Skip to content

examples.md

Latest commit

 

History

History
262 lines (226 loc) · 4.99 KB

examples.md

File metadata and controls

262 lines (226 loc) · 4.99 KB
Raw

VFB3-MCP Examples

This file contains examples of how to use the VFB3-MCP server tools.

Tool Examples

1. Get Term Info

{
  "method": "tools/call",
  "params": {
    "name": "get_term_info",
    "arguments": {
      "id": "VFB_jrcv0i43"
    }
  }
}

2. Run Query

{
  "method": "tools/call",
  "params": {
    "name": "run_query",
    "arguments": {
      "id": "VFB_00101567",
      "query_type": "PaintedDomains"
    }
  }
}

3. Search Terms

{
  "method": "tools/call",
  "params": {
    "name": "search_terms",
    "arguments": {
      "query": "medulla"
    }
  }
}

4. Search Terms with Filtering

{
  "method": "tools/call",
  "params": {
    "name": "search_terms",
    "arguments": {
      "query": "medulla",
      "filter_types": ["neuron", "adult"],
      "exclude_types": ["deprecated"]
    }
  }
}

5. Search Terms with Minimization

{
  "method": "tools/call",
  "params": {
    "name": "search_terms",
    "arguments": {
      "query": "medulla",
      "minimize_results": true,
      "rows": 20
    }
  }
}

6. Search Terms with Auto Term Info

{
  "method": "tools/call",
  "params": {
    "name": "search_terms",
    "arguments": {
      "query": "antennal lobe",
      "auto_fetch_term_info": true
    }
  }
}

7. Resolve an Entity Name

{
  "method": "tools/call",
  "params": {
    "name": "resolve_entity",
    "arguments": {
      "name": "P{VT054895-GAL4.DBD}"
    }
  }
}

8. Find Stocks for a Feature ID

{
  "method": "tools/call",
  "params": {
    "name": "run_query",
    "arguments": {
      "id": "FBst123456",
      "query_type": "FindStocks"
    }
  }
}

9. Resolve a Split-GAL4 Combination

{
  "method": "tools/call",
  "params": {
    "name": "resolve_combination",
    "arguments": {
      "name": "SS04495" 
    }
  }
}

10. Find Publications for a Split-GAL4 Combination

{
  "method": "tools/call",
  "params": {
    "name": "run_query",
    "arguments": {
      "id": "FBco_0001234",
      "query_type": "FindComboPublications"
    }
  }
}

11. List Connectome Datasets

{
  "method": "tools/call",
  "params": {
    "name": "list_connectome_datasets",
    "arguments": {}
  }
}

12. Query Connectivity

{
  "method": "tools/call",
  "params": {
    "name": "query_connectivity",
    "arguments": {
      "upstream_type": "FBbt_00000001",
      "downstream_type": "FBbt_00000002",
      "weight": 5,
      "group_by_class": true,
      "exclude_dbs": ["hemibrain"]
    }
  }
}

13. Get Ontology Hierarchy

Traverse part_of (region structure) or subclass_of (cell-type taxonomy) for a VFB term. relationship is required; direction defaults to both and max_depth to 1.

{
  "name": "get_hierarchy",
  "arguments": {
    "id": "FBbt_00005801",
    "relationship": "part_of",
    "direction": "descendants",
    "max_depth": 1
  }
}

Returns the direct parts of the mushroom body (FBbt_00005801). Use relationship: "subclass_of" for cell-type taxonomies (e.g. FBbt_00003686, Kenyon cell), and increase max_depth to expand deeper.

Integration with MCP Clients

Claude Desktop

Add to your MCP server configuration:

{
  "mcpServers": {
    "vfb3-mcp": {
      "type": "http",
      "url": "https://vfb3-mcp.virtualflybrain.org",
      "tools": ["*"]
    }
  }
}

Claude Code

Add to your claude.json:

{
  "mcpServers": {
    "vfb3-mcp": {
      "type": "http",
      "url": "https://vfb3-mcp.virtualflybrain.org",
      "tools": ["*"]
    }
  }
}

GitHub Copilot

Configure the MCP server URL in your Copilot settings pointing to https://vfb3-mcp.virtualflybrain.org.

Gemini

To use the Virtual Fly Brain (VFB) Model Context Protocol (MCP) server with AI on Google Search, connect through the Gemini CLI or a custom Python/Node.js client.

The Gemini web interface does not directly support custom MCP integration. Developer tools are needed to connect the two.

Option 1: Using Gemini CLI

The Gemini CLI allows direct registration of remote MCP servers.

Install the CLI: Ensure Node.js is installed, then run:

npm install -g @google/gemini-cli

Add the VFB Server: Use the add command with the VFB URL.

gemini mcp add vfb https://vfb3-mcp.virtualflybrain.org

Verify & Use: Run the CLI by typing gemini. Check the connection with /mcp. AI on Google Search will then call VFB's neuroanatomy and connectivity tools when questions about Drosophila are asked.

Option 2: Using Python

For application development, use the mcp and google-genai libraries to connect.

Setup: pip install google-genai mcp

Implementation: Use a streamable HTTP transport in JSON response mode (e.g. enableJsonResponse: true) to connect to the VFB URL, list its tools, and pass their schemas to the Gemini model as Function Declarations.

Docker Usage

# Build and run
docker-compose up --build

# Or manually
docker build -t vfb3-mcp .
docker run -it vfb3-mcp