Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Jun 30 06:42
    nijel assigned #4668
  • Jun 30 06:42
    nijel assigned #4669
  • Jun 30 06:42
    nijel closed #4669
  • Jun 30 06:42

    nijel on pip

    (compare)

  • Jun 30 06:42

    nijel on master

    build(deps-dev): bump virtualen… (compare)

  • Jun 30 06:42
    nijel closed #4668
  • Jun 30 06:42

    nijel on master

    build(deps-dev): bump setuptool… (compare)

  • Jun 30 06:42

    nijel on pip

    (compare)

  • Jun 29 23:04
    dependabot[bot] labeled #4669
  • Jun 29 23:04
    dependabot[bot] opened #4669
  • Jun 29 23:04
    dependabot[bot] labeled #4669
  • Jun 29 23:04

    dependabot[bot] on pip

    build(deps-dev): bump virtualen… (compare)

  • Jun 29 23:04
    dependabot[bot] labeled #4668
  • Jun 29 23:04
    dependabot[bot] labeled #4668
  • Jun 29 23:04
    dependabot[bot] opened #4668
  • Jun 29 23:04

    dependabot[bot] on pip

    build(deps-dev): bump setuptool… (compare)

  • Jun 29 12:22
    nijel commented #4667
  • Jun 29 12:22
    nijel closed #4667
  • Jun 29 12:22

    nijel on master

    Handle '\r' in rc file strings … (compare)

  • Jun 29 12:16
    nijel assigned #4667
Stuart Prescott
@llimeht
Any thoughts on https://bugs.debian.org/916745 / translate/translate#3838 ? I really don't want to diverge in Debian without this being reviewed and preferably merged.
@nijel (simply because you seem most active at the moment....) any thoughts on the above?
Michal Čihař
@nijel
I think having tests covering the unicode strings handling as mentioned by @unho in review is way to go - that way we can clearly see whether it doesn't break anything
Ilia Shifrin
@protranslating
hi, looking for the best way to get source XLIFF segments parsed in python sending them one at a time to GCP AutoML via Google Translate API for a given custom-trained NMT model and then serializing the API responses back into the file
Got to a point when I can read a file and send it to Google Cloud to have translated with a specific custom-trained model and output into a separate file
but the file is plain text file and I need to do a bunch of custom model testing by getting XLIFFs translated with it retaining the format
found this guy's repo https://github.com/translate/pyliff
Ilia Shifrin
@protranslating
but it was last updated like 3 years ago and has this description "Currently XLIFF data can only be read. In future versions you will be able to manipulate the XLIFF data and serialize it back to a file." so it seems the dev gave up on it for whichever reason
anyone knows if a similar solution exists?
Ilia Shifrin
@protranslating
thinking of getting around with an XLIFF to CSV conversion but thought maybe there is an existing solution to work directly with an XLIFF
Michal Čihař
@nijel
@protranslating translate-toolkit can both read and write xliff
friedelwolff
@friedelwolff
@jason-p-pickering : I see you mentioned problems with amagama a while ago. I'm working on that now, so hopefully things are better soon.
It might already work now, but I'm still busy, so expect it go up and down for a little while longer.
Stuart Prescott
@llimeht
So given the latest release of ttkit breaks virtaal, is there any chance of a virtaal release soon? ;)
Stuart Prescott
@llimeht
(translate/virtaal@77c43a9 on its own cannot be cherry-picked; it needs some of the preceding patches too, which makes me nervous)
Leandro Regueiro
@unho
@llimeht Maybe you need to ping Friedel Wolff in GitHub regarding a new Virtaal release. I am not sure he reads this channel.
Allan Nordhøy
@kingu_gitlab
A new Virtaal would really do nicely
friedelwolff
@friedelwolff
I would really like to work on Virtaal again
Virtaal isn't packaged for some distros anymore (just recently saw it is not on fedora, for example)
I guess a flatpak would solve parts of the distribution problem
there are already several nice improvements in git since the last release
It is unfortunate that backwards compatibility was not important enough for the toolkit
friedelwolff
@friedelwolff
@unho By the way: I see you didn't announce the release on the mailing list. Did you manage to post? I recently was unable to post to a sourceforge mailing list.
Stuart Prescott
@llimeht
@friedelwolff we've still got a virtaal package for Debian and Ubuntu. I can drop in the newer version of that file into virtaal and it is probably OK (essentially cherrypicking 4 commits just to it).
However, Python 2 and gtk2 will both be going away in Debian and Ubuntu reasonably soon, I think, and that will force other action.
friedelwolff
@friedelwolff
@llimeht Yes. that is a bigger problem. The rewrite is a non-trivial amount of work. It comes down to three ports (pygtk -> gobject introspection, gtk2 to gtk3, python2 -> python3)
we struggle with providing long-term platforms in our world, it seems.
Stuart Prescott
@llimeht
yes, although I don't think we should say that any of the existing ones were short term platforms, hanging around for more than a decade ;)
(And it's better than Qt, which seems determined to kill off all Qt-using software with incompatible upgrades)
Mixing these 3 big transitions up is certainly a nasty challenge though. I've never done any gtk programming at all to know how difficult that might be. ttkit and friends are probably amongst the hardest things to do the py2→3 transition too :(
friedelwolff
@friedelwolff
well, I'm very comfortable with Python, so I don't mind that py2-3 migration too much
but it requires time
Leandro Regueiro
@unho
@friedelwolff I honestly didn't try to post on the mailing list.
@friedelwolff Regarding backwards compatibility there is a point when we should stop supporting everything. I sincerely hope this is another kick to get another Virtaal release out, this time including fixed requirements to avoid new TTK releases breaking compatibility for already rolled out Virtaal installs.
BTW, I have plans for removing support for Python 2 in any TTK release after Python 2.7 support goes away, which basically is 1st January 2020, which is in less than 6 months from now: https://pythonclock.org/
friedelwolff
@friedelwolff
@unho Of course there is such a point. We just disagree on where that point is.
fixed requirements won't help in the case of distros that want to upgrade the toolkit
I'd like to do a million things on virtaal, but my life went through multiple very disruptive changes in the last few years
Let's hope I can look into it more
friedelwolff
@friedelwolff
petersen from #fedora-i18n just said that he is updating the translate toolkit packages in Fedora!
prog20901
@prog20901

