fix: download tesseract traineddata files directly from source

build.sh

@@ -5,50 +5,73 @@ set -e
 
 echo "Starting build process..."
 
-# Install system dependencies for tesseract with more specific packages
+# Install system dependencies for tesseract
 echo "Installing Tesseract and dependencies..."
 apt-get update && apt-get install -y \
     tesseract-ocr \
     tesseract-ocr-eng \
     libtesseract-dev \
     libleptonica-dev \
-    pkg-config
+    pkg-config \
+    wget
 
-# Verify tesseract installation
+# Create tessdata directory
+TESSDATA_DIR="/usr/share/tesseract-ocr/4.00/tessdata"
+mkdir -p "$TESSDATA_DIR"
+
+# Download traineddata files directly from the official repository
+echo "Downloading Tesseract traineddata files..."
+wget -O "$TESSDATA_DIR/eng.traineddata" "https://github.com/tesseract-ocr/tessdata/raw/main/eng.traineddata"
+wget -O "$TESSDATA_DIR/osd.traineddata" "https://github.com/tesseract-ocr/tessdata/raw/main/osd.traineddata"
+
+# Set and verify TESSDATA_PREFIX
+export TESSDATA_PREFIX="$TESSDATA_DIR"
+echo "TESSDATA_PREFIX=${TESSDATA_PREFIX}" >> /etc/environment
+
+# Verify tesseract installation and data files
+echo "Verifying Tesseract installation..."
 if ! command -v tesseract &> /dev/null; then
     echo "Tesseract installation failed!"
     exit 1
 fi
 echo "Tesseract version: $(tesseract --version)"
 
-# Set TESSDATA_PREFIX environment variable
-TESSDATA_PREFIX=$(dpkg -L tesseract-ocr-eng | grep tessdata$)
-if [ -z "$TESSDATA_PREFIX" ]; then
-    echo "Could not find tessdata directory!"
+# Verify traineddata files
+echo "Verifying traineddata files..."
+if [ ! -f "$TESSDATA_DIR/eng.traineddata" ]; then
+    echo "eng.traineddata is missing!"
+    exit 1
+fi
+if [ ! -f "$TESSDATA_DIR/osd.traineddata" ]; then
+    echo "osd.traineddata is missing!"
     exit 1
 fi
-echo "Set TESSDATA_PREFIX=${TESSDATA_PREFIX}"
-export TESSDATA_PREFIX
 
-# Add TESSDATA_PREFIX to environment for persistence
-echo "TESSDATA_PREFIX=${TESSDATA_PREFIX}" >> /etc/environment
+echo "Traineddata files in $TESSDATA_DIR:"
+ls -l "$TESSDATA_DIR"
 
-# Verify tessdata directory
-if [ ! -d "$TESSDATA_PREFIX" ]; then
-    echo "Tessdata directory does not exist!"
+# 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
-echo "Tessdata directory contents:"
-ls -l $TESSDATA_PREFIX
+rm test.png
 
-# Uninstall any existing tesserocr and install from source
+# Clean and install tesserocr from source
 echo "Installing tesserocr from source..."
 pip uninstall -y tesserocr || true
-pip install --no-binary :all: tesserocr
+CPPFLAGS="-I/usr/include/tesseract" LDFLAGS="-L/usr/lib/x86_64-linux-gnu/" pip install --no-binary :all: tesserocr
 
-# Install ocrmac
-echo "Installing ocrmac..."
-pip install ocrmac
+# Verify tesserocr installation
+echo "Verifying tesserocr installation..."
+python3 -c "
+import tesserocr
+print(f'tesserocr version: {tesserocr.__version__}')
+print(f'Available languages: {tesserocr.get_languages()}')
+print(f'TESSDATA_PREFIX: {tesserocr.get_languages()[1]}')
+"
 
 # Install Python dependencies
 echo "Installing Python dependencies..."
@@ -57,7 +80,7 @@ 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!"