How to create your own mobipocket dictionary for any language:

drago · 02-12-2008, 06:17 AM

I was very gratefull to Adam B for porting the Stardict dictonary tool for the Iliad, but I allways longed for the kind of instant lookup, which is possible from mobipocket. Unfortunately mobipocket dictionaries are not available for more exotic languages, so I allways wanted to create my own mobipocket dictionaries. It seems I have now found a solution:

First you need to get hold of a tab delimited dictionary file. (example. dictionary.txt)

(Stardict dictionary files are available from: http://stardict.sourceforge.net/Dictionaries.php
These can be converted into a tab delimited file by the stardict editor programe, which you can find here: http://stardict.sourceforge.net/other.php )

Once you have such a file, use tab2opf.py from http://www.klokan.cz/projects/stardict-lingea/
This is a free python script by Petr Klokan, so you’ll need python isntalled.

Type tab2opf.py dictionary.txt into the command line to run the script.

This will give you a dictionary.opf file, and a number of dictionary0.html files

Edit the dictionary.opf file, to specify the name of the dictionary, and the impot output languages.

Get mobigen.exe form - http://www.mobipocket.com/soft/prcgen/mobigen.zip

Extract mobigen exe

Run mobigen.exe dictionary.opf

If you did everything right, you’ll get dictionary.mobi which’ll be the desired mobipocket dictionary.

Cheers!

D.

templario · 02-12-2008, 06:33 AM

That´s very interesting and helpful. Thanks for sharing. Just one question, can this be done in a windows machine?

Thanks

drago · 02-12-2008, 06:44 AM

Templario,

yes, I have done it on a win machine. Stardict editor has a win version, python can be instaled for windows, and mobigen. exe is windows only.

Btw. it seems mobigen.exe runs under wine as well, so this can also be done on a Linux box.

Cheers!

D.

ericshliao · 02-12-2008, 06:57 AM

Mobipocket Reader on iLiad still have problem displaying non-western glyphs, such as Chinese glyphs. I think Japanese and Korean glyphs are no exception.

Edited:
With proper modification, CJK glyphs can be displayed on Mobipocket Reader. See this:
http://forum.irexnet.com/viewtopic.php?t=2281

hellosim · 02-13-2008, 02:48 AM

I followed your step in winXP.

installed stardict3.0, installed GTK2 Runtime, downloaded dictionary files, extracted with 7zip, run stardict editor.

after I select *.dict.dz file, It verifies. BUT If I select decompile, it says

Building...
Error, file version is not 2.4.2
Done!

What does it mean?

drago · 02-13-2008, 06:02 AM

This happened to me as well. What i did, was I extracted the .dict file from the .dz (i don't know what kind of encryption this is, but ctrl-pgdown in total commander dealt with it.) In stardict editor under decompile you need to select the ifo file. Sometimes it refused to decompile giving me the error message sametypesequence not m, but after changing the sametypesequence=x line in the ifo file to sametypesequence=m it decompilled.

Good luck!

D

ericshliao · 02-15-2008, 01:03 AM

The .dz file is actually a .gz format archive. You should rename the .dz to .gz, then unpack it. There should be a .dict file in it. The .dict file is the uncompressed dictionary file for decompiling.

drago · 02-16-2008, 03:07 AM

Update: creating the dictionaries with Mobipocket Creator Publisher Edition gave better and more consistent results than using the command line mobigen.exe.
Get it from:
http://www.mobipocket.com/en/Downloa...tion=Publisher

Create new publication
----create presonal dictionary-glosary
important: specify the language, and select the input-output languages at the book-setting tabs, without this, instant lookup won't work. (Also I got the best results with unicode encoding.)
drop all the numbered .html files created by the tab2opf.py (leave they .opf file alone)
click on build

Enjoy!

D.

wagnerian · 02-25-2008, 01:34 AM

I have built a yet another tab2mobi script because the tab2opf.py does not support Asian characters.

Install BeeLiner first:
http://wing4bee.tistory.com

And put this script package and mobigen.exe into the directory where BeeLiner is installed.
Modify the MyDict.opf file and run:
tab2mobi.cmd your_source_file.txt

ericshliao · 02-25-2008, 08:13 AM

By my experience, with a "-utf" option, the original tab2opf.py can deal with Chinese characters.

wagnerian · 02-25-2008, 01:02 PM

Quote:

Originally Posted by ericshliao

By my experience, with a "-utf" option, the original tab2opf.py can deal with Chinese characters.

Oops! It does work!

PointBoom · 04-16-2008, 05:15 PM

Thanks for this great post! It allowed me to find and get a dictionary running very quickly.