Hi,

I have huge text files - raw text. I want to convert them to a different language. Google/Microsoft/Yandex is best for online conversions. However we need to copy and paste and moreover it has size limitation as well.

Is there is any tool which i can use for bulk conversion at one shot which can give the output in multiple files.

Thanks in advance.

Allan Nordhøy
@comradekingu
@phlax I am all for it, the pipeline structure seems more complete over at GitLab
prog20901
@prog20901

Save the word as text file after applying font with UTF-8

i have a word file which shows boxes when i opened. However when choosing a particular font, the text is being displayed. I selected all the content of the document with ctr+a and changed the font and save as .txt file with UTF-8.

Everything works fine.

Is there is any macro can do this automatically? Basically applying a font for entire document and save as .txt file with UTF-8?

Any tool or software does this?

Tual Maxime
@filozofer
Hey ! I'm trying to install a local tm server using elasticsearch. My translations has been successfully indexed inside my ES instance but I don't see any change on my pootle website. There is no suggestion provide when i'm translating. Is there something to enable ? Have you a screen or videos showing this ? Thanx !
Tual Maxime
@filozofer
I finally understand how it's work. The tm_suggestion goes directly under the translations when there is results. In my case there was no result because I only have one project and my unit source are technical keys.
Is it possible to configure how the tm server store the translations and retrieve them ? Because in my case the feature is useless. My .po files use a technical key as "msgid" and become the unit.source in pootle (exemple: company.edit.title). In this case, pootle try to find on the pm server how I translate "company.edit.title" which is a nonsence. I want him to search for the english translation for exemple: How did I translate "Edit company:" string in my previous project.
Is this possible without changing my po files ?
Tual Maxime
@filozofer
One other thing, when I translate my italian units, pootle show me tm_suggestion in technical to french language. Why ?
prog20901
@prog20901
What is the difference between Markdown vs Docx vs HTML files?
Bachstelze
@Bachstelze
Markdown has very easy syntax and is often used for documentations. Docx is commercial garbage and HTML is the format for nearly every website.