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

# NornicDB Build System # ============================================================================== # Cross-platform Docker image build and deployment # # Architectures: # - arm64-metal : Apple Silicon with Metal acceleration # - amd64-cuda : x86_64 with NVIDIA CUDA acceleration # # Each architecture has two variants: # - Base (BYOM) : Bring Your Own Model - mount models at runtime # - BGE : Pre-packaged with bge-m3.gguf embedding model # # Usage: # make build-arm64-metal # Build base image (no model) # make build-arm64-metal-bge # Build with embedded BGE model # make deploy-arm64-metal # Build + Push base image # make deploy-arm64-metal-bge # Build + Push BGE image # make deploy-all # Deploy both variants for current arch # ============================================================================== # Configuration REGISTRY ?= timothyswt VERSION ?= latest # Docker build flags (use NO_CACHE=1 to force rebuild without cache) # Usage: make build-arm64-metal NO_CACHE=1 ifdef NO_CACHE DOCKER_BUILD_FLAGS := --no-cache else DOCKER_BUILD_FLAGS := endif # Detect OS first (Windows doesn't have uname) ifeq ($(OS),Windows_NT) HOST_OS := windows # Windows is always amd64 for our purposes (or override with HOST_ARCH=arm64) HOST_ARCH ?= amd64 # Windows binaries need .exe extension BIN_EXT := .exe else HOST_OS := $(shell uname -s | tr '[:upper:]' '[:lower:]') # Detect architecture: arm64 (Apple Silicon) or x86_64/amd64 (Intel/AMD) UNAME_M := $(shell uname -m) ifeq ($(UNAME_M),arm64) HOST_ARCH := arm64 else ifeq ($(UNAME_M),aarch64) HOST_ARCH := arm64 else HOST_ARCH := amd64 endif # Unix binaries have no extension BIN_EXT := endif # Image names: nornicdb-{architecture}[-{feature}]:latest IMAGE_ARM64 := $(REGISTRY)/nornicdb-arm64-metal:$(VERSION) IMAGE_ARM64_BGE := $(REGISTRY)/nornicdb-arm64-metal-bge:$(VERSION) IMAGE_ARM64_BGE_HEIMDALL := $(REGISTRY)/nornicdb-arm64-metal-bge-heimdall:$(VERSION) IMAGE_ARM64_HEADLESS := $(REGISTRY)/nornicdb-arm64-metal-headless:$(VERSION) IMAGE_AMD64 := $(REGISTRY)/nornicdb-amd64-cuda:$(VERSION) IMAGE_AMD64_BGE := $(REGISTRY)/nornicdb-amd64-cuda-bge:$(VERSION) IMAGE_AMD64_BGE_HEIMDALL := $(REGISTRY)/nornicdb-amd64-cuda-bge-heimdall:$(VERSION) IMAGE_AMD64_HEADLESS := $(REGISTRY)/nornicdb-amd64-cuda-headless:$(VERSION) IMAGE_AMD64_CPU := $(REGISTRY)/nornicdb-amd64-cpu:$(VERSION) IMAGE_AMD64_CPU_HEADLESS := $(REGISTRY)/nornicdb-amd64-cpu-headless:$(VERSION) LLAMA_CUDA := $(REGISTRY)/llama-cuda-libs:b4785 # Dockerfiles DOCKER_DIR := docker # Model URLs and paths MODELS_DIR := models BGE_MODEL := $(MODELS_DIR)/bge-m3.gguf QWEN_MODEL := $(MODELS_DIR)/qwen2.5-0.5b-instruct-q4_k_m.gguf BGE_URL := https://huggingface.co/gpustack/bge-m3-GGUF/resolve/main/bge-m3-Q4_K_M.gguf QWEN_URL := https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct-GGUF/resolve/main/qwen2.5-0.5b-instruct-q4_k_m.gguf .PHONY: build-arm64-metal build-arm64-metal-bge build-arm64-metal-bge-heimdall build-arm64-metal-headless .PHONY: build-amd64-cuda build-amd64-cuda-bge build-amd64-cuda-bge-heimdall build-amd64-cuda-headless .PHONY: build-amd64-cpu build-amd64-cpu-headless .PHONY: build-all build-arm64-all build-amd64-all .PHONY: push-arm64-metal push-arm64-metal-bge push-arm64-metal-bge-heimdall push-arm64-metal-headless .PHONY: push-amd64-cuda push-amd64-cuda-bge push-amd64-cuda-bge-heimdall push-amd64-cuda-headless .PHONY: push-amd64-cpu push-amd64-cpu-headless .PHONY: deploy-arm64-metal deploy-arm64-metal-bge