metorial-google 1.0.0rc5

Google (Gemini) provider for Metorial

metorial-google

Google (Gemini) provider integration for Metorial.

Installation

pip install metorial-google
# or
uv add metorial-google
# or
poetry add metorial-google

Features

• 🤖 Gemini Integration: Full support for Gemini Pro, Gemini Flash, and other Google AI models
• 📡 Session Management: Automatic tool lifecycle handling
• 🔄 Format Conversion: Converts Metorial tools to Google function declaration format
• ⚡ Async Support: Full async/await support

Supported Models

All Google Gemini models that support function calling:

• gemini-1.5-pro: Most capable Gemini model with 2M context window
• gemini-1.5-flash: Fast and efficient Gemini model
• gemini-pro: Standard Gemini Pro model
• gemini-pro-vision: Gemini Pro with vision capabilities

Usage

Quick Start (Recommended)

import asyncio
import google.generativeai as genai
from metorial import Metorial

async def main():
  # Initialize clients
  metorial = Metorial(api_key="...your-metorial-api-key...") # async by default
  genai.configure(api_key="...your-google-api-key...")
  google_client = genai.GenerativeModel('gemini-pro')
  
  # One-liner chat with automatic session management
  response = await metorial.run(
    "What are the latest commits in the metorial/websocket-explorer repository?",
    "...your-mcp-server-deployment-id...", # can also be list
    google_client,
    model="gemini-pro",
    max_iterations=25
  )
  
  print("Response:", response)

asyncio.run(main())

Streaming Chat

import asyncio
import google.generativeai as genai
from metorial import Metorial
from metorial.types import StreamEventType

async def streaming_example():
  # Initialize clients
  metorial = Metorial(api_key="...your-metorial-api-key...")
  genai.configure(api_key="...your-google-api-key...")
  google_client = genai.GenerativeModel('gemini-pro')
  
  # Streaming chat with real-time responses
  async def stream_action(session):
    messages = [
      {"role": "user", "content": "Explain quantum computing"}
    ]
    
    async for event in metorial.stream(
      google_client, session, messages, 
      model="gemini-pro",
      max_iterations=25
    ):
      if event.type == StreamEventType.CONTENT:
        print(f"🤖 {event.content}", end="", flush=True)
      elif event.type == StreamEventType.TOOL_CALL:
        print(f"\n🔧 Executing {len(event.tool_calls)} tool(s)...")
      elif event.type == StreamEventType.COMPLETE:
        print(f"\n✅ Complete!")
  
  await metorial.with_session("...your-server-deployment-id...", stream_action)

asyncio.run(streaming_example())

Advanced Usage with Session Management

import asyncio
import google.generativeai as genai
from metorial import Metorial
from metorial_google import MetorialGoogleSession

async def main():
  # Initialize clients
  metorial = Metorial(api_key="...your-metorial-api-key...")
  genai.configure(api_key="...your-google-api-key...")
  
  # Create session with your server deployments
  async with metorial.session(["...your-server-deployment-id..."]) as session:
    # Create Google-specific wrapper
    google_session = MetorialGoogleSession(session.tool_manager)
    
    model = genai.GenerativeModel(
      model_name="gemini-pro",
      tools=google_session.tools
    )
    
    response = model.generate_content("What are the latest commits?")
    
    # Handle function calls if present
    if response.candidates[0].content.parts:
      function_calls = [
        part.function_call for part in response.candidates[0].content.parts
        if hasattr(part, 'function_call') and part.function_call
      ]
      
      if function_calls:
        tool_response = await google_session.call_tools(function_calls)
        # Continue conversation with tool_response

asyncio.run(main())

Using Convenience Functions

from metorial_google import build_google_tools, call_google_tools

async def example_with_functions():
  # Get tools in Google format
  tools = build_google_tools(tool_manager)
  
  # Call tools from Google response
  response = await call_google_tools(tool_manager, function_calls)

API Reference

MetorialGoogleSession

Main session class for Google integration.

session = MetorialGoogleSession(tool_manager)

Properties:

• tools: List of tools in Google function declaration format

Methods:

• async call_tools(function_calls): Execute function calls and return user content

build_google_tools(tool_mgr)

Build Google-compatible tool definitions.

Returns: List of tool definitions in Google format

call_google_tools(tool_mgr, function_calls)

Execute function calls from Google response.

Returns: User content with function responses

Tool Format

Tools are converted to Google's function declaration format:

[{
  "function_declarations": [
    {
      "name": "tool_name",
      "description": "Tool description",
      "parameters": {
        "type": "object",
        "properties": {...},
        "required": [...]
      }
    }
  ]
}]

Error Handling

try:
    response = await google_session.call_tools(function_calls)
except Exception as e:
    print(f"Tool execution failed: {e}")

Tool errors are returned as error objects in the response format.

License

MIT License - see LICENSE file for details.