The code behind https://kantoniko.com/

Development environment

Create a folder called kantoniko. Inside that folder clone the repositories of https://github.com/kantoniko/

Clone this project that contains the code. This is enought to run the tests and develop the code.

git clone https://github.com/kantoniko/ladino-diksionaryo-code.git

In order to work with the real data of the dictionary clone https://github.com/kantoniko/ladino-diksionaryo-data/ using the following command:

git clone https://github.com/kantoniko/ladino-diksionaryo-data.git

Install the dependencies.

pip install -r requirements.txt

Testing

pytest -vvs tests/test_generate.py
pytest -vvs -rA -x --log-cli-level=DEBUG --random-order tests/test_generate.py

the expected output files are in the tests subdirectory.

When the expected output changes we can update the files with the following command:

pytest -vvs tests/test_generate.py --save

Generate test coverage report:

pytest -vvs -rA -x --log-cli-level=DEBUG --random-order --cov=ladino --cov-report html --cov-report term --cov-branch tests/test_generate.py

Generate the site locally

These commands will generate the static files in the docs subdrirectory. Below you'll find the command to start a local web server to view the results.

Generate test pages

Some selected pieces of data (words, examples, etc.) that we stored in the repository of the code to serve us for testing.

./generate_sample.sh

Generate only the dictionary

For this we need the data repository cloned.

PYTHONPATH=. python ladino/generate.py --dictionary ../ladino-diksionaryo-data/ --html docs --all --log

This took 3.20 on my computer wheren there were 3321 words and 1556 example in the ladino-diksionaryo-data repository.

Generate the whole site locally

For this we need all the repositories to be cloned.

./generate.sh

time PYTHONPATH=. python ladino/generate.py --dictionary ../ladino-diksionaryo-data/ --html docs --all --log --whatsapp ../ladino-estamos-whatsapeando/ --sounds ../ladino-diksionaryo-sounds/ --unafraza ../ladino-una-fraza-al-diya/ --pages ../ladino-pages --books ../ladino-salu-lulu/  --ladinadores ../ladino-los-ladinadores/ --limit 10

Launch a static web server with the following command:

Run locally

./app.py

or

FLASK_DEBUG=1 flask run

Language considerations

• Verbs in Ladino have a lot of conjugations. (are there as many as in modern Spanish?)

• Several forms of the same verb translate to the same form in English: yo komo I eat to komes you eat

• On the other hand they translate to both feminine and masculine in Hebrew yo komo אני אוכל / אני אוכלת

• "komo" has other meanings as well, for exampl "like" and "how?"

• Some nouns, such as "meza" = "table" have a single gender but they have both singular (meza) and plural form ("mezas").

• Some nouns, such as "elevo" = "student" have both masculine (elevo) and feminine (eleva) form in singular and (elevos, elevas) in plural.

• Some words, such as "klaro" = "clear" can act noun, adjective, or adverb (at least in modern Spanish). The English translation (clear) can also act as verb which then translates back to aklarar.

Design

From the yaml files we generate several json files and several static html files.

• dictionary.json

    "ladino": {
        "kome": {
            "ladino": "kome",
            "translations": {
                "inglez": "eats",
                "kasteyano": "come"
            }
        },
    },
    "inglez": {
        "eat": "komen",
        "eats": "kome"
    },

Rashi fonts

https://fonts.google.com/noto/specimen/Noto+Rashi+Hebrew/about