AWS Textract issue

What's about splitting document to different pages and process them in parallel ?

Use a tool or library (like PyPDF2 for Python) to split the PDF into individual pages or smaller chunks.

from PyPDF2 import PdfFileReader, PdfFileWriter

def split_pdf(input_pdf):
    pdf = PdfFileReader(input_pdf)
    for page_num in range(pdf.numPages):
        pdf_writer = PdfFileWriter()
        pdf_writer.addPage(pdf.getPage(page_num))
        
        output_filename = f'page_{page_num + 1}.pdf'
        with open(output_filename, 'wb') as out:
            pdf_writer.write(out)
        yield output_filename

Use AWS Lambda or Step Functions to process each split PDF concurrently using Textract’s asynchronous API.

import boto3
import concurrent.futures

textract = boto3.client('textract')

def process_pdf(file_name):
    with open(file_name, 'rb') as document:
        response = textract.start_document_text_detection(Document={'Bytes': document.read()})
        return response['JobId']

with concurrent.futures.ThreadPoolExecutor() as executor:
    file_names = list(split_pdf('large_document.pdf'))
    future_to_file = {executor.submit(process_pdf, file): file for file in file_names}
    for future in concurrent.futures.as_completed(future_to_file):
        file = future_to_file[future]
        try:
            job_id = future.result()
            print(f'Job ID for {file}: {job_id}')
        except Exception as exc:
            print(f'{file} generated an exception: {exc}')