tesseract ocr fix

build.sh

@@ -10,9 +10,14 @@ echo "Installing Tesseract and dependencies..."
 apt-get update && apt-get install -y \
     tesseract-ocr \
     tesseract-ocr-eng \
+    tesseract-ocr-osd \
     libtesseract-dev \
     libleptonica-dev \
-    pkg-config
+    pkg-config \
+    build-essential
+
+# Create tessdata directory if it doesn't exist
+mkdir -p /usr/share/tesseract-ocr/4.00/tessdata
 
 # Verify tesseract installation
 if ! command -v tesseract &> /dev/null; then
@@ -21,34 +26,42 @@ if ! command -v tesseract &> /dev/null; then
 fi
 echo "Tesseract version: $(tesseract --version)"
 
-# Set TESSDATA_PREFIX environment variable
-TESSDATA_PREFIX=$(dpkg -L tesseract-ocr-eng | grep tessdata$)
-if [ -z "$TESSDATA_PREFIX" ]; then
-    echo "Could not find tessdata directory!"
-    exit 1
-fi
+# Set and export TESSDATA_PREFIX
+export TESSDATA_PREFIX="/usr/share/tesseract-ocr/4.00/tessdata"
 echo "Set TESSDATA_PREFIX=${TESSDATA_PREFIX}"
-export TESSDATA_PREFIX
 
 # Add TESSDATA_PREFIX to environment for persistence
 echo "TESSDATA_PREFIX=${TESSDATA_PREFIX}" >> /etc/environment
 
-# Verify tessdata directory
+# Verify tessdata directory and contents
 if [ ! -d "$TESSDATA_PREFIX" ]; then
-    echo "Tessdata directory does not exist!"
-    exit 1
+    echo "Creating tessdata directory..."
+    mkdir -p "$TESSDATA_PREFIX"
 fi
+
 echo "Tessdata directory contents:"
 ls -l $TESSDATA_PREFIX
 
-# Uninstall any existing tesserocr and install from source
-echo "Installing tesserocr from source..."
+# Test Tesseract functionality
+echo "Testing Tesseract functionality..."
+echo "Hello World" > test.png
+if ! tesseract test.png stdout; then
+    echo "Tesseract test failed!"
+    exit 1
+fi
+rm test.png
+
+# Clean any existing tesserocr installation
+echo "Cleaning existing tesserocr installation..."
 pip uninstall -y tesserocr || true
-pip install --no-binary :all: tesserocr
 
-# Install ocrmac
-echo "Installing ocrmac..."
-pip install ocrmac
+# Install tesserocr from source with proper configuration
+echo "Installing tesserocr from source..."
+CPPFLAGS=-I/usr/include/tesseract/ LDFLAGS=-L/usr/lib/x86_64-linux-gnu/ pip install --no-binary :all: tesserocr
+
+# Verify tesserocr installation
+echo "Verifying tesserocr installation..."
+python3 -c "import tesserocr; print(f'tesserocr version: {tesserocr.__version__}')"
 
 # Install Python dependencies
 echo "Installing Python dependencies..."
@@ -57,7 +70,7 @@ pip install -e .
 # Create .env file if it doesn't exist
 if [ ! -f .env ]; then
     echo "Creating .env file..."
-    cp .env.example .env
+    cp .env.example .env || echo "Warning: .env.example not found"
 fi
 
 echo "Build process completed successfully!"

requirements.txt

@@ -1,4 +1,4 @@
-docling==2.18.0
+docling==2.25.0
 gradio==5.14.0
 grpcio-status==1.70.0
 markdown==3.7
@@ -8,8 +8,11 @@ openai==1.61.1
 pipdeptree==2.25.0
 pytesseract==0.3.13
 semchunk==2.2.2
+Pillow>=9.0.0
+numpy>=1.21.0
+# Tesseract dependencies
 tesseract==0.1.3
 tesserocr>=2.5.0; platform_system != "Windows"  # Only install on non-Windows systems
-Pillow>=9.0.0  # Required for image processing
-numpy>=1.21.0  # Required for image processing
-# Use pytesseract instead of tesserocr for cross-platform compatibility
+# Additional dependencies for image processing
+opencv-python-headless>=4.5.0  # Headless version for server environments
+pdf2image>=1.16.0  # For PDF processing

src/main.py

@@ -186,7 +186,13 @@ def main():
             [chatbot, chatbot]
         )
 
-    demo.launch(server_name="0.0.0.0", server_port=7860)
+    demo.launch(
+        server_name="0.0.0.0",
+        server_port=7860,
+        root_path="",
+        show_error=True,
+        share=False  # Explicitly disable sharing on Hugging Face
+    )
 
 
 if __name__ == "__main__":