ocr_processor.py

import pytesseract
import os
import csv
import re
import tempfile
from pdf2image import convert_from_path
from datetime import datetime

pytesseract.pytesseract.tesseract_cmd = "/usr/bin/tesseract"
POPPLER_PATH = "/usr/bin"
OCR_LANG = "eng"

def parse_structured_table_line(line):
    pattern = re.compile(
        r'(?P<cin>U\d{5}[A-Z]{2}\d{4}PTC\d{6})\W*' +
        r'(?P<company>.+?)\s+' +
        r'(?P<din>\d{6,10})\W*' +
        r'(?P<director>[A-Z\s\.]+?)\s+' +
        r'(?P<roc>\d{3})\W*' +
        r'(?P<status>Disqualified|Qualified)\W*' +
        r'(?P<from>\d{2}\.\d{2}\.\d{4})\W*' +
        r'(?P<till>\d{2}\.\d{2}\.\d{4})',
        re.IGNORECASE
    )
    match = pattern.search(line)
    if match:
        return [
            match.group("cin").strip(),
            match.group("company").strip(),
            match.group("din").strip(),
            match.group("director").strip(),
            match.group("roc").strip(),
            match.group("status").strip(),
            match.group("from").strip(),
            match.group("till").strip()
        ]
    return None

def process_pdf(file, filename):
    results = []
    try:
        with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf") as tmp:
            tmp.write(file.read())
            tmp_path = tmp.name

        images = convert_from_path(tmp_path, dpi=200, poppler_path=POPPLER_PATH)
        for page_num, image in enumerate(images, start=1):
            text = pytesseract.image_to_string(image, lang=OCR_LANG)
            lines = [line.strip() for line in text.splitlines() if line.strip()]
            for line_num, line in enumerate(lines, start=1):
                parsed = parse_structured_table_line(line)
                if parsed:
                    results.append((filename, page_num, line_num, *parsed))
                else:
                    results.append((filename, page_num, line_num, '', '', '', '', '', '', '', line))
        os.remove(tmp_path)
    except Exception as e:
        print(f"❌ Error with {filename}: {e}")
    return results

def process_pdf_files(uploaded_files):
    all_results = []
    for uploaded_file in uploaded_files:
        all_results.extend(process_pdf(uploaded_file, uploaded_file.name))

    output_file = os.path.join("output", f"structured_output.csv")
    os.makedirs("output", exist_ok=True)
    with open(output_file, "w", newline="", encoding="utf-8") as out_csv:
        writer = csv.writer(out_csv)
        writer.writerow([
            "Filename", "Page", "Line",
            "CIN", "Company Name", "DIN", "Director Name",
            "ROC Code", "Status", "Period From", "Period Till",
            "Raw Line (Fallback)"
        ])
        writer.writerows(all_results)

    return output_file