The client examples provided are vague and the sampling examples, non-existant. The best I could find was:

  async def handle_sampling_message(
      message: types.CreateMessageRequestParams,
  ) -> types.CreateMessageResult:
      return types.CreateMessageResult(
          role="assistant",
          content=types.TextContent(
              type="text",
              text="Hello, world! from model",
          ),
          model="gpt-3.5-turbo",
          stopReason="endTurn",
      )

this barely explains how it works, and worse, doesn't show how to invoke sampling on the server-side. The documentation on the mcp website is even more vague.

Then there seem to be multiple ways to create a client:

• creating an MCPClient class like in https://modelcontextprotocol.io/quickstart/client
• aysnc with ClientSession like in your readme https://github.com/modelcontextprotocol/python-sdk/blob/main/README.md#writing-mcp-clients

I'm just trying to learn, but its confusing