Spaces:
Sleeping
Sleeping
krishnachoudhary-hclguvi commited on
Fix EasyOCR crash before metadata generation
Browse files
extractors/ocr_extractor.py
CHANGED
|
@@ -136,6 +136,13 @@ def _reconstruct_from_boxes(results: list) -> str:
|
|
| 136 |
def extract_image(file_path: str) -> ExtractionResult:
|
| 137 |
"""Extract text from an image using the best available OCR engine."""
|
| 138 |
start_time = time.time()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 139 |
|
| 140 |
# 1. Check for EasyOCR (Preferred)
|
| 141 |
if EASYOCR_AVAILABLE:
|
|
|
|
| 136 |
def extract_image(file_path: str) -> ExtractionResult:
|
| 137 |
"""Extract text from an image using the best available OCR engine."""
|
| 138 |
start_time = time.time()
|
| 139 |
+
original_size = (0, 0)
|
| 140 |
+
try:
|
| 141 |
+
with Image.open(file_path) as img:
|
| 142 |
+
original_size = img.size
|
| 143 |
+
except Exception:
|
| 144 |
+
# Keep defaults; OCR engines will surface the real file/open errors.
|
| 145 |
+
pass
|
| 146 |
|
| 147 |
# 1. Check for EasyOCR (Preferred)
|
| 148 |
if EASYOCR_AVAILABLE:
|