and silly after processing is still needed.
ri...@gmail.com
How can this still be present? no activity to solve this bug?
va...@gmail.com
Encountering this same issue when wrapping text with a <Span translate="no"> tag that is followed by punctuation.
ni...@babbels.nl
Encountering this issue too, it seems it's been 3 years since someone with a google.com email said in this thread that the team is working on it and any progress would be shared here.
Any progress in the last 3 years?
Any progress in the last 3 years?
me...@gmail.com
Any update on this? this is bad.. Content moderation is difficult .. code snippets are having additional spaces :(
Description
Google break add extra space to some tags
Example: <a>text</a><p>Hello <a>text</a></p><li>Hello</li><li>World</li>
Get: <a>Text</a> <p> Hallo <a>Text</a> </p><li> Hallo </li><li> Welt </li>
What you expected to happen:
Should be translated: <a>Text</a><p>Hallo <a>text</a></p><li>Hallo</li><li>Welt</li>
But Google break word by added space between letters
Steps to reproduce:
The input text to translate - <a>text</a><p>Hello <a>text</a></p><li>Hello</li><li>World</li>
Target: de or ru
Source: en