rn0x / transtexa

A versatile library for text translation using Hugging Face's Transformers

transtexa

translate.py

Handles text translation using pretrained models from Hugging Face's Transformers library

• Loads or downloads models based on provided IDs.
• Translates input text using loaded models and tokenizers.
• Supports command-line arguments for model ID, input text, and storage directory.

index.js

Manages communication with translate.py from Node.js:

• Locates Python executable dynamically.
• Installs Python dependencies from requirements.txt.
• Uses config.json for configuration and models_with_languages.json for model selection.
• Spawns Python subprocesses for translation tasks.

models_with_languages.json

Metadata about available translation models:

• Lists model IDs, source and target languages.

Helsinki-NLP

Refers to NLP research at the University of Helsinki, known for AI leadership. It focuses on advancements in language processing.

Usage

1. Setup:

   • Ensure Python 3.x and Node.js are installed.
   • Run npm install transtexa

2. Translation:

   • Use TranslationModel methods (translateText, getSupportedLanguages) in Node.js.

3. Customization:

   • Modify models_with_languages.json to adjust models and languages.
   • Configure config.json for Python paths.

Example

import TranslationModel from './index.js';

const translator = new TranslationModel();

translator.translateText('zh', 'en', '你好吗？').then(translated => {
    console.log(translated); // Outputs: "How are you?"
}).catch(err => {
    console.error('Translation error:', err);
});

Node.js Integration with Telegram Bot

import Telegraf from 'telegraf';
import TranslationModel from './index.js';

const bot = new Telegraf('YOUR_TELEGRAM_BOT_TOKEN');
const translator = new TranslationModel();

bot.on('text', async (ctx) => {
    const sourceLanguage = 'en'; // Assuming English as input language
    const targetLanguage = 'fr'; // Assuming French as output language
    const textToTranslate = ctx.message.text;

    try {
        const translatedText = await translator.translateText(sourceLanguage, targetLanguage, textToTranslate);
        ctx.reply(translatedText);
    } catch (error) {
        console.error('Translation error:', error);
        ctx.reply('Translation failed. Please try again later.');
    }
});

bot.launch();

Node.js Integration with Express.js Server

import express from 'express';
import TranslationModel from './index.js';

const app = express();
const port = 3000;
const translator = new TranslationModel();

app.use(express.json());

app.post('/translate', async (req, res) => {
    const { sourceLanguage, targetLanguage, text } = req.body;

    try {
        const translatedText = await translator.translateText(sourceLanguage, targetLanguage, text);
        res.json({ translatedText });
    } catch (error) {
        console.error('Translation error:', error);
        res.status(500).json({ error: 'Translation failed. Please try again later.' });
    }
});

app.listen(port, () => {
    console.log(`Server is running on http://localhost:${port}`);
});