Add Ollama setup to script and use for in HF space

Dockerfile

@@ -21,14 +21,19 @@ RUN touch /app/.env.local
 
 USER root
 RUN apt-get update
-RUN apt-get install -y libgomp1 libcurl4
+RUN apt-get install -y libgomp1 libcurl4 curl
+
+# Install Ollama
+RUN curl -fsSL https://ollama.ai/install.sh | sh
+
+# ensure ollama cache dir exists before adjusting ownership
+RUN mkdir -p /home/user/.ollama && chown -R 1000:1000 /home/user/.ollama
 
 # ensure npm cache dir exists before adjusting ownership
 RUN mkdir -p /home/user/.npm && chown -R 1000:1000 /home/user/.npm
 
 USER user
 
-
 COPY --chown=1000 .env /app/.env
 COPY --chown=1000 entrypoint.sh /app/entrypoint.sh
 COPY --chown=1000 package.json /app/package.json

entrypoint.sh

@@ -14,6 +14,39 @@ if [ "$INCLUDE_DB" = "true" ] ; then
     nohup mongod &
 fi;
 
+# Start Ollama service for HF space (local gpu)
+echo "Starting local Ollama service"
+nohup ollama serve > /tmp/ollama.log 2>&1 &
+OLLAMA_PID=$!
+
+# Wait for Ollama to be ready
+MAX_RETRIES=30
+RETRY_COUNT=0
+until curl -s http://localhost:11434/api/tags > /dev/null 2>&1; do
+    RETRY_COUNT=$((RETRY_COUNT + 1))
+    if [ $RETRY_COUNT -ge $MAX_RETRIES ]; then
+        echo "Ollama failed to start after $MAX_RETRIES attempts"
+        cat /tmp/ollama.log
+        exit 1
+    fi
+    sleep 2
+done
+
+# Pull models, ex.: OLLAMA_MODELS="llama3.1:8b,mistral:7b,codellama:13b"
+OLLAMA_MODELS=${OLLAMA_MODELS:-llama3.1:8b}
+
+IFS=',' read -ra MODEL_ARRAY <<< "$OLLAMA_MODELS"
+for MODEL in "${MODEL_ARRAY[@]}"; do
+    MODEL=$(echo "$MODEL" | xargs) # trim whitespace
+    if ! ollama list | grep -q "$MODEL"; then
+        echo "  Pulling model: $MODEL (this may take several minutes)..."
+        ollama pull "$MODEL"
+        echo "  $MODEL pulled successfully!"
+    else
+        echo "  $MODEL already exists"
+    fi
+done
+
 export PUBLIC_VERSION=$(node -p "require('./package.json').version")
 
 dotenv -e /app/.env -c -- node /app/build/index.js -- --host 0.0.0.0 --port 3000