M.I.M.I.R - Multi-agent Intelligent Memory & Insight Repository

#!/bin/bash # Build and publish llama.cpp server for ARM64 with Nomic Embed Multimodal set -e # Configuration IMAGE_NAME="timothyswt/llama-cpp-server-arm64-nomic-multimodal" VERSION="${1:-latest}" DOCKER_USERNAME="${DOCKER_USERNAME:-timothyswt}" MODEL_NAME="nomic-embed-multimodal" echo "🔨 Building llama.cpp server for ARM64 with Nomic Embed Multimodal..." echo "Image: $IMAGE_NAME:latest" echo "" # Target directory for model TARGET_DIR="docker/llama-cpp/models" TARGET_PATH="$TARGET_DIR/${MODEL_NAME}.gguf" # Create models directory if it doesn't exist mkdir -p "$TARGET_DIR" echo "📥 Checking for Nomic Embed Multimodal GGUF model..." echo "" # Check if model already exists if [ -f "$TARGET_PATH" ]; then echo "✅ Model already exists: $TARGET_PATH" MODEL_SIZE=$(du -h "$TARGET_PATH" | cut -f1) echo " Size: $MODEL_SIZE" DOWNLOAD_SUCCESS=true else echo "Model not found, attempting download..." echo "" # Try to download from HuggingFace or other sources # NOTE: As of this script, Nomic Embed Multimodal GGUF may not be publicly available yet # We'll try multiple sources and provide clear error messages DOWNLOAD_SUCCESS=false # Attempt 1: Try HuggingFace (adjust URL when model is published) echo "Attempting to download from HuggingFace..." HUGGINGFACE_URL="https://huggingface.co/nomic-ai/nomic-embed-multimodal-v1-GGUF/resolve/main/nomic-embed-multimodal-v1.Q4_K_M.gguf" if curl -L -f -# -o "$TARGET_PATH" "$HUGGINGFACE_URL" 2>/dev/null; then DOWNLOAD_SUCCESS=true echo "✅ Downloaded from HuggingFace" else echo "⚠️ HuggingFace download failed (model may not be published yet)" fi # Attempt 2: Try alternative quantization if [ "$DOWNLOAD_SUCCESS" = false ]; then echo "" echo "Attempting alternative model source..." ALT_URL="https://huggingface.co/nomic-ai/nomic-embed-vision-v1.5-GGUF/resolve/main/nomic-embed-vision-v1.5.Q4_K_M.gguf" if curl -L -f -# -o "$TARGET_PATH" "$ALT_URL" 2>/dev/null; then DOWNLOAD_SUCCESS=true echo "✅ Downloaded Nomic Embed Vision (alternative)" MODEL_NAME="nomic-embed-vision-v1.5" fi fi # If download failed, provide manual instructions if [ "$DOWNLOAD_SUCCESS" = false ]; then echo "" echo "❌ Automatic download failed. Manual download required." echo "" echo "📋 Please manually download the model GGUF file:" echo "" echo "Option 1 - Nomic Embed Multimodal (RECOMMENDED if available):" echo " Visit: https://huggingface.co/nomic-ai" echo " Look for: nomic-embed-multimodal GGUF files" echo " Download: Q4_K_M or Q8_0 quantization" echo " Save to: $TARGET_PATH" echo "" echo "Option 2 - Nomic Embed Vision (Alternative):" echo " Visit: https://huggingface.co/nomic-ai/nomic-embed-vision-v1.5" echo " Download the GGUF version" echo " Save to: $TARGET_PATH" echo "" echo "Option 3 - Use curl with direct link:" echo " curl -L -o '$TARGET_PATH' [MODEL_URL]" echo "" echo "After downloading, run this script again." exit 1 fi fi # Verify file exists if [ ! -f "$TARGET_PATH" ]; then echo "❌ Model file not found: $TARGET_PATH" exit 1 fi echo "" echo "✅ Model file ready: $TARGET_PATH" MODEL_SIZE=$(du -h "$TARGET_PATH" | cut -f1) echo " Size: $MODEL_SIZE" echo "" # Detect