Update app.py

app.py

@@ -158,7 +158,7 @@ class OCRSystem:
         """Initialize OCR models configuration based on available hardware"""
         self.models = {}
         self.model_performance = {
-            'microsoft_trocr': {'success': 0, 'fail': 0},  # Added Microsoft TrOCR
+            'microsoft_trocr': {'success': 0, 'fail': 0},
             'pretrained_model': {'success': 0, 'fail': 0},
             'mT5_OCR_fa': {'success': 0, 'fail': 0},
             'LayoutLMv3_fa': {'success': 0, 'fail': 0},
@@ -172,7 +172,7 @@ class OCRSystem:
         self.logger.info(f"Using device: {self.device}, Max workers: {self.max_workers}")
 
         self.model_configs = {
-            'microsoft_trocr': {  # Microsoft TrOCR for typed documents
+            'microsoft_trocr': {
                 'name': "microsoft/trocr-base-printed",
                 'type': "transformer",
                 'threshold': 0.85,
@@ -223,7 +223,7 @@ class OCRSystem:
             self.logger.warning(f"WARNING: Cannot create {tessdata_dir}: {str(e)}. Tesseract may fail without language data.")
             return False
         
-        base_url = "https://github.com/tesseract-ocr/tessdata/raw/master/"
+        base_url = "https://github.com/tesseract-ocr/tessdata_best/raw/main/"  # Updated URL
         
         for lang in languages:
             file_path = os.path.join(tessdata_dir, f"{lang}.traineddata")
@@ -269,13 +269,15 @@ class OCRSystem:
                     return False
 
             # Download Tesseract language data
-            self.download_tessdata(['eng', 'fas'])
-            tessdata_prefix = TESSDATA_LOCAL
+            tessdata_downloaded = self.download_tessdata(['eng', 'fas'])
+            tessdata_prefix = TESSDATA_LOCAL if tessdata_downloaded else TESSDATA_PREFIX_DEFAULT
             os.environ['TESSDATA_PREFIX'] = tessdata_prefix
             pytesseract.pytesseract.tesseract_cmd = TESSERACT_CMD
             
             version = subprocess.check_output([TESSERACT_CMD, '--version'])
             self.logger.info(f"SUCCESS: Tesseract installed: {version.decode()[:50]}")
+            if not os.path.exists(os.path.join(tessdata_prefix, 'eng.traineddata')):
+                self.logger.warning(f"WARNING: Tesseract language data not found in {tessdata_prefix}, functionality may be limited")
             self.logger.info("SUCCESS: System dependencies setup completed")
             return True