Defect report [78317807]

Fixed

Bug

Status Update

No update yet.

Description

oi...@gmail.com

created issue #1

Apr 20, 2018 08:51PM

HTML API requests returns incorrect HTML:

Here's the original Text that I was translating (this is from Wikipedia - It was a long page, so I was breaking it up to less than 5000 Characters per piece as Translate API requests, being careful to keep it valid, well-formatted HTML):

<h3 class="in-block"><span id="1968–present:_Graphical_User_Interface"></span><span class="mw-headline" id="1968.E2.80.93present:_Graphical_User_Interface">1968–present: Graphical User Interface</span><span><a class="mw-ui-icon mw-ui-icon-element mw-ui-icon-edit-enabled edit-page" data-section="7" href="/w/index.php?title=User_interface&action=edit&section=7" title="Edit section: 1968–present: Graphical User Interface">Edit</a></span></h3>

I was translating from English to Swahili, with the format = html and using the python client library.
The result was as follows:

<h3 class="in-block"><span id="1968–present:_Graphical_User_Interface"></span> <span class="mw-headline" id="1968.E2.80.93present:_Graphical_User_Interface">1968-sasa:</span> <span><a class="mw-ui-icon mw-ui-icon-element mw-ui-icon-edit-enabled edit-page" data-section="7" href="/w/index.php?title=User_interface&action=edit&section=7" title="Hariri sehemu: 1968-sasa: Interface User Graphic">Edit</a></span> <span class="mw-headline" id="1968.E2.80.93present:_Graphical_User_Interface">Graphical User Interface</span> <span><a class="mw-ui-icon mw-ui-icon-element mw-ui-icon-edit-enabled edit-page" data-section="7" href="/w/index.php?title=User_interface&action=edit&section=7" title="Hariri sehemu: 1968-sasa: Interface User Graphic">Edit</a></span></h3>

As you can see the "1968–present: Graphical User Interface" was somehow made into 2 elements with a split on the ":" And the "EDIT" element was duplicated.

This bug has serious visual and functional implications.

Comments

oi...@gmail.com <oi...@gmail.com> #2Apr 20, 2018 08:57PM

Easiest way to visually see the issue is to highlight "span" or "<a" and see the different number of those tags in the 2 snippets.

This snippet is originally from Wikipedia's Mobile User interface page:

https://en.m.wikipedia.org/wiki/User_interface

This bug shows up even when translating the page with the

translate.google.com interface here:

https://translate.google.com/translate?sl=en&tl=sw&js=y&prev=_t&hl=en&ie=UTF-8&u=https%3A%2F%2Fen.m.wikipedia.org%2Fwiki%2FUser_interface&edit-text=&act=url

notice the double pencil icon in the 1968-sasa section and the separate "Graphical User Interface" tag instead of all one part of the title.

Message last modified on Apr 20, 2018 09:04PM

kn...@google.com <kn...@google.com> Apr 21, 2018 09:43PM

Assigned to kn...@google.com.

kn...@google.com <kn...@google.com> #3Apr 22, 2018 12:12AM

Reassigned to gc...@google.com.

Thanks for reporting this issue. This issue is reported to the Cloud Translation API team so that they may investigate it, however there is no ETA for the fix at this time. All further updates should occur in this report.

I should note that here we support Google Cloud Platform-end products and for any support on Google Translate, I recommend to contact to Google Translate Help Center [1].

[1]

https://support.google.com/translate/?hl=en#topic=7011755

mm...@google.com <mm...@google.com> #4Apr 23, 2018 01:43PM

I reproduced the scenario and can confirm that I got the same behavior. The issue seems to be in the colon where the html splits and makes two elements. I was looking for a pattern but couldn't find any. I tried using one and more colons on different locations, but in all attempts I got the expected response.
examples:

input: <h3 class="in-block"><span id="1968–present:_Graphical_User_Interface"></span><span class="mw-headline" id="1968.E2.80.93present:_Graphical_User_Interface">1968–present Graphical: User Interface</span><span><a class="mw-ui-icon mw-u
i-icon-element mw-ui-icon-edit-enabled edit-page" data-section="7" href="/w/index.php?title=User_interface&action=edit&section=7" title="Edit section: 1968–present: Graphical User Interface">Edit</a></span></h3>

output: <h3 class="in-block"><span id="1968–present:_Graphical_User_Interface"></span> <span class="mw-headline" id="1968.E2.80.93present:_Graphical_User_Interface">1968-sasa Graphical: User Interface</span> <span><a class="mw-ui-ico
n mw-ui-icon-element mw-ui-icon-edit-enabled edit-page" data-section="7" href="/w/index.php?title=User_interface&action=edit&section=7" title="Hariri sehemu: 1968-sasa: Interface User Graphic">Edit</a></span> </h3>

input: <h3 class="in-block"><span id="1968–present:_Graphical_User_Interface"></span><span class="mw-headline" id="1968.E2.80.93present:_Graphical_User_Interface">1968–present: Graphical User: Interface</span><span><a class="mw-ui-icon mw-
ui-icon-element mw-ui-icon-edit-enabled edit-page" data-section="7" href="/w/index.php?title=User_interface&action=edit&section=7" title="Edit section: 1968–present: Graphical User Interface">Edit</a></span></h3>

output: <h3 class="in-block"><span id="1968–present:_Graphical_User_Interface"></span> <span class="mw-headline" id="1968.E2.80.93present:_Graphical_User_Interface">1968-sasa: Mtumiaji wa Graphical: Kihariri</span> <span><a class="mw
-ui-icon mw-ui-icon-element mw-ui-icon-edit-enabled edit-page" data-section="7" href="/w/index.php?title=User_interface&action=edit&section=7" title="Hariri sehemu: 1968-sasa: Interface User Graphic">Hariri</a></span> </h3>

yi...@google.com <yi...@google.com> #5Dec 5, 2019 11:50PM

actually it is not related to colons, I will try to see how to fix this problem.

sd...@google.com <sd...@google.com> #6Jun 9, 2020 05:54PM

Any updates?

ti...@google.com <ti...@google.com> #7Sep 16, 2020 07:37AM

A change has been made to fix the issue and it is being tracked.

al...@google.com <al...@google.com> #8Nov 11, 2020 08:23AM

Hello,

The product team is still working on a fix for this. They have made progress but is still not fully fixed as it will take some more time. further information should be posted here as soon as it is available.

mk...@google.com <mk...@google.com> #9Jan 22, 2021 09:26AM

Marked as fixed.

Hello,

The fix patch has been recently introduced to GA in v3 Translation API, therefore the current report can be closed. If the issue requires more further attention this public case can be re-opened at any time.