python-pdf2txt/pdf_to_word.py

import logging
from pdf2image import convert_from_path
import pytesseract
from PIL import Image
import tempfile
import os
from docx import Document
from datetime import datetime

# Configure logging
logging.basicConfig(level=logging.INFO, filename='/app/logs/pdf_to_word.log', filemode='w',
                    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s')

def pdf_to_word_txt(pdf_path, output_dir, lang='fas+eng'):
    pdf_name = os.path.basename(pdf_path).split('.')[0]

    current_time = datetime.now().strftime("%Y%m%d_%H%M%S")
    output_path = os.path.join(output_dir, f'{pdf_name}_{current_time}.txt')

    try:
        pages = convert_from_path(pdf_path, 600)
        logging.info(f'Converted PDF to images, number of pages: {len(pages)}')
        all_text = []
        for i, page in enumerate(pages):
            with tempfile.NamedTemporaryFile(suffix='.jpg') as temp_image:
                page.save(temp_image.name, 'JPEG')
                text = pytesseract.image_to_string(Image.open(temp_image.name), lang=lang, config='--oem 3 --psm 3')
                logging.info("\n------"+f'Page {i+1}'+"------\n"+text+"\n------*------\n")
                all_text.append(text)
        
        with open(output_path, 'w', encoding='utf-8') as f:
            f.write("\n".join(all_text))

        logging.info(f'Successfully converted {pdf_path} to text at {output_path}')
        return output_path
    except Exception as e:
        logging.error(f'Error converting PDF to text: {e}', exc_info=True)
        return None




def pdf_to_word_docx(pdf_path, output_dir, lang='eng+fas'):
    pdf_name = os.path.basename(pdf_path).split('.')[0]
    # output_path = os.path.join(output_dir, f'{pdf_name}.docx')
    current_time = datetime.now().strftime("%Y%m%d_%H%M%S")
    output_path = os.path.join(output_dir, f'{pdf_name}_{current_time}.docx')

    try:
        # Increase DPI for higher image quality
        pages = convert_from_path(pdf_path, 300)
        logging.info(f'Converted PDF to images, number of pages: {len(pages)}')
        
        # PSM 4 might be more suitable for mixed content like your provided image
        custom_config = r'--oem 3 --psm 4'
        # custom_config = r''
        all_text = []
        
        for i, page in enumerate(pages):
            with tempfile.NamedTemporaryFile(suffix='.jpg') as temp_image:
                page.save(temp_image.name, 'JPEG')
                text = pytesseract.image_to_string(Image.open(temp_image.name), lang=lang, config=custom_config)
                logging.info("\n------"+f'Page {i+1}'+"------\n"+text+"\n------*------\n")
                all_text.append(text)
        
        # Save to a .docx file using a library like python-docx
        doc = Document()
        for text in all_text:
            doc.add_paragraph(text)
        doc.save(output_path)
        
        logging.info(f'Successfully converted {pdf_path} to a .docx file at {output_path}')
        return output_path
    except Exception as e:
        logging.error(f'Error converting PDF to text: {e}', exc_info=True)
        return None