Add MXCP user configuration for LLM API keys

Added missing MXCP user configuration that was needed for LLM functionality:

- Created deployment/mxcp-user-config.yml.template with:
  - Transport config for Docker (streamable-http on 0.0.0.0:8001)
  - LLM model configurations (OpenAI GPT-4o, GPT-3.5, Anthropic Claude)
  - Environment variable references for API keys

- Updated Dockerfile to:
  - Create /home/mxcp/.mxcp directory
  - Copy user config template to ~/.mxcp/config.yml
  - Set proper ownership for mxcp user

- Updated setup-project.sh to copy user config template
- Updated README.md to document:
  - New template file in customizable components
  - Required OPENAI_API_KEY and ANTHROPIC_API_KEY secrets

This ensures MXCP projects have proper LLM configuration out of the box.

Author: ppolydoras

README.md

@@ -24,6 +24,7 @@ This repository provides standardized deployment infrastructure for MXCP project
 - `config.env.template` - AWS account, ECR repository, service names
 - `mxcp-site-docker.yml.template` - MXCP configuration with project name
 - `profiles-docker.yml.template` - dbt profiles with project name
+- `mxcp-user-config.yml.template` - MXCP user config with LLM API keys
 - `Dockerfile` - Generic container build pattern
 - `start.sh` - Generic container startup script
 - `requirements.txt` - Base MXCP dependencies
@@ -38,6 +39,8 @@ This repository provides standardized deployment infrastructure for MXCP project
    - `AWS_SECRET_ACCESS_KEY` - For deployment access
    - `MXCP_DATA_ACCESS_KEY_ID` - For data download (if needed)
    - `MXCP_DATA_SECRET_ACCESS_KEY` - For data download (if needed)
+   - `OPENAI_API_KEY` - For LLM functionality (optional)
+   - `ANTHROPIC_API_KEY` - For Claude models (optional)
 
 ### Required Tools
 - **Git** - For repository management

deployment/Dockerfile

@@ -27,6 +27,10 @@ COPY . .
 COPY deployment/mxcp-site-docker.yml ./mxcp-site.yml
 COPY deployment/profiles-docker.yml ./profiles.yml
 
+# Create MXCP user configuration directory and copy config
+RUN mkdir -p /home/mxcp/.mxcp
+COPY deployment/mxcp-user-config.yml.template /home/mxcp/.mxcp/config.yml
+
 # PROJECT-SPECIFIC: UAE Business License Data Preparation
 # Squirro can modify this section for their data sources
 ARG MXCP_DATA_ACCESS_KEY_ID
@@ -44,7 +48,7 @@ ENV AWS_ACCESS_KEY_ID=""
 ENV AWS_SECRET_ACCESS_KEY=""
 
 # Ensure directories have correct permissions
-RUN chmod -R 755 ./data ./logs && chown -R mxcp:mxcp ./data ./logs
+RUN chmod -R 755 ./data ./logs && chown -R mxcp:mxcp ./data ./logs /home/mxcp/.mxcp
 
 # Set environment variables
 ENV PYTHONPATH=/app

deployment/mxcp-user-config.yml.template

@@ -0,0 +1,24 @@
+mxcp: 1
+
+# Transport configuration for Docker deployment
+transport:
+  provider: streamable-http
+  http:
+    host: 0.0.0.0
+    port: 8001
+
+# LLM Models Configuration
+# Add your API keys here or use environment variables
+models:
+  gpt-4o:
+    provider: openai
+    api_key: ${OPENAI_API_KEY}
+    model: gpt-4o
+  gpt-3.5-turbo:
+    provider: openai
+    api_key: ${OPENAI_API_KEY}
+    model: gpt-3.5-turbo
+  claude-3-5-sonnet-20241022:
+    provider: anthropic
+    api_key: ${ANTHROPIC_API_KEY}
+    model: claude-3-5-sonnet-20241022

setup-project.sh

@@ -98,6 +98,11 @@ sed -i "s/{{PROJECT_NAME}}/$PROJECT_NAME/g; s/{{AWS_REGION}}/$AWS_REGION/g" depl
 
 print_success "Created deployment/profiles-docker.yml"
 
+# Copy MXCP user configuration
+cp deployment/mxcp-user-config.yml.template deployment/mxcp-user-config.yml
+
+print_success "Created deployment/mxcp-user-config.yml"
+
 # Summary
 echo ""
 print_success "🎉 Project setup complete!"
@@ -113,6 +118,7 @@ echo "- deployment/config.env"
 echo "- justfile"
 echo "- deployment/mxcp-site-docker.yml"
 echo "- deployment/profiles-docker.yml"
+echo "- deployment/mxcp-user-config.yml"
 echo ""
 echo -e "${BLUE}Project: ${GREEN}$PROJECT_NAME${NC}"
 echo -e "${BLUE}Region:  ${GREEN}$AWS_REGION${NC}"