Add ollama-init service for auto model pull in docker compose

New ollama-init service pulls nomic-embed-text and qwen2.5-coder:1.5b
before CodeRAG starts. Fix ollama image tag 0.5 → latest.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: maciek-O-digiaidev

docker-compose.yml

@@ -9,6 +9,54 @@
 # ============================================================================
 
 services:
+  # --------------------------------------------------------------------------
+  # Ollama — local LLM / embedding model server
+  # --------------------------------------------------------------------------
+  ollama:
+    image: ollama/ollama:latest
+    container_name: coderag-ollama
+    ports:
+      - "11434:11434"
+    volumes:
+      - ollama-data:/root/.ollama
+    # GPU support (uncomment for NVIDIA GPU)
+    # deploy:
+    #   resources:
+    #     reservations:
+    #       devices:
+    #         - driver: nvidia
+    #           count: all
+    #           capabilities: [gpu]
+    restart: unless-stopped
+    networks:
+      - coderag-net
+    healthcheck:
+      test: ["CMD", "ollama", "list"]
+      interval: 30s
+      timeout: 10s
+      retries: 3
+
+  # --------------------------------------------------------------------------
+  # Ollama Init — pull required models before CodeRAG starts
+  # Uses the same ollama image (has CLI) and shares the model volume.
+  # --------------------------------------------------------------------------
+  ollama-init:
+    image: ollama/ollama:latest
+    container_name: coderag-ollama-init
+    volumes:
+      - ./scripts/ollama-init.sh:/init.sh:ro
+    entrypoint: ["/bin/sh", "/init.sh"]
+    environment:
+      OLLAMA_HOST: http://ollama:11434
+      CODERAG_EMBEDDING_MODEL: ${CODERAG_EMBEDDING_MODEL:-nomic-embed-text}
+      CODERAG_ENRICHMENT_MODEL: ${CODERAG_ENRICHMENT_MODEL:-qwen2.5-coder:1.5b}
+    depends_on:
+      ollama:
+        condition: service_healthy
+    networks:
+      - coderag-net
+    restart: "no"
+
   # --------------------------------------------------------------------------
   # CodeRAG — MCP server + API server + CLI
   # --------------------------------------------------------------------------
@@ -37,6 +85,8 @@ services:
     depends_on:
       ollama:
         condition: service_healthy
+      ollama-init:
+        condition: service_completed_successfully
     networks:
       - coderag-net
     healthcheck:
@@ -49,33 +99,6 @@ services:
     # Default command: start the API server (override for other commands)
     command: ["serve", "--port", "3000"]
 
-  # --------------------------------------------------------------------------
-  # Ollama — local LLM / embedding model server
-  # --------------------------------------------------------------------------
-  ollama:
-    image: ollama/ollama:0.5
-    container_name: coderag-ollama
-    ports:
-      - "11434:11434"
-    volumes:
-      - ollama-data:/root/.ollama
-    # GPU support (uncomment for NVIDIA GPU)
-    # deploy:
-    #   resources:
-    #     reservations:
-    #       devices:
-    #         - driver: nvidia
-    #           count: all
-    #           capabilities: [gpu]
-    restart: unless-stopped
-    networks:
-      - coderag-net
-    healthcheck:
-      test: ["CMD", "ollama", "list"]
-      interval: 30s
-      timeout: 10s
-      retries: 3
-
   # --------------------------------------------------------------------------
   # Viewer — CodeRAG visualization SPA (optional profile)
   # --------------------------------------------------------------------------

scripts/ollama-init.sh

@@ -0,0 +1,31 @@
+#!/bin/sh
+set -e
+
+OLLAMA_HOST="${OLLAMA_HOST:-http://ollama:11434}"
+EMBED_MODEL="${CODERAG_EMBEDDING_MODEL:-nomic-embed-text}"
+ENRICH_MODEL="${CODERAG_ENRICHMENT_MODEL:-qwen2.5-coder:1.5b}"
+
+echo "=== CodeRAG Ollama Init ==="
+echo "Ollama host: $OLLAMA_HOST"
+echo "Embedding model: $EMBED_MODEL"
+echo "Enrichment model: $ENRICH_MODEL"
+echo ""
+
+pull_model() {
+  local model="$1"
+  echo "Checking model: $model"
+  if ollama list 2>/dev/null | grep -q "$model"; then
+    echo "  Already present: $model"
+  else
+    echo "  Pulling $model (this may take a few minutes)..."
+    ollama pull "$model"
+    echo "  Done: $model"
+  fi
+}
+
+pull_model "$EMBED_MODEL"
+pull_model "$ENRICH_MODEL"
+
+echo ""
+echo "All models ready:"
+ollama list