translate.py

import os
import sys
import requests
from docx import Document
from googletrans import Translator
import PyPDF2

# function to translate text using Google Translate API
def translate_text(text, src_lang, dest_lang):
    translator = Translator()
    translation = translator.translate(text, src=src_lang, dest=dest_lang)
    return translation.text

# function to convert PDF to DOCX using PyPDF2
def convert_pdf_to_docx(file_path):
    docx_file_path = os.path.splitext(file_path)[0] + ".docx"
    with open(file_path, 'rb') as pdf_file, open(docx_file_path, 'wb') as docx_file:
        pdf_reader = PyPDF2.PdfReader(pdf_file)
        document = Document()
        for page in range(len(pdf_reader.pages)):
            page_content = pdf_reader.pages[page].extract_text()
            if page_content is not None:
                document.add_paragraph(page_content)
        document.save(docx_file)
    return docx_file_path

# function to convert DOCX to PDF using LibreOffice
def convert_docx_to_pdf(file_path):
    pdf_file_path = os.path.splitext(file_path)[0] + "_translated.pdf"
    outdir = os.path.join(os.path.dirname(os.path.abspath(file_path)), "PDF_Output")
    os.system(f"libreoffice --headless --convert-to pdf {os.path.abspath(file_path)} --outdir {outdir}")
    return pdf_file_path

# main function to translate PDF to English
def translate_pdf_to_english(file_path):
    # convert PDF to DOCX
    docx_file_path = convert_pdf_to_docx(file_path)

    # read DOCX file and translate text
    document = Document(docx_file_path)
    text = "\n".join([paragraph.text for paragraph in document.paragraphs])
    translated_text = translate_text(text, 'auto', 'en')

    # save translated text to DOCX file
    document = Document()
    paragraph = document.add_paragraph()
    paragraph.add_run(translated_text)
    translated_docx_file_path = os.path.splitext(file_path)[0] + "_translated.docx"
    document.save(translated_docx_file_path)

    # convert translated DOCX to PDF
    translated_pdf_file_path = convert_docx_to_pdf(translated_docx_file_path)

    # delete temporary files
    os.remove(docx_file_path)
    os.remove(translated_docx_file_path)

    return translated_pdf_file_path


# check if file path is provided as a command line argument
if len(sys.argv) != 2:
    print("Usage: python translate_pdf.py <file_path>")
    sys.exit(1)

# get file path from command line argument
file_path = sys.argv[1]

# check if file exists
if not os.path.isfile(file_path):
    print(f"Error: {file_path} does not exist.")
    sys.exit(1)

# example usage
translated_pdf_file = translate_pdf_to_english(file_path)
print(f"Translated PDF file saved to: {translated_pdf_file}")