mobidick

Dictionary generator for Kindle

Requirements

(All these instructions are given for Arch Linux.)

dictd

yaourt -S dictd [dict-freedict-fra-eng]
sudo systemctl start dictd

aspell

sudo pacman -S aspell [aspell-fr]

python

sudo pacman -S python virtualenv

kindlegen

Download kindlegen from Kindle web page and configure its path in your configuration file (see below):

virtualenv

sudo pacman -S virtualenv
cd <mobidick_path>
virtualenv venv
source venv/bin/activate
pip install -r requirements.txt

Configuration

Create a configuration file for your dictionary:

{
    "title": "French-English Dictionary",
    "output": "/tmp/dictionary.mobi",
    "language_from": "fr-fr",
    "language_to": "en-us",
    "dictd_hostname": "localhost",
    "dictd_port": 2628,
    "dictd_database": "fra-eng",
    "snowball_language": "french",
    "aspell_language": "fr",
    "generator_cls": "mobidick.generator.Generator",
    "kindlegen_path": "<kindlegen_path>/kindlegen",
    "logging_level": "INFO"
}

You have this example file at config/example.json.

Run

python mobidick.py --settings config/example.json

Test

Add test words with some of their inflections to your dictionary file:

{
...
    "test_words": { 
	    "aimer": ["aime", "aimerais"],
    	"trancher": ["tranchait", "trancherait"],
	    "porter": ["portèrent", "porterait"]
    }
...
}

Run the tests:

python test_mobidick.py --settings config/example.json

Extend Generator

You can extend from Generator class if you want to change the default behavior. All you need to know is that there are four steps in dictionary generation corresponding to the following Generator methods that you can override:

1. words (get words from aspell)
2. stems (group words by stem using snowball)
3. definitions (add definitions from dictd)
4. templates (compile .opf and .html required templates)
5. mobi (call kindle generator)

You can, for example, override definitions if you want to lookup at Wikipedia those words that are not found on dictd. Or you can override words if you want to get words from other source than aspell.

You just need to extend Generator as it is shown at mobidick/custom.py example:

from mobidick.generator import Generator


class SampleGenerator(Generator):

    def stems(self, words):
        stems = super(SampleGenerator, self).stems(words)
        return dict(list(stems.items())[0:25000])

And configure your class in your configuration file:

{
...
    "generator_cls": "mobidick.custom.SampleGenerator",
...
}

Then run the generator as always.

Custom Generators

You can configure any of these generators through your configuration file:

{
...
    "generator_cls": "mobidick.custom.<custom_generator>",
...
}

Of course, you can also extend from these generators as from the Generator base class, if you want a more specific behavior.

MemoizedGenerator

Use this generator if you want to memoize steps. This is specially useful for development.