app.py

import subprocess
import json
import os
import gradio as gr
from PyPDF2 import PdfReader, PdfWriter

# Function to split PDF into batches of 3 pages
def split_pdf(file_path, batch_size=3):
    pdf_reader = PdfReader(open(file_path, "rb"))
    total_pages = len(pdf_reader.pages)
    pdf_batches = []

    # Split the PDF into batches of 3 pages
    for i in range(0, total_pages, batch_size):
        pdf_writer = PdfWriter()
        for j in range(i, min(i + batch_size, total_pages)):
            pdf_writer.add_page(pdf_reader.pages[j])
        
        batch_path = f"./temp_batch_{i // batch_size}.pdf"
        with open(batch_path, "wb") as batch_file:
            pdf_writer.write(batch_file)
        
        pdf_batches.append(batch_path)
    
    return pdf_batches

# Function to process the PDF batch using subprocess
def process_pdf_batch(batch_path, output_dir):
    # Extract the base name of the batch file
    pdf_name = os.path.basename(batch_path).split('.')[0]
    result_path = os.path.join(output_dir, pdf_name, "results.json")
    
    # Build the OCR command
    ocr_command = ["surya_ocr", batch_path, "--results_dir", output_dir]
    
    # Run the command using subprocess
    try:
        result = subprocess.run(ocr_command, check=True, text=True, capture_output=True,encoding="utf-8")
        print("OCR Command Output:", result.stdout)
    except subprocess.CalledProcessError as e:
        return f"OCR processing failed: {e.stderr}"
    
    # After OCR processing, read the results from the JSON file
    if os.path.exists(result_path):
        with open(result_path, 'r', encoding="utf-8") as f:
            data = json.load(f)
        
        # Extract text from the JSON
        result_text = ''
        for page_data in data[pdf_name]:
            for line in page_data['text_lines']:
                result_text += line['text'] + '\n'
                
        return result_text
    else:
        return "OCR processing completed, but result file not found."

# Main function to process the entire PDF in batches
def process_pdf(file):
    # Define output directory
    output_dir = "./result"
    
    # Split the uploaded PDF into batches of 3 pages
    pdf_batches = split_pdf(file.name, batch_size=3)
    
    # Process each batch and accumulate results
    final_text = ""
    for batch_path in pdf_batches:
        batch_result = process_pdf_batch(batch_path, output_dir)
        final_text += batch_result + "\n"
    
    return final_text

# Define Gradio interface
def process_pdf_gradio(file):
    # Gradio handles the file upload differently, so process accordingly
    result = process_pdf(file)
    return result

# Gradio app
app = gr.Interface(
    fn=process_pdf_gradio,
    inputs=gr.File(label="Upload PDF"),
    outputs=gr.Textbox(label="Extracted Text"),
    title="PDF OCR Extractor"
)

# Launch the app with a specified port for Docker
app.launch(server_name="0.0.0.0", server_port=7860, share=True)