search a word in the all translations

[Axel R.-3]

Hi,
Some time, I found words which are translated by two differents way.
I would like to search a word in the all repository to find every
strings and their translation.

Thank you,

Axel (for the eo team)

[Og Maciel-3]

On Wed, Nov 4, 2009 at 8:39 AM, Axel R. <axel@...> wrote:
> Some time, I found words which are translated by two differents way.
> I would like to search a word in the all repository to find every strings
> and their translation.

Not 100% what you asked for but it does the trick:  http://open-tran.eu/

Cheers,
--
Og B. Maciel

omaciel@...
ogmaciel@...
ogmaciel@...

GPG Keys: D5CFC202

http://www.ogmaciel.com (en_US)
http://blog.ogmaciel.com (pt_BR)

[Nicolas Ternisien]

The page

http://l10n.kde.org/dictionary/search-translations.php

used to do this, but I've disabled the database behind, because it
consumed too many resources.

Sorry...

On Wed, Nov 4, 2009 at 2:45 PM, Og Maciel <ogmaciel@...> wrote:

> On Wed, Nov 4, 2009 at 8:39 AM, Axel R. <axel@...> wrote:
>> Some time, I found words which are translated by two differents way.
>> I would like to search a word in the all repository to find every strings
>> and their translation.
>
> Not 100% what you asked for but it does the trick:  http://open-tran.eu/
>
> Cheers,
> --
> Og B. Maciel
>
> omaciel@...
> ogmaciel@...
> ogmaciel@...
>
> GPG Keys: D5CFC202
>
> http://www.ogmaciel.com (en_US)
> http://blog.ogmaciel.com (pt_BR)
>

[Martin Schlander]

Onsdag den 4. november 2009 14:39:44 skrev Axel R.:
> Some time, I found words which are translated by two differents way.
> I would like to search a word in the all repository to find every
> strings and their translation.

grep is one simple option.

But let me advertise another nifty tool.

pyg3t is a little suite of python scripts for gettext translators.
https://launchpad.net/pyg3t

It includes among a few other things a neat thing called 'gtgrep'. This has
some very neat features over "normal" grep. For example you can 'grep' in
msgstr or msgid only, or both at the same time and you can choose to ignore
accelerators.

Example: 'gtgrep -f -i scanner -s scanner *'

This command will all strings where the word 'scanner' exists in both msgid
and msgstr - even if there's an accelerator used (e.g. 'sca&nner' will be
found too.)

However gtgrep doesn't support recursive grepping - so you'll have to gtgrep
in every module of your checkout - or do some bash magic.

On Wednesday 04 November 2009 15:16:59 Nicolas Ternisien wrote:
> The page
>
> http://l10n.kde.org/dictionary/search-translations.php
>
> used to do this, but I've disabled the database behind, because it
> consumed too many resources.
>
> Sorry...
>

That's why it doesn't work all the time... :-D
Luckily Lokalize has the same feature :-) (except that Lokalize crashes *all the time* for me)

Greetings

:-)

[marce-5]

2009/11/4 Nicolas Ternisien <nicolas.ternisien@...>:
> The page
>
> http://l10n.kde.org/dictionary/search-translations.php
>
> used to do this, but I've disabled the database behind, because it
> consumed too many resources.
>
> Sorry...

What a pity :'-((, I used to employ it frequently for terminology
work, particularly to check how a particular word or expresion was
translated on other languages (open-tran is not good for this at the
present time).

[Nicolas Ternisien]

>
> What a pity :'-((, I used to employ it frequently for terminology
> work, particularly to check how a particular word or expresion was
> translated on other languages (open-tran is not good for this at the
> present time).
>

True, I will need to retest it someday to see if the server supports
the charge related to the database filling.

Nicolas

Onsdag 4. november 2009 skreiv Martin Schlander:
>It includes among a few other things a neat thing called 'gtgrep'. This has
>some very neat features over "normal" grep. For example you can 'grep' in
>msgstr or msgid only, or both at the same time and you can choose to ignore
>accelerators.

You can do the same (and much more!) with the Pology tools (the
‘find-messages’ sieve), available in KDE’s SVN repository.

(Pology is a suite of PO tools all KDE translators should be very familiar
with.)

--
Karl Ove Hufthammer
http://huftis.org/
Jabber: karl@...

[Orestes Mas-2]

A Dimecres 04 Novembre 2009 21:13:48, Karl Ove Hufthammer va escriure:

> Onsdag 4. november 2009 skreiv Martin Schlander:
> >It includes among a few other things a neat thing called 'gtgrep'. This
> > has some very neat features over "normal" grep. For example you can
> > 'grep' in msgstr or msgid only, or both at the same time and you can
> > choose to ignore accelerators.
>
> You can do the same (and much more!) with the Pology tools (the
> ‘find-messages’ sieve), available in KDE’s SVN repository.
>
> (Pology is a suite of PO tools all KDE translators should be very familiar
> with.)

