Index of Custom Dictionaries for Kobo eReader

[rggdsepp]

hi guys, i'm new here...can someone help me to convert a bgl file to kobo format? it's the english2italian dictionary from babylon... I'm not able to do it

[rmreader]

Hi rggdsepp,

Sorry, I did not see your private message, but I have converted your en-it dictionary.
Please try it out and report back here if it works. Then Alberto can add it to the list of dictionaries for everyone to use.

Kind regards,
Ruud

[rggdsepp]

@rmrreader it works! great, thank you so much

[myshkin]

Hey how's it going everyone!?

First a huge thanks to Alberto for the guide and program to make dictionaries for Kobos! Anyway I have been looking like mad to find a danish danish or danish english one and couldn't find anything so I have used the guide and made one.. only problem is that I couldn't find a big enough source file to change into the Kobo format?

If anyone out there has a Danish dictionary file or a link to one (ideally monolingual but bi would be good as well) can you please let me know?

I have attached the little one I have made (bit basic but works fine!) and have to say it's been a fun process!!

Ciao ciao

[AlPe]

@rmrreader, @myshkin : added to the first post, thanks.

@myshkin : you are welcome. Yes, indeed finding good (freely usable) dictionaries is the main problem, that's why there exist publishers producing commercial dictionaries. Probably we --- as in "readers" --- might consider more contributions to projects like Wiktionary, compiling freely usable dictionaries. For my part, I decided to "donate" some time developing free software, like Penelope.

[AlPe]

In case some of you are interested: yesterday I updated Penelope (now it is at v3.0.0), with a complete code refactoring, new options, MOBI output, and a more rational command line interface. Get it from: https://github.com/pettarin/penelope

EDIT: Penelope v3.0.1 is now available on PyPI (so you can install it with: $ [sudo] pip install penelope) too! https://pypi.python.org/pypi/penelope/

[Simpetus]

Hi,

Just one quick question: Does Kobo Aura H2O has restriction for the size of a dictionary? And if it has, how big dictionary could I use? For example, if I convert 177 mb dict dictionary format into Kobo dictionary format will it work on Kobo?

[Simpetus]

Thanks in advance

[AlPe]

I have no idea. Kobo has not published official documentation, so it is impossible to tell for sure.

What I can say is that I tried ~50MB dictionaries on my Kobo Glo, and they worked.

AFAICT, Kobo software loads only the index file into memory (called "words" --- it uses a very efficient data structure, called MARISA trie; its size only depends on the number of words in your dictionaries, not the size/length of the definitions), while it unpacks only the relevant definition file ??.html when the lookup in the index succeeds for the word selected by the user via the UI.

My suggestion is: build your dictionary and see what happens.

[Simpetus]

Originally Posted by AlPe

I have no idea. Kobo has not published official documentation, so it is impossible to tell for sure.

What I can say is that I tried ~50MB dictionaries on my Kobo Glo, and they worked.

AFAICT, Kobo software loads only the index file into memory (called "words" --- it uses a very efficient data structure, called MARISA trie; its size only depends on the number of words in your dictionaries, not the size/length of the definitions), while it unpacks only the relevant definition file ??.html when the lookup in the index succeeds for the word selected by the user via the UI.

My suggestion is: build your dictionary and see what happens.

Got it. Big thanks for quick reply. I tried to convert dict format to kobo's, the process seems good at the start, but then I come up with that:

C:\Penelope>penelope.py -p zho_rus -f zho -t rus --output-kobo
[INFO] Input dictionary has sequence type 'x'.
[INFO] Starting conversion with the following parameters:

[INFO] Dictionary file: dicthtml-zho-rus
[INFO] Index file:      words
[INFO] Language from:   zho
[INFO] Language to:     rus
[INFO] License:         GNU GPL 3
[INFO] Copyright:       GNU GPL 3
[INFO] Title:           Dictionary zho -> rus
[INFO] Description:     Dictionary zho -> rus
[INFO] Year:            2012

[INFO] Reading input dictionaries...
[INFO] Parsing the input data...
[INFO] Using the built-in parser...
[INFO] Outputting in Kobo format to file...
[INFO] Creating compressed index file words...
Traceback (most recent call last):
  File "C:\Penelope\penelope.py", line 2093, in <module>
    main()
  File "C:\Penelope\penelope.py", line 2069, in main
    write_to_kobo_format(config, parsed_data, debug)
  File "C:\Penelope\penelope.py", line 1218, in write_to_kobo_format
    p = subprocess.Popen([MARISA_BUILD_PATH, "-l", "-o", index_filename], stdout
=subprocess.PIPE, stdin=subprocess.PIPE, stderr=subprocess.PIPE)
  File "C:\Python27\lib\subprocess.py", line 672, in __init__
    errread, errwrite)
  File "C:\Python27\lib\subprocess.py", line 882, in _execute_child
    startupinfo)
WindowsError: [Error 2] The system cannot find the file specified

C:\Penelope>

I am using:

Penelope 1.20
Python 2.7.0

What went wrong here and how to fix it?
Thanks.

[tshering]

Originally Posted by Simpetus

What went wrong here and how to fix it?
Thanks.

Did you edit penelope.py, so that MARISA_BUILD_PATH (and MARISA_REVERSE_LOOKUP_PATH) point to the correct locations?

[Simpetus]

Originally Posted by tshering

Did you edit penelope.py, so that MARISA_BUILD_PATH (and MARISA_REVERSE_LOOKUP_PATH) point to the correct locations?

Thank you! I am messed up path address and wrote tpm instead of tmp in script file! It works now!

[AlPe]

EDIT: I missed the message above. Good. Anyway, updating Penelope will not hurt. v1.20 is a really old version.

=== === ===

Update penelope to the new version, and try again. With the new version, if you install via pip, you do not need to download MARISA separately, it will be installed automatically by pip:

$ pip install penelope

(I do not remember if on Windows you have a GUI for pip or you have pip on the console)

=== === ===

As tshering mentioned, with the previous versions of Penelope you needed to modify the .py file, assigning to the MARISA_BUILD_PATH variable the path of the marisa-build executable, that you needed to download and install separately. As said above, install the new version of Penelope, and MARISA will come automatically.

[oren64]

Hi Alberto

I try to install with Win7 but get an error, this is what i did:

1) extract the file
2)copy pip-Win_1.7, to the folder.
3)run the command "pip install penelope" and run.
4)get an error "error:microsoft visual c++ 9.0 is required".
5)download and install microsoft visual c++ 9.0.
6)run and get an error see pic1-(Python 2.7)
7)download Python 3.5 get error pic2

[AlPe]

EDIT: I did not read the screenshots carefully enough. It looks you have the compiler installed, but compiling lxml fails for some reason.

It seems a very recent problem, see if these posts on StackOverflow help you:

http://stackoverflow.com/questions/3...scrapy-install

or

http://stackoverflow.com/questions/3...85756#33785756

Once you have lxml installed, try pip install penelope again.

=== === ===

Alternatively, you can get Penelope by source: https://github.com/pettarin/penelope#from-source-code If you do not read/write XML files, not having lxml should not prevent Penelope from running. (You might need to pip install marisa-trie separately, if you read/write in Kobo format)