levalencia commited on
Commit
cbcf1e4
·
1 Parent(s): 9626485

Configure EasyOCR model storage with enhanced error handling and logging

Browse files

- Set up a writable directory for EasyOCR model storage, with a primary path in a temporary directory and fallbacks to /tmp and the current working directory.
- Improved logging to provide feedback on the model directory creation process and any issues encountered.

Files changed (2) hide show
  1. Dockerfile +3 -2
  2. src/streamlit_app.py +20 -0
Dockerfile CHANGED
@@ -10,9 +10,10 @@ RUN apt-get update && apt-get install -y \
10
  && rm -rf /var/lib/apt/lists/*
11
 
12
  # Create necessary directories with proper permissions
13
- RUN mkdir -p /app/.streamlit /tmp/docling_temp && \
14
  chmod 755 /app/.streamlit && \
15
- chmod 777 /tmp/docling_temp
 
16
 
17
  COPY requirements.txt ./
18
  COPY src/ ./src/
 
10
  && rm -rf /var/lib/apt/lists/*
11
 
12
  # Create necessary directories with proper permissions
13
+ RUN mkdir -p /app/.streamlit /tmp/docling_temp /tmp/easyocr_models && \
14
  chmod 755 /app/.streamlit && \
15
+ chmod 777 /tmp/docling_temp && \
16
+ chmod 777 /tmp/easyocr_models
17
 
18
  COPY requirements.txt ./
19
  COPY src/ ./src/
src/streamlit_app.py CHANGED
@@ -63,10 +63,30 @@ os.environ['STREAMLIT_BROWSER_GATHER_USAGE_STATS'] = 'false'
63
  os.environ['STREAMLIT_SERVER_ENABLE_CORS'] = 'false'
64
  os.environ['STREAMLIT_SERVER_ENABLE_XSRF_PROTECTION'] = 'false'
65
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
66
  # Log startup information
67
  logging.info("=" * 50)
68
  logging.info("Docling Streamlit App Starting")
69
  logging.info(f"Temp directory: {TEMP_DIR}")
 
70
  logging.info(f"Current working directory: {os.getcwd()}")
71
  logging.info(f"Python version: {sys.version}")
72
  logging.info("=" * 50)
 
63
  os.environ['STREAMLIT_SERVER_ENABLE_CORS'] = 'false'
64
  os.environ['STREAMLIT_SERVER_ENABLE_XSRF_PROTECTION'] = 'false'
65
 
66
+ # Configure EasyOCR to use writable directory for model storage
67
+ os.environ['EASYOCR_MODULE_PATH'] = os.path.join(TEMP_DIR, 'easyocr_models')
68
+ try:
69
+ os.makedirs(os.environ['EASYOCR_MODULE_PATH'], exist_ok=True)
70
+ logging.info(f"EasyOCR model directory: {os.environ['EASYOCR_MODULE_PATH']}")
71
+ except Exception as e:
72
+ logging.warning(f"Could not create EasyOCR model directory: {e}")
73
+ # Fallback to /tmp if the temp directory fails
74
+ os.environ['EASYOCR_MODULE_PATH'] = '/tmp/easyocr_models'
75
+ try:
76
+ os.makedirs(os.environ['EASYOCR_MODULE_PATH'], exist_ok=True)
77
+ logging.info(f"Using fallback EasyOCR model directory: {os.environ['EASYOCR_MODULE_PATH']}")
78
+ except Exception as e2:
79
+ logging.error(f"Failed to create EasyOCR model directory: {e2}")
80
+ # Last resort: use current directory
81
+ os.environ['EASYOCR_MODULE_PATH'] = os.path.join(os.getcwd(), 'easyocr_models')
82
+ os.makedirs(os.environ['EASYOCR_MODULE_PATH'], exist_ok=True)
83
+ logging.warning(f"Using current directory for EasyOCR models: {os.environ['EASYOCR_MODULE_PATH']}")
84
+
85
  # Log startup information
86
  logging.info("=" * 50)
87
  logging.info("Docling Streamlit App Starting")
88
  logging.info(f"Temp directory: {TEMP_DIR}")
89
+ logging.info(f"EasyOCR model directory: {os.environ.get('EASYOCR_MODULE_PATH', 'NOT_SET')}")
90
  logging.info(f"Current working directory: {os.getcwd()}")
91
  logging.info(f"Python version: {sys.version}")
92
  logging.info("=" * 50)