Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Sep 30 20:44
    Yorwba commented #2544
  • Sep 30 19:35
    AndiPersti commented #2544
  • Sep 29 19:38
    Yorwba commented #2544
  • Sep 29 19:19
    Yorwba labeled #2567
  • Sep 29 19:19
    Yorwba opened #2567
  • Sep 29 17:20
    AndiPersti edited #2528
  • Sep 29 13:40
    AndiPersti commented #2544
  • Sep 29 10:51
    AndiPersti synchronize #2566
  • Sep 29 10:20
    AndiPersti synchronize #2566
  • Sep 28 14:35
    AndiPersti opened #2566
  • Sep 28 13:39
    AndiPersti opened #2565
  • Sep 28 11:50

    AndiPersti on prod_2020-09-27

    (compare)

  • Sep 28 11:49

    AndiPersti on master

    Added language Old Frisian Added language Nuer Use correct language code for f… and 4 more (compare)

  • Sep 28 11:47

    AndiPersti on dev

    Pull translations from Transifex (compare)

  • Sep 27 07:21
    sabretou commented #2529
  • Sep 27 07:18
    soliloquist-tatoeba commented #2540
  • Sep 26 21:43
    soliloquist-tatoeba commented #2528
  • Sep 26 17:19
    AndiPersti opened #2564
  • Sep 26 16:16
    trang commented #2529
  • Sep 26 15:25

    AndiPersti on dev

    Update POT files (compare)

arslan
@arslun
the last step (ansible-playbook ./local.yml) failed with the following error message:
Error! vars file h was not found
any suggestions?
Thanks
Trang
@trang
@arslun We replied to you via email, as you also contacted us via email. Just pinging you here too, in case.
ghlai9665
@ghlai9665
Hi guys I'm new to Tatoeba. I would like to get a corpus of example sentences of Chinese with English translations. What's the best way to do that? I downloaded the csv file from here https://tatoeba.org/eng/downloads, but I don't seem to be able to find Chinese sentences in it.
Thanks so mcuh
Trang
@trang
@ghlai9665 Hi, the Chinese sentences (assuming you're looking for Mandarin Chinese) are under the code cmn. There should be around 60,000 sentences, but you have to extract them from the sentences.csv file (which contains all sentences in all languages). Please refer to our FAQ: https://en.wiki.tatoeba.org/articles/show/faq#how-can-i-download-all-sentences-and-translations- (the last question). If this doesn't help, then please let me know more precisely what issues you run into exactly.
Ricardo14
@RyckRichards

Please could you help me?
I stucked on https://getcomposer.org/download/ -

It says http://prntscr.com/ndr8gz

but do I create one?

**how do I create one?
Thank you!
Trang
@trang
@RyckRichards It should be C:\xampp\php\php.exe. Just click on "Browse" and paste this path.
Yanis Batura
@Inego
Hey-hey everyone. Gitter shows me the last message here is dated Apr 19, or am I missing something?
Is this chat abandoned? I'll try the XMPP chatroom, maybe more luck there :)
Trang
@trang
It's not exactly abandoned, just not used much these days :)
Yanis Batura
@Inego
In the frontend / JS, what is the correct way to get the root path to the site?
For prod, this is https://tatoeba.org, and for dev, it's localhost:8080
This string is required to decide whether the text in the clipboard is indeed a URL of a Tatoeba sentence.
Also, if I will have a regex for that containing /sentences/show part, will it be robust in case when the URL schema / naming changes in the future?
Yanis Batura
@Inego
Currently, the regex to extract the sentence number only looks for the first sequence of digits after a slash:
image.png
Is it because there are several possible kinds of URLS which may contain the sentence number?
/sentences/show should be stable, yes
There's so far only one URL that contains the sentence number, I think... but there could be more in the future
Yanis Batura
@Inego
What about browser compatibility? Are browser compatibility requirements stated anywhere?
Trang
@trang
We don't have specific requirements when it comes to browser compatibility. We make things as compatible as possible with whatever amount of effort we are able to put into this :)
danail-georgiev
@danail-georgiev
@trang hi
Iwan Aucamp
@aucampia
hi
Trang
@trang
@aucampia You don't have to merge them. They are merge automatically by our bot Horus.
Iwan Aucamp
@aucampia
great thanks
Brendan Martin
@BrendanMartin
Hey, I was wondering if anyone could shed light on how the search query picks up on all conjugations of a word?
Trang
@trang
@BrendanMartin We're using a stemming library. In case you're not familiar with stemming: https://en.wikipedia.org/wiki/Stemming
Yorwba
@Yorwba
Specifically, the search is powered by Manticore and the documentation for their stemming functionality is here: https://docs.manticoresearch.com/latest/html/conf_options_reference/index_configuration_options.html#morphology
Brendan Martin
@BrendanMartin
@Yorwba how do you index and search character-based languages, like Chinese?
Yorwba
@Yorwba
We use Manticore's support for ngram extraction, although we set n=1 so we're really just searching characters individually. It works well enough for Mandarin, I think. Here's the Manticore documentation for ngrams: https://docs.manticoresearch.com/latest/html/conf_options_reference/index_configuration_options.html#ngram-chars
Brendan Martin
@BrendanMartin
Oh I see. Thanks for the info.
I'm trying to implement a similar text search in Postgres but can't seem to get it as fast as the Tatoeba search
Safa Alfulaij
@SafaAlfulaij
image.png
What?!
Trang
@trang
@SafaAlfulaij I'm not sure what you are trying to say exactly...
Safa Alfulaij
@SafaAlfulaij
That is clearly not proper Arabic
العربية
Trang
@trang
I see. Indeed the language name changed between the old config file and the new one... It was written العربية before, not sure how it became ﺎﻠﻋﺮﺒﻳﺓ
I'll check this. Thanks.
Safa Alfulaij
@SafaAlfulaij
Exactly. I mentioned the commit where that happened
Perhaps the text editor used messed it up
Thanks for your efforts :)
Trang
@trang
@SafaAlfulaij The Arabic language name should be fixed now. Could you confirm?
Safa Alfulaij
@SafaAlfulaij
Confirmed, sorry for the long deley :(
Safa Alfulaij
@SafaAlfulaij
Nice improvments to the website :)