translate2english

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
# Filename: translate2english.py
# Author: Jeoi Reqi

"""
Chinese To English Translator

This Python script translates Chinese text into English.

Usage:
1. Ensure you have Python 3 installed.
2. Replace 'input_cn.txt' with the name of your file containing Chinese text.
3. Replace 'output_en.txt' with the desired name for the translated English text output file.
4. Run the script, and the translated text will be saved in the specified output file.

Requirements:
- Python 3
"""

import os
from deep_translator import GoogleTranslator
import re

def detect_language(text):
    # Use a simple regex to detect if the text contains Chinese characters
    if re.search(r'[\u4e00-\u9fff]', text):
        return 'zh'
    else:
        return 'en'

def translate_text(text, target_language='en'):
    translator = GoogleTranslator(source='auto', target=target_language)
    translated_text = translator.translate(text)
    return translated_text

def split_text_into_sentences(text):
    # Split the text into sentences based on punctuation
    sentences = re.split(r'(?<!\w\.\w.)(?<![A-Z][a-z]\.)(?<=\.|\?)\s', text)
    return sentences

def main(input_file, output_file):
    with open(input_file, 'r', encoding='utf-8') as file:
        mixed_text = file.read()

    sentences = split_text_into_sentences(mixed_text)

    translated_sentences = []
    for sentence in sentences:
        source_language = detect_language(sentence)
        if source_language == 'zh':
            translated_sentence = translate_text(sentence)
            translated_sentences.append(translated_sentence)
        else:
            translated_sentences.append(sentence)

    translated_text = ' '.join(translated_sentences)

    with open(output_file, 'w', encoding='utf-8') as output_file:
        output_file.write(translated_text)

if __name__ == "__main__":
    current_directory = os.getcwd()
    input_file_path = os.path.join(current_directory, 'input_cn.txt')  # Replace 'input_cn.txt' with the text file containing the Chinese text.
    output_file_path = os.path.join(current_directory, 'output_en.txt')  # Replace 'outpu_en.txt' with the desired name for the translated text.
    main(input_file_path, output_file_path)