From 9cf8d1b07c3893affc455e3e014b30fd600a3350 Mon Sep 17 00:00:00 2001
From: crowetic <jason@crowetic.com>
Date: Mon, 19 May 2025 19:16:39 -0700
Subject: [PATCH] switching back to AIO build of localai

---
 setup-ai-stack.sh | 200 ++++++++++++++++++++++------------------------
 1 file changed, 94 insertions(+), 106 deletions(-)

diff --git a/setup-ai-stack.sh b/setup-ai-stack.sh
index b834951..6004dad 100644
--- a/setup-ai-stack.sh
+++ b/setup-ai-stack.sh
@@ -69,69 +69,30 @@ if [[ "$1" == "--update" ]]; then
   echo "🔧 Running make prepare..."
   make prepare
 
-  echo "🧠 Building LocalAI binary with GPU, localrecall, and localagi..."
-  make build \
-    TAGS="localrecall,localagi,gpu" \
-    BACKENDS="llama-cuda whispercpp stablediffusion ollama" \
-    GPU_SUPPORT="cuda"
+  # echo "🧠 Building LocalAI binary with GPU, localrecall, and localagi..."
+  # make build \
+  #   TAGS="localrecall,localagi,gpu" \
+  #   BACKENDS="llama-cuda whispercpp stablediffusion ollama" \
+  #   GPU_SUPPORT="cuda"
 
 
-  # Download your custom Dockerfile
-  echo "📥 Downloading custom Dockerfile..."
-  curl -fsSL "$DOCKERFILE_URL" -o Dockerfile
+  # # Download your custom Dockerfile
+  # echo "📥 Downloading custom Dockerfile..."
+  # curl -fsSL "$DOCKERFILE_URL" -o Dockerfile
 
-  # Build the container
-  echo "🔨 Building custom LocalAI Docker image..."
-  docker build -t localai:custom .
+  # # Build the container
+  # echo "🔨 Building custom LocalAI Docker image..."
+  # docker build -t localai:custom .
 
-  # Remove existing container
-  docker stop localai 2>/dev/null || true
-  docker rm localai 2>/dev/null || true
+  # # Remove existing container
+  # docker stop localai 2>/dev/null || true
+  # docker rm localai 2>/dev/null || true
 
-  # Create model and config directories if needed
-  mkdir -p ~/ai-stack/localai/models
-  mkdir -p ~/ai-stack/localai/config
-
-  # Run new container
-  docker run -d \
-    --name localai \
-    --gpus all \
-    -p 8080:8080 \
-    -v ~/ai-stack/localai/models:/models \
-    -v ~/ai-stack/localai/config:/config \
-    -e ENABLE_BACKENDS=llama-cuda,ollama \
-    -e INCLUDE_DEFAULT_MODELS=true \
-    -e AUTOLOAD_MODELS=true \
-    -e MODEL_PATH=/models \
-    --restart unless-stopped \
-    localai:custom
-
-  echo "✅ LocalAI (custom) is now running at http://localhost:8080"
-
-
-  #   # 🔁 Updating LocalAI (GPU build)
-  #   if check_container_running "localai"; then
-  #     echo "🧠 Updating LocalAI..."
-  #     docker stop localai
-  #     docker rm localai
-  #   else
-  #     echo "➕ Installing LocalAI..."
-  #   fi
-
-  #   docker pull localai/localai:latest-aio-gpu-nvidia-cuda-12
-
-  #   # Ensure default models.yaml exists
-  #   MODELS_YAML=~/ai-stack/localai/config/models.yaml
-  #   if [ ! -f "$MODELS_YAML" ]; then
-  #     echo "📄 Creating default models.yaml..."
-  # cat > "$MODELS_YAML" <<EOF
-  # - name: phi
-  #   backend: llama-cuda
-  #   parameters:
-  #     model: /models/phi-2.gguf
-  # EOF
-  #   fi
+  # # Create model and config directories if needed
+  # mkdir -p ~/ai-stack/localai/models
+  # mkdir -p ~/ai-stack/localai/config
 
+  # # Run new container
   # docker run -d \
   #   --name localai \
   #   --gpus all \
@@ -143,9 +104,36 @@ if [[ "$1" == "--update" ]]; then
   #   -e AUTOLOAD_MODELS=true \
   #   -e MODEL_PATH=/models \
   #   --restart unless-stopped \
-  #   localai/localai:latest-aio-gpu-nvidia-cuda-12
+  #   localai:custom
 
-  # echo "✅ LocalAI running at http://localhost:8080"
+  # echo "✅ LocalAI (custom) is now running at http://localhost:8080"
+
+
+    # 🔁 Updating LocalAI (GPU build)
+    if check_container_running "localai"; then
+      echo "🧠 Updating LocalAI..."
+      docker stop localai
+      docker rm localai
+    else
+      echo "➕ Installing LocalAI..."
+    fi
+
+    docker pull localai/localai:latest-aio-gpu-nvidia-cuda-12
+
+  docker run -d \
+    --name localai \
+    --gpus all \
+    -p 8080:8080 \
+    -v ~/ai-stack/localai/models:/models \
+    -v ~/ai-stack/localai/config:/config \
+    -e ENABLE_BACKENDS=llama-cuda,ollama \
+    -e INCLUDE_DEFAULT_MODELS=true \
+    -e AUTOLOAD_MODELS=true \
+    -e MODEL_PATH=/models \
+    --restart unless-stopped \
+    localai/localai:latest-aio-gpu-nvidia-cuda-12
+
+  echo "✅ LocalAI running at http://localhost:8080"
 
