Hi Mikel:
Thank you for your feedback Mikel! I've fixed some of the issues that
you have found (the new version is already uploaded):
Cool!
Cool. Works like a charm. A minor quibble: I said I wanted it
installed in /tmp and it decided that every file in there was an
already installed language pair. Perhaps a regex identifying
language pairs by name would be very helpful.
I now look at the file extension to filter JAR files. I think that it
should be enough. It would also be possible to look at the whole
filename, but this would make manually installed pairs unusable unless
they are properly named...
The standard style is apertium-[a-z][a-z][a-z]?-[a-z][a-z][a-z]?.jar I
think it would be no problem for your program to require that.
_*Apertium plug-in for OmegaT*_
The menu appearde but this one crashed on me: I think the problem
is that I had a previous installation somewhere else (they use the
same .java/.userPrefs/org/apertium/ . I remove these.
They now use independent preferences, so we shouldn't have that
problem anymore.
Then on launching OmegaT I get a dialog that may be confusing for
some, as it does not identify itself as an Apertium warning. It
should...
Now it does!
I'll take a look. Thanks a lot!
It kindly asks me to decide where to install itself, and where to
install language pairs. Then I open a project, mark "Apertium
offline" in the Machine Translation options, and it crashes.
java.lang.IndexOutOfBoundsException: Index: 13, Size: 7
at java.util.ArrayList.rangeCheck(ArrayList.java:571)
at java.util.ArrayList.remove(ArrayList.java:412)
at
org.apertium.lttoolbox.process.State.nodeStatePool_get(State.java:101)
at org.apertium.lttoolbox.process.State.copy(State.java:145)
at org.apertium.lttoolbox.process.State.copy(State.java:158)
at
org.apertium.lttoolbox.process.FSTProcessor.analysis(FSTProcessor.java:857)
at org.apertium.lttoolbox.LTProc.doMain(LTProc.java:287)
at org.apertium.pipeline.Dispatcher.doLTProc(Dispatcher.java:259)
at org.apertium.pipeline.Dispatcher.dispatch(Dispatcher.java:333)
at org.apertium.Translator.translate(Translator.java:305)
at
org.omegat.plugin.machinetranslators.ApertiumTranslate.translate(ApertiumTranslate.java:62)
at
org.omegat.core.machinetranslators.BaseTranslate.getTranslation(BaseTranslate.java:64)
at
org.omegat.gui.exttrans.MachineTranslateTextArea$FindThread.search(MachineTranslateTextArea.java:128)
at
org.omegat.gui.exttrans.MachineTranslateTextArea$FindThread.search(MachineTranslateTextArea.java:103)
at
org.omegat.gui.common.EntryInfoSearchThread.run(EntryInfoSearchThread.java:95)
I'm not sure if I have correctly identified the problem, but it is
probably fixed now.
I will test it again, to see what happens.
There is one thing that could be easily solved. Víctor Sánchez
(cc-ed) maybe can help you. When one uses the Apertium webservice
from inside OmegaT, we avoid translating the tags (<u0>, etc.).
Some minor changes were made to the code that calls Apertium as a
webservice (you'll easily find them, but if not, I can help) and
some changes were made in the webservice itself (Víctor can help
here). I think it is a matter of using some regular expressions to
hide these in some way...
I guess that you are talking about this
<http://omegat.svn.sourceforge.net/viewvc/omegat/trunk/src/org/omegat/core/machinetranslators/ApertiumTranslate.java?revision=4434&view=markup>.
I might be blind, but I haven't been able to identify the relevant
piece of code there...
You're right. Most of the work is done at the Apertium server when it
receives format=omegat.
The code that carries out the translation in the plug-in can be found
here
<http://apertium.svn.sourceforge.net/viewvc/apertium/branches/gsoc2012/artetxem/apertium-omegat/src/org/omegat/plugin/machinetranslators/ApertiumTranslate.java?revision=39499&view=markup>
from line 56 to 66. The function is really simple, so it should be
easy to apply the necessary changes there (if we know what this
necessary changes are!).
I'm sure Víctor can tell you. It's a matter of catching and escaping the
<> tags generated by OmegaT so that their name does not get translated.
You should consider sending a message to the OmegaT list. I can
help here, as I am subscribed. I think this is very, very
important! And perhaps you can get OmegaT people to help with some
of the issues.
I will do it then!
Excellent.
It is important to note that these settings as well as the
installed language pairs are shared with Apertium Caffeine, which
means that, if you install, uninstall or update a language pair,
the changes will be reflected in both programs. This is a design
decision that I took, but it would be simple to make them
independent if you prefer it.
Oops I hadn't read that. I think it would be nice to have separate
options, yes.
Now they are completely independent. But this could be problematic in
some special cases: if the same directory is chosen for both programs,
they would conflict and we would probably get a strange behaviour...
You could, perhaps, add some kind of file that says which (OmegaT or
Caffeine) is using the directory, and announce the conflict when the
other one tries to use it too.
Please, note that both are still under development, so you might
find that some stuff is not working as it should. Most notably,
the maintenance of online language pair packages is being
discussed in another thread, and we haven't taken a definitive
solution yet, so both apps are currently using some testing
language pairs that I keep in my branch.
Nice. We should solve this.
Yes. We should probably create a new directory in SVN and start
creating and uploading packages for every language pair. The question
is how to maintain it in long-term: we could integrate my script in
the makefiles of each language pair to make things easier (although
the dependency of Android-SDK and lttoolbox-java can still be a
problem for some people), but we would still need the implication of
every language pair developer in Apertium (or some responsible to take
care of the whole maintenance).
This deserves a deeper thought. Any ideas?
Mikel
--
Mikel L. Forcada (http://www.dlsi.ua.es/~mlf/)
Departament de Llenguatges i Sistemes Informàtics
Universitat d'Alacant
E-03071 Alacant, Spain
Phone: +34 96 590 9776
Fax: +34 96 590 9326
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Apertium-stuff mailing list
Apertium-stuff@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/apertium-stuff
