Penelope

Penelope is a multi-tool for creating, editing and converting dictionaries, especially for eReader devices.

• Version: 3.1.3
• Date: 2016-09-23
• Developer: Alberto Pettarin
• License: the MIT License (MIT)
• Contact: click here

With the current version you can:

• convert a dictionary from/to the following formats:
  • Bookeen Cybook Odyssey (R/W)
  • CSV (R/W)
  • EPUB (W only)
  • MOBI (Kindle, W only)
  • Kobo (R index only, W unencrypted/unobfuscated only)
  • StarDict (R/W)
  • XML (R/W)
• merge several dictionaries of the same type into a single dictionary
• merge several definitions for the same headword
• sort by headword and/or by definition
• define your own input parser to merge/sort/edit definitions
• define your own collation function (bookeen output format only)
• output an EPUB file containing the dictionary (e.g., to cope with the lack of a search function of your eReader)
• output a MOBI (Kindle) dictionary

Important Update

• 2016-04-17 Sadly, I can no longer afford to spend time working on Penelope, as my other FLOSS projects take 100% of my FLOSS time, and I still need to pay rent and bills, spend time with family and friends, etc., as anyone else. Therefore, I will not work on issues or pull requests, please do not expect they will be dealt with at all. I am actively looking for other developers to take this project over. (This notice should be removed when the changeover happens.) If you need to convert a dictionary and the current version of Penelope does not work for you, you might want to have a look at PyGlossary. My most sincere apologies for the inconvenience.

Installation

Using pip

1. Open a console and type:

   $ [sudo] pip install penelope

2. That's it! Just run without arguments (or with -h or --help) to get the manual:

   $ penelope

This procedure will install lxml and marisa-trie. You might need to install dictzip (StarDict output) and kindlegen (MOBI output) separately, see below.

From source code

1. Get the source code:

   • clone this repo with git:

     $ git clone https://github.com/pettarin/penelope.git

   • or download the latest release and uncompress it somewhere,

   • or download the current master ZIP and uncompress it somewhere.

2. Open a console and enter the penelope (cloned) directory:

   $ cd /path/to/penelope

3. That's it! Just run without arguments (or with -h or --help) to get the manual:

   $ python -m penelope

This procedure will not install any dependencies: you will need to do that manually, see below.

Dependencies

• Python, version 2.7.x or 3.4.x (or above)

• to write StarDict dictionaries: the dictzip executable, available in your $PATH or specified with --dictzip-path:

  $ [sudo] apt-get install dictzip

• to read/write Kobo dictionaries: the Python module marisa-trie:

  $ [sudo] pip install marisa-trie

  or MARISA executables available in your $PATH or specified with --marisa-bin-path

• to write MOBI Kindle dictionaries: the kindlegen executable, available in your $PATH or specified with --kindlegen-path

• to read/write XML dictionaries: the Python module lxml:

  $ [sudo] pip install lxml

Usage

usage: 
  $ penelope -h
  $ penelope -i INPUT_FILE -j INPUT_FORMAT -f LANGUAGE_FROM -t LANGUAGE_TO -p OUTPUT_FORMAT -o OUTPUT_FILE [OPTIONS]
  $ penelope -i IN1,IN2[,IN3...] -j INPUT_FORMAT -f LANGUAGE_FROM -t LANGUAGE_TO -p OUTPUT_FORMAT -o OUTPUT_FILE [OPTIONS]

description:
  Convert dictionary file(s) with file name prefix INPUT_FILE from format INPUT_FORMAT to format OUTPUT_FORMAT, saving it as OUTPUT_FILE.
  The dictionary is from LANGUAGE_FROM to LANGUAGE_TO, possibly the same.
  You can merge several dictionaries (with the same format), by providing a list of comma-separated prefixes, as shown by the third synopsis above.