One query though:
I created a French-English Dictionary. And it works great except when the word is not found and a 'manual search' option is offered. Clicking on that throws an error. For example:

Could not open: c:\path to dictionary\french-english-dict.prc(window.index_search('','','notamment'))

I found that by deleting the idx:key tags and recompiling things that the manual search started working nicely.

My question is: What was hoped to be accomplished with the idx:key tags?

I noticed that (some of/all of?) the accents had been removed in the idx:key tags. If this was meant to allow searching without accents maybe this could be done with the 'inflections' tagging? Any ideas? I'm wondering if the script needs modification (or if I buggered something up making the change).

moontaxi · 04-20-2008, 11:58 AM

drago, wagnerian, ericshliao, pointboom

can any of you, who succeed this work, convert babylon files using dictconv and make a mobipocket dictionary for me? i'm running windows xp and linux is the only system that i can convert babylon files.

this is the link for dictconv (to convert babylon to stardict format)
http://linux.softpedia.com/get/Educa...nv-23446.shtml

i uploaded 5 babylon dictionaries here: (3,64mb)
http://rapidshare.com/files/108999822/Babylon.rar.html

i'll be glad.

wagnerian · 04-25-2008, 03:14 AM

moontaxi, it's extremely difficult to me to figure out the match-the-cat-pictures task of the RapidShare.
I've tried and failed three times and finally given up.

artbatista · 05-12-2008, 09:31 PM

It's silly, but it makes me feel better to know I am not the only one that can not decode the damned cats..

Art

Quote:

Originally Posted by wagnerian

moontaxi, it's extremely difficult to me to figure out the match-the-cat-pictures task of the RapidShare.
I've tried and failed three times and finally given up.

02-12-2008, 06:17 AM	#1
drago Enthusiast Posts: 28 Karma: 10 Join Date: Jul 2006 Location: Budapest, Hungary Device: KPW, BrokenfirstgeniLiad	How to create your own mobipocket dictionary for any language: I was very gratefull to Adam B for porting the Stardict dictonary tool for the Iliad, but I allways longed for the kind of instant lookup, which is possible from mobipocket. Unfortunately mobipocket dictionaries are not available for more exotic languages, so I allways wanted to create my own mobipocket dictionaries. It seems I have now found a solution: First you need to get hold of a tab delimited dictionary file. (example. dictionary.txt) (Stardict dictionary files are available from: http://stardict.sourceforge.net/Dictionaries.php These can be converted into a tab delimited file by the stardict editor programe, which you can find here: http://stardict.sourceforge.net/other.php ) Once you have such a file, use tab2opf.py from http://www.klokan.cz/projects/stardict-lingea/ This is a free python script by Petr Klokan, so you’ll need python isntalled. Type tab2opf.py dictionary.txt into the command line to run the script. This will give you a dictionary.opf file, and a number of dictionary0.html files Edit the dictionary.opf file, to specify the name of the dictionary, and the impot output languages. Get mobigen.exe form - http://www.mobipocket.com/soft/prcgen/mobigen.zip Extract mobigen exe Run mobigen.exe dictionary.opf If you did everything right, you’ll get dictionary.mobi which’ll be the desired mobipocket dictionary. Cheers! D.

02-12-2008, 06:57 AM	#4
ericshliao Guru Posts: 976 Karma: 687 Join Date: Nov 2007 Device: Dell X51v; iLiad v2	Mobipocket Reader on iLiad still have problem displaying non-western glyphs, such as Chinese glyphs. I think Japanese and Korean glyphs are no exception. Edited: With proper modification, CJK glyphs can be displayed on Mobipocket Reader. See this: http://forum.irexnet.com/viewtopic.php?t=2281 Last edited by ericshliao; 02-15-2008 at 12:46 PM.

02-16-2008, 03:07 AM	#8
drago Enthusiast Posts: 28 Karma: 10 Join Date: Jul 2006 Location: Budapest, Hungary Device: KPW, BrokenfirstgeniLiad	Update: creating the dictionaries with Mobipocket Creator Publisher Edition gave better and more consistent results than using the command line mobigen.exe. Get it from: http://www.mobipocket.com/en/Downloa...tion=Publisher Create new publication ----create presonal dictionary-glosary important: specify the language, and select the input-output languages at the book-setting tabs, without this, instant lookup won't work. (Also I got the best results with unicode encoding.) drop all the numbered .html files created by the tab2opf.py (leave they .opf file alone) click on build Enjoy! D. Last edited by drago; 02-16-2008 at 03:08 AM. Reason: speling

