Enhance Dockerfile and Streamlit app for EasyOCR directory management

- Updated Dockerfile to create additional temporary directories with proper permissions for EasyOCR.
- Improved Streamlit app to set environment variables for EasyOCR and create necessary directories, enhancing error handling and logging for directory creation.
- Documented changes in TROUBLESHOOTING.md to address potential permission errors and environment variable configurations.

Dockerfile

@@ -10,10 +10,13 @@ RUN apt-get update && apt-get install -y \
     && rm -rf /var/lib/apt/lists/*
 
 # Create necessary directories with proper permissions
-RUN mkdir -p /app/.streamlit /tmp/docling_temp /tmp/easyocr_models && \
+RUN mkdir -p /app/.streamlit /tmp/docling_temp /tmp/easyocr_models /tmp/cache /tmp/config /tmp/data && \
     chmod 755 /app/.streamlit && \
     chmod 777 /tmp/docling_temp && \
-    chmod 777 /tmp/easyocr_models
+    chmod 777 /tmp/easyocr_models && \
+    chmod 777 /tmp/cache && \
+    chmod 777 /tmp/config && \
+    chmod 777 /tmp/data
 
 COPY requirements.txt ./
 COPY src/ ./src/

TROUBLESHOOTING.md

@@ -68,4 +68,31 @@ The app creates these directories:
 The Dockerfile has been updated to:
 - Create necessary directories with proper permissions
 - Copy Streamlit configuration files
-- Set up proper environment variables 
+- Set up proper environment variables 
+
+### EasyOCR Permission Errors
+
+If you encounter EasyOCR permission errors like:
+```
+PermissionError: [Errno 13] Permission denied: '/.EasyOCR'
+```
+
+The app now handles these by:
+1. Setting `EASYOCR_MODULE_PATH` to a writable directory
+2. Setting `HOME`, `USERPROFILE`, and XDG directories to temp locations
+3. Creating all necessary directories with proper permissions
+4. Using fallback directories if the primary ones fail
+
+### Environment Variables
+
+The app automatically sets these environment variables:
+- `STREAMLIT_SERVER_FILE_WATCHER_TYPE=none`
+- `STREAMLIT_SERVER_HEADLESS=true`
+- `STREAMLIT_BROWSER_GATHER_USAGE_STATS=false`
+- `STREAMLIT_SERVER_ENABLE_CORS=false`
+- `STREAMLIT_SERVER_ENABLE_XSRF_PROTECTION=false`
+- `EASYOCR_MODULE_PATH=/tmp/easyocr_models` (or fallback)
+- `HOME=/tmp/docling_temp` (or fallback)
+- `XDG_CACHE_HOME=/tmp/cache` (or fallback)
+- `XDG_CONFIG_HOME=/tmp/config` (or fallback)
+- `XDG_DATA_HOME=/tmp/data` (or fallback) 

src/streamlit_app.py

@@ -82,6 +82,21 @@ except Exception as e:
         os.makedirs(os.environ['EASYOCR_MODULE_PATH'], exist_ok=True)
         logging.warning(f"Using current directory for EasyOCR models: {os.environ['EASYOCR_MODULE_PATH']}")
 
+# Additional EasyOCR environment variables to prevent root directory access
+os.environ['HOME'] = TEMP_DIR  # Set HOME to temp directory
+os.environ['USERPROFILE'] = TEMP_DIR  # For Windows compatibility
+os.environ['XDG_CACHE_HOME'] = os.path.join(TEMP_DIR, 'cache')
+os.environ['XDG_CONFIG_HOME'] = os.path.join(TEMP_DIR, 'config')
+os.environ['XDG_DATA_HOME'] = os.path.join(TEMP_DIR, 'data')
+
+# Create additional directories that EasyOCR might need
+for env_var in ['XDG_CACHE_HOME', 'XDG_CONFIG_HOME', 'XDG_DATA_HOME']:
+    try:
+        os.makedirs(os.environ[env_var], exist_ok=True)
+        logging.info(f"Created directory for {env_var}: {os.environ[env_var]}")
+    except Exception as e:
+        logging.warning(f"Could not create directory for {env_var}: {e}")
+
 # Log startup information
 logging.info("=" * 50)
 logging.info("Docling Streamlit App Starting")