Stardict, Lingea
Dictionaries, language tools and free software

Decoder for Lingea Dictionary (.trd) binary format

Converts into Stardict, MobiPocket (Nokia Symbian cellphones, devices like Psion, Blackberry, Windows Mobile, Palm) and weDict (Apple iPhone)

Use the script only on legally bought dictionaries!

To use this software you must be legal owner of the license for the dictionary!
Unless you are: buy the dictionary!
Read the license! It probably limits the usage to one computer and one user! Produced dictionary has the same copyright as original. You are not owner of the data - you have to follow the license!

You are not allowed to distribute the produced dictionaries. Script is only for personal use. No warranty. GNU GPL License.

Created for devices/platforms not supported by Lingea, like PowerPC Linux, Psion 5MX, Apple iPhone, Nokia cellphones with Symbian, ...

The decoding follows Nomad's specification, my work was done just by guessing and reverse engineering of .trd file, NOT BY CRACKING LINGEA BINARIES. This software was NOT written by Lingea s.r.o. and there is no connection to that company.

Copyright © 2007 Klokan Petr Přidal.

Lingea Dictionary Format

Description of binary format of Lingea dictionaries comes from CobuildConv, written by Nomad in Japanese. English translation by babelfish.
Tags used for records are described in Czech.

Decoder for Lingea TRD

Script for decoding Lingea Dictionary (.trd) file. Result is <header>\t<definition> file, convertible easily by stardict-editor (or by tabfile command from package stardict-tools) into native Stardict dictionary (stardict.sf.net and www.stardict.org).
To run the decoder you need a Python interpreter.

• lingea-trd-decoder.py

Save the file and run:

python lingea-trd-decoder.py lg_XXXX.trd

For now, only 100% supported dictionaries are:

These dictionaries are available in bookstores or you can order them from www.lingea.cz. Czech-English dictionary was also on the DVD of CHIP 10/2007.

Other Lingea dictionaries could follow (byte streams for records are read, so problem is only in decoding of structure for more complicated records). Study the source and Nomad's docs if you would like to help.

Patches for the source code are welcome.

I do NOT provide any user support for the script, it was mine free time activity and I did it to have a nice dictionary on one of unsupported devices (like Apple iPhone, Nokia with Symbian or Psion 5mx). Let me know if you do some changes.

Stardict to Apple iPhone

An application weDict which uses the stardict format directly.

Stardict to Apple Mac OS X built-in Dictionary client

You can convert any stardict dictionary into the format supported by native Apple Dictionary.app. Then the stardict dictionary appear in the Mac OS X 10.5+ in Services menu, in a Dashboard as well as in the standard /Applications/Dictionary.app client.

There is a mac application DictUnifier and sdconv command line utility doing this conversion.

It expects .bz2 archive of the dictionary, therefore you have to create the archive of your Lingea dictionary.
A huge collection of dictionaries is available here, so download and run the DictUnifier is usually enough.

Stardict to MobiPocket OPF (PRC, MOBI)

Script for conversion of Stardict tabfile (<header>\t<definition> per line) into the OPF file for MobiPocket Dictionary

Scripts for conversion:

• tab2opf.py

Produced .opf dictionary must be converted to binary file with extension .mobi (= .prc) by running:

(wine) mobigen.exe DICTIONARY.opf

mobigen.exe is available at: http://www.mobipocket.com/soft/prcgen/mobigen.zip

MobiPocket project supports platforms: PalmOs, Windows Mobile, Symbian (Series 60, Series 80, 90, UIQ), Psion, Blackberry, Franklin, iLiad (by iRex), BenQ-Siemens, Pepper Pad..

Download the Reader

GNU Lexicon

I had an idea to improve existing free software dictionary applications, and create a new version (successor) of DictOSX project, or even merge DictOSX and Stardict, and submit some patches into stardict tree. But because it is a free time (not paid) activity, don't expect miracles soon.

There is a simple proof of concept done in python/GTK called GNU Lexicon, but I think from this project it make sense to use only the idea of better GUI for stardict and usage of regular expressions for conversion of text based dictd/stardict dictionaries into a html form. I planed to include native support for WordNet and for flection, so I did a small research into freely available natural language tools.

Czech and European free natural language tools

• NLTools from Prague
• (ajka is under GPL but not the data) from Brno NLP
• RussianPOSTagger lemmatizing in Russian, English, and German.
• Verbiste - a French conjugation system
• Compjugador is a Spanish verbs conjugator
• Sarf - Arabic Morphology System
• aspell, ispell dictionaries. But we need more - stammers, lematizers or morphology analyzers.

Most used words:

• German, English, French, Dutch, 100, 1000, 10000, ordered by frequency
• Oxford 3000 and other wordlists
• Czech National Corpus frequency analysis. There is also a CD from publication: Čermák, F. - Křen, M. (eds.): Frekvenční slovník češtiny. Nakladatelství Lidové noviny, Praha 2004. ISBN 80-7106-676-1

I was also thinking about embedding some free speech synthesis like Festival or eSpeak, on platform where native synthesis is not supported. BTW There is also an EPOS synthesis and a free czech synthesis based on festival.

It should be done the same way like commercial voices from Infovox iVox do.

Do you know about some freely available tools? Let me know. ;-)