Spaces:
Running
Running
| """Lightweight OCR model using Tesseract + RapidOCR fallback. | |
| No PyTorch or heavy ML deps — works on HF free tier. | |
| """ | |
| import logging | |
| import os | |
| import shutil | |
| from pathlib import Path | |
| try: | |
| import pytesseract | |
| from PIL import Image | |
| HAS_TESSERACT = True | |
| except ImportError: | |
| HAS_TESSERACT = False | |
| try: | |
| from rapidocr_onnxruntime import RapidOCR | |
| HAS_RAPIDOCR = True | |
| except ImportError: | |
| HAS_RAPIDOCR = False | |
| class OCRModel: | |
| def __init__(self, tesseract_cmd=""): | |
| self._rapidocr_engine = None | |
| self._tesseract_path = self._find_tesseract(tesseract_cmd) if HAS_TESSERACT else "" | |
| logging.getLogger("RapidOCR").disabled = True | |
| def _find_tesseract(self, tesseract_cmd=""): | |
| env_path = tesseract_cmd.strip() or os.environ.get("TESSERACT_CMD", "").strip() | |
| if env_path and Path(env_path).exists(): | |
| return env_path | |
| cmd = shutil.which("tesseract") | |
| if cmd: | |
| return cmd | |
| for p in (r"C:\Program Files\Tesseract-OCR\tesseract.exe", | |
| r"C:\Program Files (x86)\Tesseract-OCR\tesseract.exe", | |
| "/usr/bin/tesseract"): | |
| if Path(p).exists(): | |
| return p | |
| return "" | |
| def extract_text(self, image_path): | |
| """Extract text from image using Tesseract, falling back to RapidOCR.""" | |
| text = self._tesseract_ocr(image_path) | |
| if text: | |
| return text | |
| return self._rapidocr(image_path) | |
| def _tesseract_ocr(self, image_path): | |
| if not HAS_TESSERACT or not self._tesseract_path: | |
| return "" | |
| pytesseract.pytesseract.tesseract_cmd = self._tesseract_path | |
| try: | |
| image = Image.open(image_path) | |
| return pytesseract.image_to_string(image).strip() | |
| except Exception: | |
| return "" | |
| def _rapidocr(self, image_path): | |
| if not HAS_RAPIDOCR: | |
| return "" | |
| if self._rapidocr_engine is None: | |
| self._rapidocr_engine = RapidOCR() | |
| try: | |
| result = self._rapidocr_engine(str(image_path)) | |
| if result and result[0]: | |
| texts = [item[1].strip() for item in result[0] if item[1].strip()] | |
| return "\n".join(texts).strip() | |
| except Exception: | |
| pass | |
| return "" | |