embedding dimensions from model metadata (if possible) echo "🔍 Analyzing model..." # Note: This is a placeholder - actual dimension detection would require parsing GGUF metadata echo " Model: $MODEL_NAME" echo " Note: Update MIMIR_EMBEDDINGS_DIMENSIONS based on actual model specs" echo "" # Build the Docker image echo "🏗️ Building Docker image..." echo "" docker build \ --platform linux/arm64 \ -t "$IMAGE_NAME:$VERSION" \ -t "$IMAGE_NAME:latest" \ -f docker/Dockerfile.multimodal \ . echo "" echo "✅ Build complete!" echo "" echo "🔍 Image details:" docker images | grep llama-cpp-server-arm64-multimodal | head -n 2 echo "" echo "🧪 Testing the image locally..." echo " Starting container on port 8080..." # Start container for testing CONTAINER_ID=$(docker run -d -p 8080:8080 "$IMAGE_NAME:$VERSION") echo " Container ID: $CONTAINER_ID" echo " Waiting for server to start..." # Wait for health check sleep 10 # Test health endpoint if curl -f http://localhost:8080/health > /dev/null 2>&1; then echo " ✅ Health check passed" # Test embeddings endpoint echo " Testing embeddings..." EMBED_RESULT=$(curl -s -X POST http://localhost:8080/v1/embeddings \ -H "Content-Type: application/json" \ -d '{"model": "'"$MODEL_NAME"'", "input": "test embedding"}' || echo "failed") if [ "$EMBED_RESULT" != "failed" ] && [ -n "$EMBED_RESULT" ]; then echo " ✅ Embeddings test passed" # Try to extract dimensions from response DIMENSIONS=$(echo "$EMBED_RESULT" | grep -o '"embedding":\[[^]]*\]' | grep -o ',' | wc -l) if [ -n "$DIMENSIONS" ] && [ "$DIMENSIONS" -gt 0 ]; then DIMENSIONS=$((DIMENSIONS + 1)) echo " 📊 Detected dimensions: $DIMENSIONS" fi else echo " ⚠️ Embeddings test failed (this may be normal during startup)" fi else echo " ⚠️ Health check failed (server may need more time)" fi # Stop test container echo " Stopping test container..." docker stop "$CONTAINER_ID" > /dev/null 2>&1 docker rm "$CONTAINER_ID" > /dev/null 2>&1 echo "" echo "📦 Ready to push to Docker Hub..." echo " (Make sure you're logged in: docker login)" echo "" # Ask for confirmation read -p "Push to Docker Hub? (y/N) " -n 1 -r echo if [[ $REPLY =~ ^[Yy]$ ]]; then docker push "$IMAGE_NAME:$VERSION" if [ "$VERSION" != "latest" ]; then docker push "$IMAGE_NAME:latest" fi echo "✅ Published to Docker Hub!" else echo "⏭️ Skipped push. To push manually:" echo " docker push $IMAGE_NAME:$VERSION" echo " docker push $IMAGE_NAME:latest" fi # Clean up downloaded model echo "" echo "🧹 Cleaning up..." read -p "Remove temporary model file? (y/N) " -n 1 -r echo if [[ $REPLY =~ ^[Yy]$ ]]; then rm -f "$TARGET_PATH" echo "✅ Removed temporary model copy" else echo "⏭️ Keeping model file at: $TARGET_PATH" fi echo "" echo "🎉 Done! To use this image in docker-compose.arm64.yml:" echo " llama-server:" echo " image: $IMAGE_NAME:$VERSION" echo " # ... rest of config ..." echo "" echo " And update environment variables:" echo " - MIMIR_EMBEDDINGS_MODEL=$MODEL_NAME" echo " - MIMIR_EMBEDDINGS_DIMENSIONS=[CHECK_MODEL_DIMENSIONS]" echo "" echo "⚠️ IMPORTANT: Verify embedding dimensions and update Neo4j vector index:" echo " npm run embeddings:check" echo " npm run embeddings:reset # if dimensions changed" echo ""