sequenceDiagram.md•2.22 kB
```mermaid
sequenceDiagram
participant User
participant test.py
participant OpenAI API
participant MCPClient (in test.py)
participant whatsapp_server.py (subprocess)
participant Twilio API
User->>+test.py: Run script `python /home/rj/Code/mcp-generated/test.py`
test.py->>test.py: Load .env (OpenAI Key, Twilio details implicitly needed by server)
test.py->>test.py: Define 'send_whatsapp' tool structure
test.py->>+OpenAI API: Send user_prompt + tool definition
OpenAI API-->>-test.py: Return tool_call request ('send_whatsapp', args)
alt Tool Call Requested
test.py->>+MCPClient: Create MCPClient(run_command='python /home/rj/Code/mcp-generated/whatsapp_server.py')
Note over MCPClient, whatsapp_server.py (subprocess): MCPClient starts whatsapp_server.py in a parallel subprocess
MCPClient->>+whatsapp_server.py (subprocess): Start Process
whatsapp_server.py (subprocess)->>whatsapp_server.py (subprocess): Initialize (Load settings, Twilio Client, Define tool)
whatsapp_server.py (subprocess)-->>-MCPClient: Ready (Implicitly, listening on stdio)
test.py->>MCPClient: call_tool('send_whatsapp', **args)
MCPClient->>+whatsapp_server.py (subprocess): Send tool call request via stdio
whatsapp_server.py (subprocess)->>whatsapp_server.py (subprocess): Execute send_whatsapp(to_number, message)
whatsapp_server.py (subprocess)->>+Twilio API: Create message (from_, to, body)
Twilio API-->>-whatsapp_server.py (subprocess): Return message SID or Error
whatsapp_server.py (subprocess)->>whatsapp_server.py (subprocess): Format success/error result string
whatsapp_server.py (subprocess)-->>-MCPClient: Send result string via stdio
MCPClient-->>-test.py: Return result from server
test.py->>User: Print result from MCP Server
Note over MCPClient, whatsapp_server.py (subprocess): MCPClient terminates the subprocess when done
MCPClient->>whatsapp_server.py (subprocess): Terminate Process (Implicit)
deactivate MCPClient
else No Tool Call
test.py->>User: Print OpenAI's text response
end
deactivate test.py
```