aalars / django-translate Goto Github PK

Few helpers that somewhat automate the process of translating a Python (Django) project.

On high level the process is (with some manual steps in between)

• Extract the strings to be translated from the code
• Get the translations from .po file or translate the strings using DeepL Translate API
• Replace the strings in the code with the translated strings wrapped in gettext() or gettext_lazy() function
• Save the translations into a dictionary in the _translations.py file
• Translate the strings in the .po files using the dictionary

Uses libcst to parse the Python code and extract the strings to be translated.

These strings will be translated to English using DeepL Translate API and replaced with the translated strings in the code.

gettext_lazy() is only used if the path has models. or models/ in it.

Migration files are ignored (checks if migrations is in path).

Translations will be saved into a dictionary in the _translations.py file in the format:

translations = {
    '"Puidu tüübid"': '"Types of wood"',
}

If you only want to translate missing strings in .po files, use -to/--translate_only flag. In order to use different target language than EN, use -l/--language flag.

poetry run python main.py -to -l ET -p path/to/your/po/file

Uses polib to parse the .po files and replace the strings with the translated strings from the _translations.py file.

1. Clone the repo
2. Run poetry install to install the dependencies
3. Create yourself a DeepL API key and save it in the DEEPL_API_KEY environment variable

Run the command for the first time, check logs for errors, revert those changes and fix the errors manually. Commit fixes. Use optional parameter -p with path to a .po file to first find for translations in tha file.

poetry run python main.py -d path/to/your/directory -p path/to/your/po/file

Why? Because libcst doesn't handle some strings well, and our code is not so good either. Also, our code does not detect if the string is already in a gettext function.

If there are some repeating strings that don't need to be translated, add them to the IGNORE_WORDS list in the constants.py file.

If you want then you can also improve the PATTERN regex in the constants.py file.

When you are somewhat happy with the results, run the command last time. Manually go over the changes and fix any errors, keeping the translations in the dictionary in sync with the changes. Run poetry run black . to format the dictionary.

Run ./manage.py makemessages to update the .po files.

Run the command below. This takes the values from the dictionary in the _translations.py file and replaces the strings in the .po files. Manually go over the changes to make sure everything is good to go.

poetry run python update_po.py path/to/your/po/file

• Make it recognize and ignore if the string is already in a gettext function
• Better logic to detect if gettext or gettext_lazy should be used
• Would be better if EN strings were keys in the dictionary
• Possibly can skip the _translations.py file and just use the .po files
• Add tests
• Add possibility of a dry run -- gettext linter?
• Add imports
• Replace _() with respective gettext function
• Do not save translations in the dictionary with double quotes, use ast.literal_eval(string)
• Update doc to reflect the changes (update_po_from_another_po.py, ...)