optional arguments:
  -h, --help            show this help message and exit
  -d, --debug           enable debug mode (default: False)
  -f LANGUAGE_FROM, --language-from LANGUAGE_FROM
                        from language (ISO 639-1 code)
  -i INPUT_FILE, --input-file INPUT_FILE
                        input file name prefix(es). Multiple prefixes must be
                        comma-separated.
  -j INPUT_FORMAT, --input-format INPUT_FORMAT
                        from format (values: bookeen|csv|kobo|stardict|xml)
  -k, --keep            keep temporary files (default: False)
  -o OUTPUT_FILE, --output-file OUTPUT_FILE
                        output file name
  -p OUTPUT_FORMAT, --output-format OUTPUT_FORMAT
                        to format (values:
                        bookeen|csv|epub|kobo|mobi|stardict|xml)
  -t LANGUAGE_TO, --language-to LANGUAGE_TO
                        to language (ISO 639-1 code)
  -v, --version         print version and exit
  --author AUTHOR       author string
  --copyright COPYRIGHT
                        copyright string
  --cover-path COVER_PATH
                        path of the cover image file
  --description DESCRIPTION
                        description string
  --email EMAIL         email string
  --identifier IDENTIFIER
                        identifier string
  --license LICENSE     license string
  --title TITLE         title string
  --website WEBSITE     website string
  --year YEAR           year string
  --apply-css APPLY_CSS
                        apply the given CSS file (epub and mobi output only)
  --bookeen-collation-function BOOKEEN_COLLATION_FUNCTION
                        use the specified collation function
  --bookeen-install-file
                        create *.install file (default: False)
  --csv-fs CSV_FS       CSV field separator (default: ',')
  --csv-ignore-first-line
                        ignore the first line of the input CSV file(s)
                        (default: False)
  --csv-ls CSV_LS       CSV line separator (default: '\n')
  --dictzip-path DICTZIP_PATH
                        path to dictzip executable
  --epub-no-compress    do not create the compressed container (epub output
                        only, default: False)
  --escape-strings      escape HTML strings (default: False)
  --flatten-synonyms    flatten synonyms, creating a new entry with
                        headword=synonym and using the definition of the
                        original headword (default: False)
  --group-by-prefix-function GROUP_BY_PREFIX_FUNCTION
                        compute the prefix of headwords using the given prefix
                        function file
  --group-by-prefix-length GROUP_BY_PREFIX_LENGTH
                        group headwords by prefix of given length (default: 2)
  --group-by-prefix-merge-across-first
                        merge headword groups even when the first character
                        changes (default: False)
  --group-by-prefix-merge-min-size GROUP_BY_PREFIX_MERGE_MIN_SIZE
                        merge headword groups until the given minimum number
                        of headwords is reached (default: 0, meaning no merge
                        will take place)
  --ignore-case         ignore headword case, all headwords will be lowercased
                        (default: False)
  --ignore-synonyms     ignore synonyms, not reading/writing them if present
                        (default: False)
  --include-index-page  include an index page (epub and mobi output only,
                        default: False)
  --input-file-encoding INPUT_FILE_ENCODING
                        use the specified encoding for reading the raw
                        contents of input file(s) (default: 'utf-8')
  --input-parser INPUT_PARSER
                        use the specified parser function after reading the
                        raw contents of input file(s)
  --kindlegen-path KINDLEGEN_PATH
                        path to kindlegen executable
  --marisa-bin-path MARISA_BIN_PATH
                        path to MARISA bin directory
  --marisa-index-size MARISA_INDEX_SIZE
                        maximum size of the MARISA index (default: 1000000)
  --merge-definitions   merge definitions for the same headword (default:
                        False)
  --merge-separator MERGE_SEPARATOR
                        add this string between merged definitions (default: '
                        | ')
  --mobi-no-kindlegen   do not run kindlegen, keep .opf and .html files
                        (default: False)
  --no-definitions      do not output definitions for EPUB and MOBI formats
                        (default: False)
  --sd-ignore-sametypesequence
                        ignore the value of sametypesequence in StarDict .ifo
                        files (default: False)
  --sd-no-dictzip       do not compress the .dict file in StarDict files
                        (default: False)
  --sort-after          sort after merging/flattening (default: False)
  --sort-before         sort before merging/flattening (default: False)
  --sort-by-definition  sort by definition (default: False)
  --sort-by-headword    sort by headword (default: False)
  --sort-ignore-case    ignore case when sorting (default: False)
  --sort-reverse        reverse the sort order (default: False)

examples:

  $ penelope -i dict.csv -j csv -f en -t it -p stardict -o output.zip
    Convert en->it dictionary dict.csv (in CSV format) into output.zip (in StarDict format)

  $ penelope -i dict.csv -j csv -f en -t it -p stardict -o output.zip --merge-definitions
    As above, but also merge definitions

  $ penelope -i d1,d2,d3 -j csv -f en -t it -p csv -o output.csv --sort-after --sort-by-headword
    Merge CSV dictionaries d1, d2, and d3 into output.csv, sorting by headword

  $ penelope -i d1,d2,d3 -j csv -f en -t it -p csv -o output.csv --sort-after --sort-by-headword --sort-ignore-case
    As above, but ignore case for sorting

  $ penelope -i d1,d2,d3 -j csv -f en -t it -p csv -o output.csv --sort-after --sort-by-headword --sort-reverse
    As above, but reverse the order

  $ penelope -i dict.zip -j stardict -f en -t it -p csv -o output.csv
    Convert en->it dictionary dict.zip (in StarDict format) into output.csv (in CSV format)

  $ penelope -i dict.zip -j stardict -f en -t it -p csv -o output.csv --ignore-synonyms
    As above, but do not read the .syn synonym file if present

  $ penelope -i dict.zip -j stardict -f en -t it -p csv -o output.csv --flatten-synonyms
    As above, but flatten synonyms

  $ penelope -i dict.zip -j stardict -f en -t it -p bookeen -o output
    Convert dict.zip into output.dict.idx and output.dict for Bookeen devices

  $ penelope -i dict.zip -j stardict -f en -t it -p kobo -o dicthtml-en-it
    Convert dict.zip into dicthtml-en-it.zip for Kobo devices

  $ penelope -i dict.csv -j csv -f en -t it -p mobi -o output.mobi --cover-path mycover.png --title "My English->Italian Dictionary"
    Convert dict.csv into a MOBI (Kindle) dictionary, using the specified cover image and title

  $ penelope -i dict.xml -j xml -f en -t it -p mobi -o output.epub
    Convert dict.xml into an EPUB dictionary

  $ penelope -i dict.xml -j xml -f en -t it -p mobi -o output.epub --epub-output-definitions
    As above, but also output definitions

You can find ISO 639-1 language codes here.

Installing the Dictionaries

Bookeen Odyssey Devices

For example, suppose you want to use an IT -> EN dictionary.

1. On your PC, produce/download the IT -> EN dictionary files it-en.dict and it-en.dict.idx.
2. Connect your Odyssey device to your PC via the USB cable.
3. Using your file manager, copy the two files it-en.dict and it-en.dict.idx from your PC into the Dictionaries/ directory on your Odyssey device.
4. Reboot your Odyssey, open a book in Italian and select a word: the definition in English should appear. (For this test, select a common word so you are sure it is present in the dictionary!)

Note that the Bookeen dictionary software will select the dictionary to use by reading the dc:language metadata of your eBook. Make sure your eBooks have the proper dc:language metadata, otherwise the correct dictionary might not be loaded.

