A PO editor

Software
1

Hi!
I’m building a simple PO editor in python and Bethon, as before it runs only in 32 bit Haiku…
I know that Haiku software uses catkeys but I hope it will be useful in translations for other projects.
It’s in an early stage, but can do some basic operations
here some screenshots:
screenshot0
screenshot1
screenshot2

screenshot3

19 Likes
2

Is it already available online somewhere?

3

No, it’s under development, but I think it will be available soon (one or two weeks if everything works)

3 Likes
4

Well, I published the code here https://github.com/tmtfx/HaiPO

It’s a bad piece of software, not optimized, slow and maybe you’ll face some bugs… This is why I consider it in alpha state
Anyhow it seems to do its job.
I’ll add some features in the near future (like handling the header of pofiles)
Notice this:

  • In github repo I placed a zipped file with gettext mimetype to place in /boot/home/config/settings/mime_db/text
  • At first run it will create a user profile which will be saved in config.ini
  • Temporary changes will be saved on a file called normalfile.temp.po
  • Load of “big” files takes some time due to double checks for multiple occurrencies (one day I’ll rewrite the source code) for example a file with 5800 entries takes more or less 7.5 seconds to load on my Intel Icelake 1.2Ghz
  • To run this application you’ll need python, Bethon (which runs only on Haiku-32bit), and polib for python which you can install with pip command

I Hope it will be usefull

Here some more screenshots:
screenshot4

screenshot7

9 Likes
5

Update:
An issue has been fixed and the workaround has been eliminated, i.e. now load-time of a file with 5800 entries takes 0,5 seconds on an older and slower machine (equipped with an Intel Pentium N3700 at 1.6GHz)

next time more speedups

5 Likes
6

screenshot1

Well, some progress…
As there’s no hunspell integration for python 2
I did it the wrong and worst way: launch hunspell as command and decode the output.
As the output of hunspell counts the position as characters and not as byte I had to analyze the string and exclude some double bytes chars (I know that’s not good) so for my language I have to analyze including characters and excluding categories of characters (will be implemented in a configuration section)…
If I find a better solution I’ll do it later. For now it works for me and maybe for other languages… didn’t try…
But at least now I have what I need for my work.
(the spellcheck works only on some calls: typing and moving the list selection with CTRL+DOWN/UP/PGUP/PGDOWN)

1 Like
7

What does PO stands for?

8

Portable Object

9

ok, thanks for clarification

10

I forgot to mention that I’ve updated Bethon integrating BTextView.FindWord function, that was not linked.

1 Like
11

I mainly just want to say, it made me smile, to randomly select a thread here and find that it’s about a Bethon project. With extra functions, which means someone figured out how that horrible apparatus works. Do you know why Bethon isn’t available on 64 bit? I don’t really know who’s supporting it these days - not me, for many years - and I suppose that’s likely the problem right there, but just wondering if there’s some real issue there.

12

Just tried to compile it on x86_64, it runs into typecasting errors where it is assumed that pointers are 32bit. These shouldn’t be too hard to fix but I don’t know if there are any other issues.

And it is still python2, right? We really need Haiku API bindings for python3 if we want to have python as a serious option for app development on Haiku.

1 Like
13

Here I am, again…
Update: with the recent updates I introduced the support for an external Translation Memory.

Translation Memory explanation

For those who don’t know what a Translation Memory is: A translation memory (abbreviation: TM) is a database that stores “segments”, which can be sentences, paragraphs or sentence-like units that have previously been translated, in order to aid human translators. (cit. from wikipedia)

As a lot of modules are missing in Python2 (and in python3 under Haiku) I thought that it would be easier to rely on an external TM server.
So, as far as you respect a simple protocol, you can create your own (and probably faster) TM server. What I provide here is a simple python3 server that uses a TMX file (Translation Memory eXchange) for finding, adding and deleting your strings.
The TMXs files are easily generated with python3 translate-tools (po2tmx command); surely a SQL database is faster…
By requesting a string from the server you’ll get the 5 most similar translations, which can copied to the translation field by pressing ctrl+1+2+3+4+5 shorcuts (or by a double click).
By Pressing Canc/Del after selecting one suggestion, you’ll erase it from the TM server.

suggestions
suggestions871×675 78.1 KB

Other features:

  • I’ve added a Launcher to use in FileTypes, so when a .po file receives a double-click it will be opened with HaiPO
  • I’ve update the x-gettext-translation filetype (you can install it with the install.sh script)

The journey:
Since I started this project I realize now that it starts to be more comfortable for my translation work. There’s still a lot of things to tune up, but I see that Haiku now is my preferred OS to get my (GNOME and TranslationProject) translations done.
A big Thank You to all Haiku Devs and to Haiku itself :slight_smile:

9 Likes