Update app.py

app.py

@@ -1,9 +1,15 @@
+import os
+import sys
+
+# Set EasyOCR environment variables BEFORE importing anything else
+os.environ['EASYOCR_MODULE_PATH'] = '/app/.cache/easyocr'
+os.environ['HOME'] = '/app'
 
+# Now import everything else
 from flask import Flask, request, jsonify
 from flask_cors import CORS
 import base64
 import io
-import os
 from PIL import Image
 import logging
 from transformers import TrOCRProcessor, VisionEncoderDecoderModel
@@ -42,11 +48,14 @@ def initialize_models():
             
             # Set cache directories
             cache_dir = os.environ.get('TRANSFORMERS_CACHE', '/app/.cache/huggingface')
-            easyocr_cache = os.environ.get('EASYOCR_MODEL_PATH', '/app/.cache/easyocr')
+            easyocr_cache = '/app/.cache/easyocr'
             
             os.makedirs(cache_dir, exist_ok=True)
             os.makedirs(easyocr_cache, exist_ok=True)
             
+            # Make sure the EasyOCR cache directory is writable
+            os.chmod(easyocr_cache, 0o777)
+            
             # Initialize TrOCR for handwritten text (Microsoft's model)
             logger.info("Loading TrOCR model for handwritten text...")
             trocr_processor = TrOCRProcessor.from_pretrained(
@@ -60,12 +69,18 @@ def initialize_models():
             
             # Initialize EasyOCR for printed text with custom model directory
             logger.info("Loading EasyOCR for printed text...")
-            easyocr_reader = easyocr.Reader(
-                ['en'], 
-                gpu=torch.cuda.is_available(),
-                model_storage_directory=easyocr_cache,
-                download_enabled=True
-            )
+            try:
+                easyocr_reader = easyocr.Reader(
+                    ['en'], 
+                    gpu=torch.cuda.is_available(),
+                    model_storage_directory=easyocr_cache,
+                    download_enabled=True
+                )
+            except Exception as e:
+                logger.warning(f"Failed to initialize EasyOCR with custom directory: {e}")
+                # Fallback: try without specifying directory
+                logger.info("Attempting EasyOCR initialization without custom directory...")
+                easyocr_reader = easyocr.Reader(['en'], gpu=torch.cuda.is_available())
             
             models_loaded = True
             logger.info("All models loaded successfully!")
@@ -352,6 +367,7 @@ def models_info():
         "supported_types": ["auto", "handwritten", "printed"],
         "supported_formats": ["PNG", "JPEG", "JPG", "BMP", "TIFF"],
         "cache_directory": os.environ.get('TRANSFORMERS_CACHE', '/app/.cache/huggingface'),
+        "easyocr_cache": os.environ.get('EASYOCR_MODULE_PATH', '/app/.cache/easyocr'),
         "gpu_available": torch.cuda.is_available(),
         "models_loaded": models_loaded
     })
@@ -396,4 +412,4 @@ if __name__ == '__main__':
     app.run(host='0.0.0.0', port=port, debug=False)
 else:
     # Running with gunicorn - just log startup, don't do anything else
-    logger.info("OCR service ready - models will load on first request")
+    logger.info("OCR service ready - models will load on first request")