Maybe, but how are translators supposed to be familiar with these kind of
tools, provided they are very little advertised (or not at all)? All I've
found is a somewhat hidden link in the translation-howto pointing to pology
doc in techbase, which is heavily incomplete. Also, this link present pology
as a tool for XML syntax checking, not a search tool. They are also some e-
mail messages scattered in KDE i18n mailing list about pology, but hard to find
by newcomers.

Similar complaints could be made against the documentation of other
translator's tools. For example, I'm not 100% sure on how to set up and use
the Lokalize "branch" feature, or what's the shortcut to copy a variable (%1,
%2, etc.) from origin (english) to destination (translated). The latter was
simple in KBabel.

Translators are not developers in most cases. They need good documentation to
use necessary tools. Please note I'm not asking for a nice DocBook
documentation. A simple text file would suffice. Without this, translator's work
could be highly inefficient.

Ok, ok. Following the best free sotware tradition, you could ask me to fill the
gaps in the documentation, but I cannot, because I'm not the developer of
these tools, and I cannot write about the answers I'm just looking for.

Orestes.

[Chusslove Illich]

> [: Orestes Mas :]
> [...] pointing to pology doc in techbase, which is heavily incomplete.

It's in the "when I get to it" bin :) However, even when "finalized", Pology
documentation on Techbase should be an overview, presentation of concepts,
short tutorial, etc. The reference documentation is found within Pology
itself, starting from doc/html/index.html (now I've mentioned this in the
Techbase article too), and it's already rather detailed in areas considered
stable. For example, for the posieve script and its find-messages or stats
sieves:

  http://nedohodnik.dyndns.org/pology/doc/html/pology.scripts.posieve-module.html
  http://nedohodnik.dyndns.org/pology/doc/html/pology.sieve.find_messages-module.html
  http://nedohodnik.dyndns.org/pology/doc/html/pology.sieve.stats-module.html

posieve can be queried for the available sieves and their parameters from
command line too:

  $ posieve -h  # global help
  $ posieve -l  # list of available sieves
  $ posieve stats -H  # help for sieve stats

posieve also has Bash shell completion defined (type first few characters of
a sieve name or a parameter, press Tab to complete it), made available by
sourcing completion/bash/pology (mentioned in Pology setup instructions).

--
Chusslove Illich (Часлав Илић)
Serbian KDE translation team

[Nick Shaforostoff-4]

Lokalize has the feature you want. To use it add all your files to
translation memory of your lokalize project, then press F7 to get
Translation Memory search tab. Lokalize ignores accelerators during
the search.

[Panagiotis Papadopoulos]
>Luckily Lokalize has the same feature :-) (except that Lokalize crashes *all the time* for me)

I see that your reports say that you're using custom-compiled trunk
version of KDE.
Do the crashes happen when you use stable (preferably,
distribution-provided) packages?

On Thursday 05 November 2009 22:47:05 Nick Shaforostoff wrote:

> Lokalize has the feature you want. To use it add all your files to
> translation memory of your lokalize project, then press F7 to get
> Translation Memory search tab. Lokalize ignores accelerators during
> the search.
>
> [Panagiotis Papadopoulos]
> >Luckily Lokalize has the same feature :-) (except that Lokalize crashes *all the time* for me)
>
> I see that your reports say that you're using custom-compiled trunk
> version of KDE.
> Do the crashes happen when you use stable (preferably,
> distribution-provided) packages?
>
>

Yeah, same crashes when using the stable package (kdesdk-lokalize 4.3.3) from Arch Linux.
If there is a way I can help, let me know. :-)

[Chusslove Illich]

For example, I was just cleaning up terminology for "boot" (boot, reboot,
boot manager...), and to find lackluster instances I used¹:

  $ posieve find-messages sr/messages/ -saccel:'&' -stransl -smsgid:'boot' -snmsgstr:'поди[зж]' -slokalize

-saccel:'&' specifies the accelerator marker to ignore, -stransl to match
only in translated messages, -smsgid:'boot' -snmsgstr:'поди[зж]' to match
messages having 'boot' in msgid but not having 'подиз' or 'подиж' in msgstr
(i.e. probably wrong terminology used), and -slokalize to pop up all matched
messages in Lokalize (so that they can be edited and saved).

[1] I lie a bit here; I didn't use -saccel because my POs are equipped with
X-Accelerator-Marker header field, which Pology reads automatically; I also
didn't use -slokalize, because I work with Kate.

--
Chusslove Illich (Часлав Илић)
Serbian KDE translation team