Spaces:
Sleeping
Sleeping
| from paddleocr import PaddleOCR | |
| import re | |
| # Initialize OCR for English | |
| ocr = PaddleOCR(use_angle_cls=True, lang='en') | |
| def _extract_dob(lines): | |
| for line in lines: | |
| m = re.search(r'\b\d{2}[./-]\d{2}[./-]\d{4}\b', line) | |
| if m: return m.group(0) | |
| for line in lines: | |
| m = re.search(r'\b\d{4}-\d{2}-\d{2}\b', line) | |
| if m: return m.group(0) | |
| for line in lines: | |
| m = re.search(r'\b(19|20)\d{2}\b', line) | |
| if m and any(lbl in line.upper() for lbl in ["YOB", "YEAR", "BIRTH"]): | |
| return m.group(0) | |
| return "Not found" | |
| def extract_kyc_fields(file_path: str) -> dict: | |
| 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_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) | |
| if pan_match: | |
| card_type = "PAN" | |
| elif aadhaar_match: | |
| card_type = "AADHAAR" | |
| else: | |
| return {"card_type": "UNKNOWN", "error": "Could not identify document as PAN or Aadhaar."} | |
| response = {"card_type": card_type} | |
| if card_type == "PAN": | |
| response["pan_number"] = pan_match.group(0) | |
| response["dob"] = _extract_dob(lines) | |
| 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.upper() for x in ["INDIA", "GOVT", "DEPARTMENT"]) | |
| and not re.search(r'\d', possible) | |
| and len(possible) >= 3 | |
| ): | |
| name = possible | |
| break | |
| break | |
| response["name"] = name | |
| else: # AADHAAR | |
| response["aadhaar_number"] = aadhaar_match.group(0) | |
| response["dob"] = _extract_dob(lines) | |
| gender = "Not found" | |
| for line in lines: | |
| up = line.upper() | |
| if "TRANSGENDER" in up: | |
| gender = "TRANSGENDER"; break | |
| if "FEMALE" in up: | |
| gender = "FEMALE"; break | |
| if "MALE" in up: | |
| gender = "MALE"; break | |
| response["gender"] = gender | |
| name = "Not found" | |
| for i, line in enumerate(lines): | |
| if re.search(r'\d{2}[./-]\d{2}[./-]\d{4}', line) and i > 0: | |
| candidate = lines[i - 1].strip() | |
| if ( | |
| not re.search(r'\d', candidate) | |
| and len(candidate.split()) >= 2 | |
| and not any(x in candidate.upper() for x in ["DOB", "INDIA", "MALE", "FEMALE", "GOVERNMENT"]) | |
| ): | |
| name = candidate | |
| break | |
| if name == "Not found": | |
| for line in lines: | |
| candidate = line.strip() | |
| if ( | |
| not re.search(r'\d', candidate) | |
| and len(candidate.split()) >= 2 | |
| and not any(x in candidate.upper() for x in ["DOB", "INDIA", "MALE", "FEMALE", "GOVERNMENT"]) | |
| ): | |
| name = candidate | |
| break | |
| response["name"] = name | |
| return response | |
| except Exception as e: | |
| return {"error": f"OCR processing failed: {str(e)}"} | |