Kobo Devices

At the time of this writing (2016-02-16), Kobo devices will load dictionaries only if the files have a file name of an official Kobo dictionaries, which are:

• dicthtml.zip (EN)
• dicthtml-de.zip (DE), dicthtml-de-en.zip (DE -> EN), dicthtml-en-de.zip (EN -> DE),
• dicthtml-es.zip (ES), dicthtml-es-en.zip (ES -> EN), dicthtml-en-es.zip (EN -> ES),
• dicthtml-fr.zip (FR), dicthtml-fr-en.zip (FR -> EN), dicthtml-en-fr.zip (EN -> FR),
• dicthtml-it.zip (IT), dicthtml-it-en.zip (IT -> EN), dicthtml-en-it.zip (EN -> IT),
• dicthtml-nl.zip (NL)
• dicthtml-ja.zip (JA), dicthtml-en-ja.zip (EN -> JA),
• dicthtml-pt.zip (PT), dicthtml-pt-en.zip (PT -> EN), dicthtml-en-pt.zip (EN -> PT)

(see this MobileRead thread)

Hence, if you want to install a custom dictionary produced with Penelope, you must choose to overwrite one of the official Kobo dictionaries, effectively loosing the possibility of using the latter.

For example, suppose you want to use a Polish dictionary (dicthtml-pl.zip), while you are not interested in using the official Portuguese one (dicthtml-pt.zip).

1. On your PC, produce/download the Polish dictionary dicthtml-pl.zip.
2. In your Kobo device, go to the settings and activate the Portuguese dictionary.
3. Connect your Kobo device to your PC via the USB cable.
4. Using your file manager, copy dicthtml-pl.zip from your PC into the .kobo/dict/ directory on your Kobo device. (Note that .kobo is a hidden directory: you might need to enable the "show hidden files/directories" setting of your file manager.)
5. Rename dicthtml-pl.zip into dicthtml-pt.zip.
6. Reboot your Kobo, open a book in Polish and select a word: the definition should appear. (For this test, select a common word so you are sure it is present in the dictionary!)

Note that if you update the firmware of your Kobo, the custom dictionaries might be overwritten with the official ones. Hence, keep a backup copy of your custom dictionaries in a safe place, e.g. your PC or a SD card.

You can find a list of custom dictionaries, mostly done with Penelope, in this MobileRead thread.

License

Penelope is released under the MIT License since version 2.0.0 (2014-06-30).

Previous versions, hosted by Google Code, were released under the GNU GPL 3 License.

Limitations and Missing Features

• Bookeen has no official documentation for its dictionary format (it has been reverse-engineered), YMMV
• Kobo has no official documentation for its dictionary format (it has been reverse-engineered), YMMV
• Reading Kobo dictionaries is partially supported (the index is read, the definitions are not, as they are encrypted/obfuscated)
• Reading EPUB (3) dictionaries is not supported; the writing part needs polishing/refactoring
• Reading PRC/MOBI (Kindle) dictionaries is not supported
• There are some limitations on StarDict files that can be read (see comments in format_stardict.py)
• Documentation is not complete
• Unit tests are missing

Sponsors

• December 2015: IngleseXpress.it, "Grazie per averci aiutato a pubblicare per Kindle il Dizionario Inglese-Italiano della Pronuncia Scritta Semplificata!"

Acknowledgments

Many thanks to:

• uwelovesdonna for contributing ideas for improving the code and for setting up many pages of the project wiki;
• Jens Sadowski for pointing out a bug with Unicode file names and for suggesting using multiset dict() instead of set dict();
• oldnat for pointing out a bug under Windows and Python 3;
• Wolfgang Miller-Reichling for providing the code for reading CSV dictionaries;
• branok for providing the idea and initial code for German collation function;
• pal for suggesting passing -l switch to MARISA_BUILD;
• Lukas Brückner for suggesting escaping & < > when outputting in XML format;
• Stephan Lichtenhagen for suggesting forcing UTF-8 encoding on Python 3;
• niconavarrete for pointing out the dependency from $CWD (issue #1), solved in v2.0.1;
• elchamaco for providing a StarDict dictionary with a .syn file for testing.