需求管理建议服务

cd requirements-advisor

# Create environment file
cp .env.example .env

# Edit .env and add your Voyage API key
nano .env  # or use your preferred editor

# Build the container
docker compose build

# Ingest content into vector store
docker compose run --rm ingestion

# Start the server
docker compose up -d

# Check logs
docker compose logs -f mcp-server

# Ensure content is ingested
docker compose run --rm ingestion

# Start server
docker compose up

# Server available at http://localhost:8000/mcp

# Install Railway CLI
npm install -g @railway/cli

# Login and deploy
railway login
railway init
railway up

# Set environment variables in Railway dashboard
# Get public URL from Railway

# Create render.yaml
# Push to GitHub
# Connect repo in Render dashboard
# Set VOYAGE_API_KEY in environment

# Build and push to container registry
docker build -t requirements-advisor .
docker tag requirements-advisor:latest <your-registry>/requirements-advisor:latest
docker push <your-registry>/requirements-advisor:latest

# Deploy to ECS/Cloud Run/Container Apps

# 1. Start everything
docker compose up

# 2. In another terminal, test the search
docker compose exec mcp-server python -m requirements_advisor.cli test-search "how to write requirements"

# 3. Make changes and rebuild
docker compose up --build

# 1. Install UV if not already installed
curl -LsSf https://astral.sh/uv/install.sh | sh

# 2. Create virtual environment and install dependencies
uv sync

# 3. Set up environment
cp .env.example .env
# Edit .env with your VOYAGE_API_KEY

# 4. Ingest content
uv run requirements-advisor ingest

# 5. Start server
uv run requirements-advisor serve

# 6. Or run directly with Python
source .venv/bin/activate
python -m requirements_advisor.cli serve

# 1. Create virtual environment
python -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate

# 2. Install package
pip install -e .

# 3. Configure and run
cp .env.example .env
# Edit .env
requirements-advisor ingest
requirements-advisor serve

{
  "mcpServers": {
    "requirements-advisor": {
      "url": "http://localhost:8000/mcp"
    }
  }
}

# Add the MCP server
claude mcp add requirements-advisor --url http://localhost:8000/mcp

http://localhost:8000/mcp

"How do I write good functional requirements?"
"What is requirements traceability and why does it matter?"
"Best practices for medical device requirements"
"Define EARS notation"
"What are non-functional requirements?"

# Start the MCP server
requirements-advisor serve [--host 0.0.0.0] [--port 8000]

# Ingest content into vector store
requirements-advisor ingest [--content-dir ./content] [--clear]

# Show configuration and status
requirements-advisor info

# Test a search query
requirements-advisor test-search "your query here" [--top-k 5]

requirements-advisor/
├── pyproject.toml              # Python package configuration
├── Dockerfile                  # Container image definition
├── docker-compose.yml          # Multi-container orchestration
├── .env.example                # Environment template
├── src/requirements_advisor/
│   ├── __init__.py
│   ├── cli.py                  # Typer CLI commands
│   ├── config.py               # Pydantic settings
│   ├── logging.py              # Loguru logging configuration
│   ├── server.py               # FastMCP server + tools
│   ├── embeddings/
│   │   ├── base.py             # Abstract interface
│   │   └── voyage.py           # Voyage AI implementation
│   ├── vectorstore/
│   │   ├── base.py             # Abstract interface
│   │   └── chroma.py           # ChromaDB implementation
│   ├── images/
│   │   ├── base.py             # Image models (CachedImage, ImageIndex)
│   │   └── cache.py            # Image fetching and caching
│   └── ingestion/
│       └── pipeline.py         # Content ingestion
├── tests/                      # Test suite
│   ├── conftest.py             # Pytest fixtures
│   └── test_*.py               # Test modules
├── content/                    # JSONL content files
│   └── requirements_management_guide.jsonl
└── data/                       # Persistent data (gitignored)
    ├── chroma/                 # Vector store
    └── images/                 # Cached images