Spaces:
Sleeping
Sleeping
File size: 1,747 Bytes
d86db02 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 | import re
from dataclasses import dataclass
import easyocr
# Lazy-loaded singleton — easyocr model is ~200MB, load once
_reader: easyocr.Reader | None = None
_KYC_PATTERNS = [
re.compile(r"\baadhaar\b", re.IGNORECASE),
re.compile(r"\buidai\b", re.IGNORECASE),
re.compile(r"\bpan\b", re.IGNORECASE),
re.compile(r"\bpassport\b", re.IGNORECASE),
re.compile(r"\bgovt\.?\s+of\s+india\b", re.IGNORECASE),
re.compile(r"\bgovernment\s+of\s+india\b", re.IGNORECASE),
re.compile(r"\bdate\s+of\s+birth\b", re.IGNORECASE),
re.compile(r"\bpermanent\s+account\s+number\b", re.IGNORECASE),
re.compile(r"\bincome\s+tax\b", re.IGNORECASE),
re.compile(r"\b[2-9]\d{3}\s?\d{4}\s?\d{4}\b"), # 12-digit Aadhaar pattern
re.compile(r"\b[A-Z]{5}[0-9]{4}[A-Z]\b"), # PAN card format
]
@dataclass
class KYCResult:
filename: str
doc_type: str | None # "kyc" or None
send_to_llm: bool
ocr_text: str
def _get_reader() -> easyocr.Reader:
global _reader
if _reader is None:
_reader = easyocr.Reader(["en"], gpu=False)
return _reader
def _is_kyc(text: str) -> bool:
return any(pattern.search(text) for pattern in _KYC_PATTERNS)
def detect_kyc(filename: str, image_bytes: bytes, reader: easyocr.Reader | None = None) -> KYCResult:
"""Run OCR on image bytes and determine if the document is a KYC document."""
r = reader if reader is not None else _get_reader()
results = r.readtext(image_bytes, detail=0)
full_text = " ".join(results)
if _is_kyc(full_text):
return KYCResult(filename=filename, doc_type="kyc", send_to_llm=False, ocr_text=full_text)
return KYCResult(filename=filename, doc_type=None, send_to_llm=True, ocr_text=full_text)
|