 
   
@@ -481,18 +469,18 @@ EOF
   fi
 fi
 
-# if [[ "$1" != "--update" ]]; then
-#   if check_container_running "localai"; then
-#     echo "✅ Skipping LocalAI install (already running)."
-#   else
-#     read -p "➕ Install LocalAI (GPU-accelerated)? [y/N]: " localai_prompt
-#     if [[ "$localai_prompt" =~ ^[Yy]$ ]]; then
-#       echo "🧠 Installing LocalAI (NVIDIA CUDA 12)..."
+if [[ "$1" != "--update" ]]; then
+  if check_container_running "localai"; then
+    echo "✅ Skipping LocalAI install (already running)."
+  else
+    read -p "➕ Install LocalAI (GPU-accelerated)? [y/N]: " localai_prompt
+    if [[ "$localai_prompt" =~ ^[Yy]$ ]]; then
+      echo "🧠 Installing LocalAI (NVIDIA CUDA 12)..."
       
-#       mkdir -p ~/ai-stack/localai/models
-#       mkdir -p ~/ai-stack/localai/config
+      mkdir -p ~/ai-stack/localai/models
+      mkdir -p ~/ai-stack/localai/config
 
-#       echo "📄 Creating default models.yaml..."
+      echo "📄 Creating default models.yaml..."
 #       cat > ~/ai-stack/localai/config/models.yaml <<EOF
 # - name: phi
 #   backend: llama-cuda
@@ -500,59 +488,59 @@ fi
 #     model: /models/phi-2.gguf
 # EOF
 
-#       docker pull localai/localai:latest-aio-gpu-nvidia-cuda-12
-
-#       docker run -d \
-#         --name localai \
-#         --gpus all \
-#         -p 8080:8080 \
-#         -v ~/ai-stack/localai/models:/models \
-#         -v ~/ai-stack/localai/config:/config \
-#         -e ENABLE_BACKENDS=llama-cuda,ollama \
-#         -e INCLUDE_DEFAULT_MODELS=true \
-#         -e AUTOLOAD_MODELS=true \
-#         -e MODEL_PATH=/models \
-#         --restart unless-stopped \
-#         localai/localai:latest-aio-gpu-nvidia-cuda-12
-
-#       echo "🌐 LocalAI is now running at http://localhost:8080"
-#     fi
-#   fi
-# fi
-
-if [[ "$1" != "--update" ]]; then
-  if check_container_running "localai"; then
-    echo "✅ Skipping LocalAI install (already running)."
-  else
-    read -p "➕ Install LocalAI (GPU-custom)? [y/N]: " localai_prompt
-    if [[ "$localai_prompt" =~ ^[Yy]$ ]]; then
-      echo "🧠 Installing LocalAI (Custom created localAI container)..."
-      
-      mkdir -p ~/ai-stack/localai/models
-      mkdir -p ~/ai-stack/localai/config
-
-      echo "📦 Using custom built LocalAI image (localai:custom)"
+      docker pull localai/localai:latest-aio-gpu-nvidia-cuda-12
 
       docker run -d \
         --name localai \
         --gpus all \
-        --network ai-stack-net \
         -p 8080:8080 \
         -v ~/ai-stack/localai/models:/models \
-        -v ~/ai-stack/localai/models:/app/models \
-        -e ENABLE_BACKENDS=llama-cuda,whispercpp,stablediffusion,ollama \
+        -v ~/ai-stack/localai/config:/config \
+        -e ENABLE_BACKENDS=llama-cuda,ollama \
         -e INCLUDE_DEFAULT_MODELS=true \
         -e AUTOLOAD_MODELS=true \
         -e MODEL_PATH=/models \
-        -e LOG_LEVEL=debug \
         --restart unless-stopped \
-        localai:custom
+        localai/localai:latest-aio-gpu-nvidia-cuda-12
 
       echo "🌐 LocalAI is now running at http://localhost:8080"
     fi
   fi
 fi
 
+# if [[ "$1" != "--update" ]]; then
+#   if check_container_running "localai"; then
+#     echo "✅ Skipping LocalAI install (already running)."
+#   else
+#     read -p "➕ Install LocalAI (GPU-custom)? [y/N]: " localai_prompt
+#     if [[ "$localai_prompt" =~ ^[Yy]$ ]]; then
+#       echo "🧠 Installing LocalAI (Custom created localAI container)..."
+      
+#       mkdir -p ~/ai-stack/localai/models
+#       mkdir -p ~/ai-stack/localai/config
+
+#       echo "📦 Using custom built LocalAI image (localai:custom)"
+
+#       docker run -d \
+#         --name localai \
+#         --gpus all \
+#         --network ai-stack-net \
+#         -p 8080:8080 \
+#         -v ~/ai-stack/localai/models:/models \
+#         -v ~/ai-stack/localai/models:/app/models \
+#         -e ENABLE_BACKENDS=llama-cuda,whispercpp,stablediffusion,ollama \
+#         -e INCLUDE_DEFAULT_MODELS=true \
+#         -e AUTOLOAD_MODELS=true \
+#         -e MODEL_PATH=/models \
+#         -e LOG_LEVEL=debug \
+#         --restart unless-stopped \
+#         localai:custom
+
+#       echo "🌐 LocalAI is now running at http://localhost:8080"
+#     fi
+#   fi
+# fi
+