Spaces:
Sleeping
Sleeping
| from paddleocr import PaddleOCR | |
| import re | |
| # Enable multilingual support (English + Hindi + Tamil) | |
| ocr = PaddleOCR(use_angle_cls=True, lang='en|hi|ta') | |
| def extract_kyc_fields(file_path): | |
| try: | |
| result = ocr.ocr(file_path, cls=True) | |
| lines = [] | |
| for block in result: | |
| for line in block: | |
| text = line[1][0].strip() | |
| if text: | |
| lines.append(text) | |
| full_text = "\n".join(lines) | |
| # PAN pattern: 5 letters + 4 digits + 1 letter | |
| pan_match = re.search(r'\b[A-Z]{5}[0-9]{4}[A-Z]\b', full_text) | |
| aadhaar_match = re.search(r'\b\d{4}[\s\-]?\d{4}[\s\-]?\d{4}\b', full_text) | |
| card_type = "UNKNOWN" | |
| if pan_match: | |
| card_type = "PAN" | |
| elif aadhaar_match: | |
| card_type = "AADHAAR" | |
| response = {"card_type": card_type} | |
| # --------- PAN CARD LOGIC --------- | |
| if card_type == "PAN": | |
| response["pan_number"] = pan_match.group(0) | |
| # Extract DOB | |
| dob = "Not found" | |
| for line in lines: | |
| match = re.search(r'\d{2}[/-]\d{2}[/-]\d{4}', line) | |
| if match: | |
| dob = match.group(0) | |
| break | |
| response["dob"] = dob | |
| # Extract Name | |
| name = "Not found" | |
| for i in range(len(lines)): | |
| if "INCOME TAX DEPARTMENT" in lines[i].upper(): | |
| for j in range(i+1, len(lines)): | |
| possible = lines[j].strip() | |
| if ( | |
| re.match(r'^[A-Z\s.]+$', possible) | |
| and not any(x in possible for x in ["INDIA", "DEPARTMENT", "GOVT"]) | |
| and not re.search(r'\d', possible) | |
| ): | |
| name = possible.strip() | |
| break | |
| break | |
| response["name"] = name | |
| # --------- AADHAAR CARD LOGIC --------- | |
| elif card_type == "AADHAAR": | |
| response["aadhaar_number"] = aadhaar_match.group(0) | |
| # Extract DOB | |
| dob = "Not found" | |
| for line in lines: | |
| match = re.search(r'\d{2}[/-]\d{2}[/-]\d{4}', line) | |
| if match: | |
| dob = match.group(0) | |
| break | |
| response["dob"] = dob | |
| # Extract Gender | |
| gender = "Not found" | |
| for line in lines: | |
| if "MALE" in line.upper(): | |
| gender = "MALE" | |
| break | |
| elif "FEMALE" in line.upper(): | |
| gender = "FEMALE" | |
| break | |
| elif "TRANSGENDER" in line.upper(): | |
| gender = "TRANSGENDER" | |
| break | |
| response["gender"] = gender | |
| # Robust name extraction | |
| name = "Not found" | |
| # First attempt: line before DOB | |
| for i, line in enumerate(lines): | |
| if re.search(r'\d{2}[/-]\d{2}[/-]\d{4}', line) and i > 0: | |
| possible_name = lines[i - 1].strip() | |
| if ( | |
| not any(x in possible_name.upper() for x in ["GOVERNMENT", "INDIA", "DOB", "MALE", "FEMALE"]) | |
| and not re.search(r'\d', possible_name) | |
| and len(possible_name.split()) >= 2 | |
| ): | |
| name = possible_name | |
| break | |
| # Fallback: best guess line with title-cased text and no digits | |
| if name == "Not found": | |
| for line in lines: | |
| if ( | |
| not re.search(r'\d', line) | |
| and len(line.split()) >= 2 | |
| and line[0].isupper() | |
| and not any(x in line.upper() for x in ["GOVERNMENT", "INDIA", "DOB", "MALE", "FEMALE"]) | |
| ): | |
| name = line.strip() | |
| break | |
| response["name"] = name | |
| else: | |
| response["error"] = "Unable to determine document type (PAN/Aadhaar)." | |
| return response | |
| except Exception as e: | |
| return {"error": f"OCR processing failed: {str(e)}"} | |