deploy-arm64-metal-bge-heimdall deploy-arm64-metal-headless .PHONY: deploy-amd64-cuda deploy-amd64-cuda-bge deploy-amd64-cuda-bge-heimdall deploy-amd64-cuda-headless .PHONY: deploy-amd64-cpu deploy-amd64-cpu-headless .PHONY: deploy-all deploy-arm64-all deploy-amd64-all .PHONY: build-llama-cuda push-llama-cuda deploy-llama-cuda .PHONY: build build-localllm build-headless build-localllm-headless test clean images help .PHONY: download-models download-bge download-qwen check-models # ============================================================================== # Model Downloads (Heimdall prerequisites) # ============================================================================== # Create models directory if it doesn't exist $(MODELS_DIR): @mkdir -p $(MODELS_DIR) # Download BGE embedding model if missing download-bge: $(MODELS_DIR) @if [ ! -f "$(BGE_MODEL)" ]; then \ echo "╔══════════════════════════════════════════════════════════════╗"; \ echo "║ Downloading BGE-M3 embedding model... ║"; \ echo "╚══════════════════════════════════════════════════════════════╝"; \ echo "Source: $(BGE_URL)"; \ echo "Target: $(BGE_MODEL)"; \ echo "Size: ~400MB (this may take a few minutes)"; \ curl -L --progress-bar "$(BGE_URL)" -o "$(BGE_MODEL)"; \ echo "✓ Downloaded $(BGE_MODEL)"; \ else \ echo "✓ BGE model already exists: $(BGE_MODEL)"; \ fi # Download Qwen LLM model if missing download-qwen: $(MODELS_DIR) @if [ ! -f "$(QWEN_MODEL)" ]; then \ echo "╔══════════════════════════════════════════════════════════════╗"; \ echo "║ Downloading Qwen2.5-0.5B-Instruct model... ║"; \ echo "╚══════════════════════════════════════════════════════════════╝"; \ echo "Source: $(QWEN_URL)"; \ echo "Target: $(QWEN_MODEL)"; \ echo "Size: ~350MB (this may take a few minutes)"; \ curl -L --progress-bar "$(QWEN_URL)" -o "$(QWEN_MODEL)"; \ echo "✓ Downloaded $(QWEN_MODEL)"; \ else \ echo "✓ Qwen model already exists: $(QWEN_MODEL)"; \ fi # Download both models if missing download-models: download-bge download-qwen @echo "" @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ ✓ All Heimdall models ready ║" @echo "╚══════════════════════════════════════════════════════════════╝" # Check if models exist (without downloading) check-models: @echo "Checking Heimdall models..." @if [ -f "$(BGE_MODEL)" ]; then \ echo "✓ BGE model: $(BGE_MODEL)"; \ else \ echo "✗ BGE model missing: $(BGE_MODEL)"; \ echo " Run: make download-bge"; \ fi @if [ -f "$(QWEN_MODEL)" ]; then \ echo "✓ Qwen model: $(QWEN_MODEL)"; \ else \ echo "✗ Qwen model missing: $(QWEN_MODEL)"; \ echo " Run: make download-qwen"; \ fi # ============================================================================== # Build (local only, no push) # ============================================================================== build-arm64-metal: @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ Building: $(IMAGE_ARM64) [BYOM]" @echo "╚══════════════════════════════════════════════════════════════╝" docker build $(DOCKER_BUILD_FLAGS) --platform linux/arm64 -t $(IMAGE_ARM64) -f $(DOCKER_DIR)/Dockerfile.arm64-metal . build-arm64-metal-bge: @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ Building: $(IMAGE_ARM64_BGE) [with BGE model]" @echo "╚══════════════════════════════════════════════════════════════╝" docker build $(DOCKER_BUILD_FLAGS) --platform linux/arm64 --build-arg EMBED_MODEL=true -t $(IMAGE_ARM64_BGE) -f $(DOCKER_DIR)/Dockerfile.arm64-metal . build-arm64-metal-bge-heimdall: download-models @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ Building: $(IMAGE_ARM64_BGE_HEIMDALL) [BGE + Heimdall]" @echo "╚══════════════════════════════════════════════════════════════╝" docker build $(DOCKER_BUILD_FLAGS) --platform linux/arm64 -t $(IMAGE_ARM64_BGE_HEIMDALL) -f $(DOCKER_DIR)/Dockerfile.arm64-metal-heimdall . build-arm64-metal-headless: @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ Building: $(IMAGE_ARM64_HEADLESS) [headless, no UI]" @echo "╚══════════════════════════════════════════════════════════════╝" docker build $(DOCKER_BUILD_FLAGS) --platform linux/arm64 --build-arg HEADLESS=true -t $(IMAGE_ARM64_HEADLESS) -f $(DOCKER_DIR)/Dockerfile.arm64-metal . build-amd64-cuda: @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ Building: $(IMAGE_AMD64) [BYOM]" @echo "╚══════════════════════════════════════════════════════════════╝" docker build $(DOCKER_BUILD_FLAGS) --platform linux/amd64 -t $(IMAGE_AMD64) -f $(DOCKER_DIR)/Dockerfile.amd64-cuda . build-amd64-cuda-bge: @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ Building: $(IMAGE_AMD64_BGE) [with BGE model]" @echo "╚══════════════════════════════════════════════════════════════╝" docker build $(DOCKER_BUILD_FLAGS) --platform linux/amd64 --build-arg EMBED_MODEL=true -t $(IMAGE_AMD64_BGE) -f $(DOCKER_DIR)/Dockerfile.amd64-cuda . build-amd64-cuda-bge-heimdall: download-models @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ Building: $(IMAGE_AMD64_BGE_HEIMDALL) [BGE + Heimdall]" @echo "╚══════════════════════════════════════════════════════════════╝" docker build $(DOCKER_BUILD_FLAGS) --platform linux/amd64 -t $(IMAGE_AMD64_BGE_HEIMDALL) -f $(DOCKER_DIR)/Dockerfile.amd64-cuda-heimdall . build-amd64-cuda-headless: @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ Building: $(IMAGE_AMD64_HEADLESS) [headless, no UI]" @echo "╚══════════════════════════════════════════════════════════════╝" docker build $(DOCKER_BUILD_FLAGS) --platform linux/amd64 --build-arg HEADLESS=true -t $(IMAGE_AMD64_HEADLESS) -f $(DOCKER_DIR)/Dockerfile.amd64-cuda . build-amd64-cpu: @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ Building: $(IMAGE_AMD64_CPU) [CPU-only, no embeddings]" @echo "╚══════════════════════════════════════════════════════════════╝" docker build $(DOCKER_BUILD_FLAGS) --platform linux/amd64 -t $(IMAGE_AMD64_CPU) -f $(DOCKER_DIR)/Dockerfile.amd64-cpu . build-amd64-cpu-headless: @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ Building: $(IMAGE_AMD64_CPU_HEADLESS) [CPU-only, headless]" @echo "╚══════════════════════════════════════════════════════════════╝" docker build $(DOCKER_BUILD_FLAGS) --platform linux/amd64 --build-arg HEADLESS=true -t $(IMAGE_AMD64_CPU_HEADLESS) -f $(DOCKER_DIR)/Dockerfile.amd64-cpu . # Build both variants for an architecture build-arm64-all: build-arm64-metal build-arm64-metal-bge build-arm64-metal-headless @echo "✓ Built all ARM64 Metal images" build-amd64-all: build-amd64-cuda build-amd64-cuda-bge build-amd64-cuda-headless build-amd64-cpu build-amd64-cpu-headless @echo "✓ Built all AMD64 images (CUDA + CPU)" # Build based on detected host architecture build-all: @echo "Detected architecture: $(HOST_ARCH)" ifeq ($(HOST_ARCH),arm64) @$(MAKE) build-arm64-all else @$(MAKE) build-amd64-all endif @echo "✓ All images for $(HOST_ARCH) built" # ============================================================================== # Push (registry only, assumes already built) # ============================================================================== push-arm64-metal: @echo "→ Pushing $(IMAGE_ARM64)" docker push $(IMAGE_ARM64) push-arm64-metal-bge: @echo "→ Pushing $(IMAGE_ARM64_BGE)" docker push $(IMAGE_ARM64_BGE) push-arm64-metal-bge-heimdall: @echo "→ Pushing $(IMAGE_ARM64_BGE_HEIMDALL)" docker push $(IMAGE_ARM64_BGE_HEIMDALL) push-arm64-metal-headless: @echo "→ Pushing $(IMAGE_ARM64_HEADLESS)" docker push $(IMAGE_ARM64_HEADLESS) push-amd64-cuda: @echo "→ Pushing $(IMAGE_AMD64)" docker push $(IMAGE_AMD64) push-amd64-cuda-bge: @echo "→ Pushing $(IMAGE_AMD64_BGE)" docker push $(IMAGE_AMD64_BGE) push-amd64-cuda-bge-heimdall: @echo "→ Pushing $(IMAGE_AMD64_BGE_HEIMDALL)" docker push $(IMAGE_AMD64_BGE_HEIMDALL) push-amd64-cuda-headless: @echo "→ Pushing $(IMAGE_AMD64_HEADLESS)" docker push $(IMAGE_AMD64_HEADLESS) push-amd64-cpu: @echo "→ Pushing $(IMAGE_AMD64_CPU)" docker push $(IMAGE_AMD64_CPU) push-amd64-cpu-headless: @echo "→ Pushing $(IMAGE_AMD64_CPU_HEADLESS)" docker push $(IMAGE_AMD64_CPU_HEADLESS) # ============================================================================== # Deploy (Build + Push) # ============================================================================== deploy-arm64-metal: build-arm64-metal push-arm64-metal @echo "✓ Deployed $(IMAGE_ARM64)" deploy-arm64-metal-bge: build-arm64-metal-bge push-arm64-metal-bge @echo "✓ Deployed $(IMAGE_ARM64_BGE)" deploy-arm64-metal-bge-heimdall: build-arm64-metal-bge-heimdall push-arm64-metal-bge-heimdall @echo "✓ Deployed $(IMAGE_ARM64_BGE_HEIMDALL)" @echo "🛡️ Heimdall cognitive features enabled - access Bifrost at /bifrost" deploy-arm64-metal-headless: build-arm64-metal-headless push-arm64-metal-headless @echo "✓ Deployed $(IMAGE_ARM64_HEADLESS)" deploy-amd64-cuda: build-amd64-cuda push-amd64-cuda @echo "✓ Deployed $(IMAGE_AMD64)" deploy-amd64-cuda-bge: build-amd64-cuda-bge push-amd64-cuda-bge @echo "✓ Deployed $(IMAGE_AMD64_BGE)" deploy-amd64-cuda-bge-heimdall: build-amd64-cuda-bge-heimdall push-amd64-cuda-bge-heimdall @echo "✓ Deployed $(IMAGE_AMD64_BGE_HEIMDALL)" @echo "🛡️ Heimdall cognitive features enabled - access Bifrost at /bifrost" deploy-amd64-cuda-headless: build-amd64-cuda-headless push-amd64-cuda-headless @echo "✓ Deployed $(IMAGE_AMD64_HEADLESS)" deploy-amd64-cpu: build-amd64-cpu push-amd64-cpu @echo "✓ Deployed $(IMAGE_AMD64_CPU)" deploy-amd64-cpu-headless: build-amd64-cpu-headless push-amd64-cpu-headless @echo "✓ Deployed $(IMAGE_AMD64_CPU_HEADLESS)" # Deploy both variants for an architecture (including headless) deploy-arm64-all: deploy-arm64-metal deploy-arm64-metal-bge deploy-arm64-metal-headless @echo "✓ Deployed all ARM64 Metal images" deploy-amd64-all: deploy-amd64-cuda deploy-amd64-cuda-bge deploy-amd64-cuda-headless deploy-amd64-cpu deploy-amd64-cpu-headless @echo "✓ Deployed all AMD64 images (CUDA + CPU)" # Deploy based on detected host architecture deploy-all: @echo "Detected architecture: $(HOST_ARCH)" ifeq ($(HOST_ARCH),arm64) @$(MAKE) deploy-arm64-all else @$(MAKE) deploy-amd64-all endif @echo "✓ All images for $(HOST_ARCH) deployed" # ============================================================================== # CUDA Prerequisite (one-time build, ~15 min) # ============================================================================== build-llama-cuda: @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ Building CUDA libs (one-time): $(LLAMA_CUDA)" @echo "╚══════════════════════════════════════════════════════════════╝" docker build --platform linux/amd64 -t $(LLAMA_CUDA) -f $(DOCKER_DIR)/Dockerfile.llama-cuda . push-llama-cuda: @echo "→ Pushing $(LLAMA_CUDA)" docker push $(LLAMA_CUDA) deploy-llama-cuda: build-llama-cuda push-llama-cuda @echo "✓ Deployed $(LLAMA_CUDA)" # ============================================================================== # Local Development (native binary, not Docker) # ============================================================================== # Build NornicDB binary + APOC plugins (on supported platforms) build: build-binary build-plugins-if-supported @echo "✓ Build complete: bin/nornicdb + plugins (if supported)" build-binary: go build -o bin/nornicdb$(BIN_EXT) ./cmd/nornicdb # Build plugins only if platform supports Go plugins (Linux/macOS, not Windows) build-plugins-if-supported: ifeq ($(OS),Windows_NT) @echo "Note: Go plugins not supported on Windows, skipping plugin build" else @$(MAKE) plugins endif build-localllm: build-plugins-if-supported CGO_ENABLED=1 go build -tags localllm -o bin/nornicdb$(BIN_EXT) ./cmd/nornicdb # Build without UI (headless mode) build-headless: build-plugins-if-supported go build -tags noui -o bin/nornicdb-headless$(BIN_EXT) ./cmd/nornicdb build-localllm-headless: build-plugins-if-supported CGO_ENABLED=1 go build -tags "localllm noui" -o bin/nornicdb-headless$(BIN_EXT) ./cmd/nornicdb test: go test ./... # ============================================================================== # Cross-Compilation (native binaries for other platforms) # ============================================================================== # Build from macOS for: Linux servers, Raspberry Pi, Windows, etc. # Note: CGO is disabled for cross-compilation (pure Go, no Metal/CUDA) .PHONY: cross-linux-amd64 cross-linux-arm64 cross-rpi cross-rpi-zero cross-windows cross-all # Linux x86_64 (standard servers, VPS, Docker hosts) cross-linux-amd64: @echo "Building for Linux x86_64..." CGO_ENABLED=0 GOOS=linux GOARCH=amd64 go build -o bin/nornicdb-linux-amd64 ./cmd/nornicdb @echo "✓ bin/nornicdb-linux-amd64" # Linux ARM64 (AWS Graviton, newer ARM servers, Jetson) cross-linux-arm64: @echo "Building for Linux ARM64..." CGO_ENABLED=0 GOOS=linux GOARCH=arm64 go build -o bin/nornicdb-linux-arm64 ./cmd/nornicdb @echo "✓ bin/nornicdb-linux-arm64" # Raspberry Pi 4/5, Pi 3B+ 64-bit, Orange Pi, etc. cross-rpi: @echo "Building for Raspberry Pi (64-bit ARM)..." CGO_ENABLED=0 GOOS=linux GOARCH=arm64 go build -o bin/nornicdb-rpi64 ./cmd/nornicdb @echo "✓ bin/nornicdb-rpi64" # Raspberry Pi 2/3/Zero 2 W (32-bit ARMv7) cross-rpi32: @echo "Building for Raspberry Pi (32-bit ARMv7)..." CGO_ENABLED=0 GOOS=linux GOARCH=arm GOARM=7 go build -o bin/nornicdb-rpi32 ./cmd/nornicdb @echo "✓ bin/nornicdb-rpi32" # Raspberry Pi 1/Zero/Zero W (ARMv6) cross-rpi-zero: @echo "Building for Raspberry Pi Zero (ARMv6)..." CGO_ENABLED=0 GOOS=linux GOARCH=arm GOARM=6 go build -o bin/nornicdb-rpi-zero ./cmd/nornicdb @echo "✓ bin/nornicdb-rpi-zero" # Windows x86_64 (CPU only, no embeddings) cross-windows: @echo "Building for Windows x86_64 (CPU-only)..." CGO_ENABLED=0 GOOS=windows GOARCH=amd64 go build -o bin/nornicdb.exe ./cmd/nornicdb @echo "✓ bin/nornicdb.exe" # Windows native builds (must run on Windows) # See: build.bat for all Windows variants cross-windows-native: @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ Windows Native Builds (run on Windows) ║" @echo "╚══════════════════════════════════════════════════════════════╝" @echo "" @echo "Variants available via build.bat:" @echo "" @echo " CPU-only (no embeddings):" @echo " build.bat cpu Smallest (~15MB)" @echo "" @echo " CPU + Local Embeddings:" @echo " build.bat cpu-localllm BYOM (~25MB)" @echo " build.bat cpu-bge With BGE model (~425MB)" @echo "" @echo " CUDA + Local Embeddings (requires NVIDIA GPU):" @echo " build.bat cuda BYOM (~30MB)" @echo " build.bat cuda-bge With BGE model (~430MB)" @echo "" @echo "Prerequisites:" @echo " - All: Go 1.23+" @echo " - localllm/bge: Pre-built llama.cpp libs (build.bat download-libs)" @echo " - cuda: CUDA Toolkit 12.x + VS2022" @echo " - bge: BGE model file (build.bat download-model)" # Build all cross-compilation targets (excludes Windows CUDA which needs native build) cross-all: cross-linux-amd64 cross-linux-arm64 cross-rpi cross-rpi32 cross-rpi-zero cross-windows @echo "" @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ Cross-compilation complete! Binaries in bin/ ║" @echo "╚══════════════════════════════════════════════════════════════╝" @ls -lh bin/nornicdb* # ============================================================================== # Utilities # ============================================================================== images: @echo "Host architecture: $(HOST_ARCH)" @echo "" @echo "ARM64 Metal images:" @echo " $(IMAGE_ARM64) [BYOM]" @echo " $(IMAGE_ARM64_BGE) [BGE embedded]" @echo " $(IMAGE_ARM64_HEADLESS) [headless, no UI]" @echo "" @echo "AMD64 CUDA images:" @echo " $(IMAGE_AMD64) [BYOM]" @echo " $(IMAGE_AMD64_BGE) [BGE embedded]" @echo " $(IMAGE_AMD64_HEADLESS) [headless, no UI]" @echo "" @echo "AMD64 CPU images (no GPU, embeddings disabled):" @echo " $(IMAGE_AMD64_CPU) [CPU-only]" @echo " $(IMAGE_AMD64_CPU_HEADLESS) [CPU-only, headless]" @echo "" @echo "CUDA prerequisite:" @echo " $(LLAMA_CUDA)" # ============================================================================== # APOC Plugin System # ============================================================================== # Go plugins (.so files) that can be dynamically loaded at runtime. # Note: Go plugins only work on Linux and macOS (not Windows). # Plugins must be built with the same Go version as the main binary. PLUGINS_DIR := apoc/built-plugins # Check if plugins are supported on this platform .PHONY: plugin-check plugin-check: ifeq ($(OS),Windows_NT) @echo "Error: Go plugins are not supported on Windows" @echo "Use static linking instead (functions are built into the binary)" @exit 1 else @echo "Platform $(shell uname -s) supports Go plugins" endif # Build APOC plugin .PHONY: plugins plugins: plugin-check plugin-apoc @echo "" @echo "╔══════════════════════════════════════════════════════════════╗" @echo "║ Plugins built successfully! ║" @echo "╚══════════════════════════════════════════════════════════════╝" @ls -lh $(PLUGINS_DIR)/*.so 2>/dev/null || echo "No plugins found" @echo "" @echo "To use: NORNICDB_PLUGINS_DIR=$(PLUGINS_DIR) ./nornicdb" # Plugin source directory PLUGINS_SRC_DIR := apoc/plugin-src # Build APOC plugin (function count determined at runtime from plugin) .PHONY: plugin-apoc plugin-apoc: plugin-check @mkdir -p $(PLUGINS_DIR) @echo "Building APOC plugin..." cd $(PLUGINS_SRC_DIR)/apoc && go build -buildmode=plugin -o ../../../$(PLUGINS_DIR)/apoc.so apoc_plugin.go @echo "Built: $(PLUGINS_DIR)/apoc.so" # Clean plugins .PHONY: plugins-clean plugins-clean: rm -rf $(PLUGINS_DIR) @echo "Cleaned plugin build artifacts" # List available plugins .PHONY: plugins-list plugins-list: @echo "Available Plugin Targets:" @echo "" @echo " make plugins Build APOC plugin" @echo " make plugin-apoc Build APOC plugin" @echo " make plugins-clean Remove built