04-16-2008, 05:15 PM	#12
PointBoom Junior Member Posts: 1 Karma: 10 Join Date: Apr 2008 Device: palm	Thanks for this great post! It allowed me to find and get a dictionary running very quickly. One query though: I created a French-English Dictionary. And it works great except when the word is not found and a 'manual search' option is offered. Clicking on that throws an error. For example: Could not open: c:\path to dictionary\french-english-dict.prc(window.index_search('','','notamment')) I found that by deleting the idx:key tags and recompiling things that the manual search started working nicely. My question is: What was hoped to be accomplished with the idx:key tags? I noticed that (some of/all of?) the accents had been removed in the idx:key tags. If this was meant to allow searching without accents maybe this could be done with the 'inflections' tagging? Any ideas? I'm wondering if the script needs modification (or if I buggered something up making the change). Last edited by PointBoom; 04-16-2008 at 05:20 PM. Reason: trying to make the question clearer.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Dictionary lookup in iBooks 1.1: "Dictionary not available for this language"	kjk	Apple Devices	71	09-18-2010 07:24 AM
best foreign language & dictionary options?	joedevivre	Which one should I buy?	2	12-13-2009 10:40 AM
How do I create headword-enabled Mobipocket dictionary?	owl123	Kindle Formats	1	07-24-2009 12:13 PM
Useful tip: How to change the BD language AFTER you create a book	HarryT	Workshop	4	04-15-2009 01:36 AM
creating a foreign language dictionary	dirtylc	Amazon Kindle	1	03-30-2009 09:40 AM

02-12-2008, 06:33 AM	#2
templario Connoisseur Posts: 86 Karma: 10 Join Date: Dec 2006 Device: Iliad	That´s very interesting and helpful. Thanks for sharing. Just one question, can this be done in a windows machine? Thanks

02-12-2008, 06:44 AM	#3
drago Enthusiast Posts: 28 Karma: 10 Join Date: Jul 2006 Location: Budapest, Hungary Device: KPW, BrokenfirstgeniLiad	Templario, yes, I have done it on a win machine. Stardict editor has a win version, python can be instaled for windows, and mobigen. exe is windows only. Btw. it seems mobigen.exe runs under wine as well, so this can also be done on a Linux box. Cheers! D.

02-13-2008, 02:48 AM	#5
hellosim Junior Member Posts: 1 Karma: 10 Join Date: Feb 2008 Device: hiebook	I followed your step in winXP. installed stardict3.0, installed GTK2 Runtime, downloaded dictionary files, extracted with 7zip, run stardict editor. after I select *.dict.dz file, It verifies. BUT If I select decompile, it says Building... Error, file version is not 2.4.2 Done! What does it mean?

02-13-2008, 06:02 AM	#6
drago Enthusiast Posts: 28 Karma: 10 Join Date: Jul 2006 Location: Budapest, Hungary Device: KPW, BrokenfirstgeniLiad	This happened to me as well. What i did, was I extracted the .dict file from the .dz (i don't know what kind of encryption this is, but ctrl-pgdown in total commander dealt with it.) In stardict editor under decompile you need to select the ifo file. Sometimes it refused to decompile giving me the error message sametypesequence not m, but after changing the sametypesequence=x line in the ifo file to sametypesequence=m it decompilled. Good luck! D

02-15-2008, 01:03 AM	#7
ericshliao Guru Posts: 976 Karma: 687 Join Date: Nov 2007 Device: Dell X51v; iLiad v2	The .dz file is actually a .gz format archive. You should rename the .dz to .gz, then unpack it. There should be a .dict file in it. The .dict file is the uncompressed dictionary file for decompiling.

02-25-2008, 08:13 AM	#10
ericshliao Guru Posts: 976 Karma: 687 Join Date: Nov 2007 Device: Dell X51v; iLiad v2	By my experience, with a "-utf" option, the original tab2opf.py can deal with Chinese characters.

04-20-2008, 11:58 AM	#13
moontaxi Junior Member Posts: 1 Karma: 10 Join Date: Apr 2008 Device: ipaq 6340	drago, wagnerian, ericshliao, pointboom can any of you, who succeed this work, convert babylon files using dictconv and make a mobipocket dictionary for me? i'm running windows xp and linux is the only system that i can convert babylon files. this is the link for dictconv (to convert babylon to stardict format) http://linux.softpedia.com/get/Educa...nv-23446.shtml i uploaded 5 babylon dictionaries here: (3,64mb) http://rapidshare.com/files/108999822/Babylon.rar.html i'll be glad.

04-25-2008, 03:14 AM	#14
wagnerian Musicology Student Posts: 55 Karma: 45 Join Date: Oct 2006 Location: Seoul, Korea (R.O.K.) Device: iLiad, BlackBerry Bold 9000, Zaurus C1000	moontaxi, it's extremely difficult to me to figure out the match-the-cat-pictures task of the RapidShare. I've tried and failed three times and finally given up.

Advert

Advert