plugins" @echo "" @if [ -d "$(PLUGINS_DIR)" ]; then \ echo "Built plugins:"; \ ls -lh $(PLUGINS_DIR)/*.so 2>/dev/null || echo " (none)"; \ else \ echo "No plugins built yet. Run 'make plugins' to build."; \ fi clean: rm -rf bin/nornicdb bin/nornicdb-headless bin/nornicdb.exe \ bin/nornicdb-linux-amd64 bin/nornicdb-linux-arm64 \ bin/nornicdb-rpi64 bin/nornicdb-rpi32 bin/nornicdb-rpi-zero help: @echo "NornicDB Build System (detected arch: $(HOST_ARCH))" @echo "" @echo "Model Downloads (Heimdall prerequisites):" @echo " make download-models Download both BGE + Qwen models" @echo " make download-bge Download BGE embedding model (~400MB)" @echo " make download-qwen Download Qwen LLM model (~350MB)" @echo " make check-models Check which models are present" @echo "" @echo "Local Development:" @echo " make build Build native binary with UI" @echo " make build-headless Build native binary without UI" @echo " make build-localllm Build with local LLM support" @echo " make build-localllm-headless Build headless with local LLM" @echo "" @echo "Cross-Compilation (from macOS to other platforms):" @echo " make cross-linux-amd64 Linux x86_64 (servers, VPS)" @echo " make cross-linux-arm64 Linux ARM64 (Graviton, Jetson)" @echo " make cross-rpi Raspberry Pi 4/5 (64-bit)" @echo " make cross-rpi32 Raspberry Pi 2/3/Zero 2 W (32-bit)" @echo " make cross-rpi-zero Raspberry Pi Zero/1 (ARMv6)" @echo " make cross-windows Windows x86_64 (CPU-only, cross-compile)" @echo " make cross-windows-native Windows builds (see all variants)" @echo " make cross-all Build ALL platforms (excl. Windows native)" @echo "" @echo "Docker Build (local only):" @echo " make build-arm64-metal Base image (BYOM)" @echo " make build-arm64-metal-bge With embedded BGE model" @echo " make build-arm64-metal-headless Headless (no UI)" @echo " make build-amd64-cuda Base image (BYOM)" @echo " make build-amd64-cuda-bge With embedded BGE model" @echo " make build-amd64-cuda-headless Headless (no UI)" @echo " make build-amd64-cpu CPU-only (no GPU, no embeddings)" @echo " make build-amd64-cpu-headless CPU-only headless" @echo " make build-arm64-all Build all ARM64 variants" @echo " make build-amd64-all Build all AMD64 variants" @echo " make build-all Build all variants for $(HOST_ARCH)" @echo "" @echo "Docker Deploy (build + push):" @echo " make deploy-arm64-metal Deploy base ARM64" @echo " make deploy-arm64-metal-bge Deploy ARM64 with BGE" @echo " make deploy-arm64-metal-headless Deploy ARM64 headless" @echo " make deploy-amd64-cuda Deploy base AMD64" @echo " make deploy-amd64-cuda-bge Deploy AMD64 with BGE" @echo " make deploy-amd64-cuda-headless Deploy AMD64 headless" @echo " make deploy-amd64-cpu Deploy AMD64 CPU-only" @echo " make deploy-amd64-cpu-headless Deploy AMD64 CPU-only headless" @echo " make deploy-arm64-all Deploy all ARM64 variants" @echo " make deploy-amd64-all Deploy all AMD64 variants" @echo " make deploy-all Deploy all variants for $(HOST_ARCH)" @echo "" @echo "CUDA prereq (one-time, run on x86 machine):" @echo " make build-llama-cuda" @echo " make deploy-llama-cuda" @echo "" @echo "Headless mode:" @echo " - Build: -tags noui (excludes UI from binary)" @echo " - Docker: --build-arg HEADLESS=true" @echo " - Runtime: --headless flag or NORNICDB_HEADLESS=true" @echo "" @echo "CPU-only mode (amd64-cpu):" @echo " - No CUDA/GPU support" @echo " - Embeddings disabled by default (NORNICDB_EMBEDDING_PROVIDER=none)" @echo " - Smallest image size" @echo "" @echo "Config: REGISTRY=name VERSION=tag make ..."