feat: Initial supervisor implementation

apps/ai_api/eda_ai_api/api/routes/supervisor.py

@@ -1,68 +1,20 @@
 from fastapi import APIRouter
 from eda_ai_api.models.supervisor import SupervisorRequest, SupervisorResponse
-from crewai import Agent, Crew, Task, Process
-from crewai.project import CrewBase, agent, crew, task
 from opportunity_finder.crew import OpportunityFinderCrew
-from crewai import LLM
-import os
+from plugins.supervisor.crew import SupervisorCrew
 
 router = APIRouter()
 
 
-# Use the same Cerebras LLM configuration as in OpportunityFinderCrew
-cerebras_llm = LLM(
-    model="sambanova/Meta-Llama-3.1-70B-Instruct",
-    api_key=os.environ.get("SAMBANOVA_API_KEY"),
-    # base_url="https://api.cerebras.ai/v1",
-    temperature=0.5,
-)
-
-
-@CrewBase
-class SupervisorCrew:
-    @agent
-    def decision_maker(self) -> Agent:
-        return Agent(
-            role="Decision Maker",
-            goal="Analyze user requests and route them to appropriate endpoints",
-            backstory="I am an AI supervisor that determines which endpoint to call based on user messages",
-            allow_delegation=False,
-            llm=cerebras_llm,  # Add explicit LLM configuration
-        )
-
-    @task
-    def route_decision_task(self, message: str) -> Task:
-        return Task(
-            description=f"""
-            Analyze this message: "{message}"
-            Choose between these endpoints:
-            1. discovery - For finding grant opportunities
-            2. heartbeat - For checking system health
-            
-            Return only: "discovery" or "heartbeat"
-            """,
-            agent=self.decision_maker,
-        )
-
-    @crew
-    def crew(self) -> Crew:
-        return Crew(
-            agents=[self.decision_maker],
-            tasks=[self.route_decision_task],
-            process=Process.sequential,
-            verbose=True,
-        )
-
-
 @router.post("/supervisor", response_model=SupervisorResponse)
 async def supervisor_route(request: SupervisorRequest) -> SupervisorResponse:
-    supervisor = SupervisorCrew()
-    decision = supervisor.crew().kickoff(inputs={"message": request.message})
+    decision = SupervisorCrew().crew().kickoff(inputs={"message": request.message})
 
     if decision.lower() == "discovery":
         topics = ["AI", "Technology"]  # Extract topics from message
-        crew = OpportunityFinderCrew()
-        result = crew.crew().kickoff(inputs={"topics": ", ".join(topics)})
+        result = (
+            OpportunityFinderCrew().crew().kickoff(inputs={"topics": ", ".join(topics)})
+        )
     else:
         result = {"is_alive": True}
 

plugins/supervisor/pyproject.toml

@@ -12,6 +12,9 @@ dependencies = [
 supervisor = "supervisor.main:run"
 run_crew = "supervisor.main:run"
 
+[tool.hatch.build.targets.wheel]
+packages = ["src/supervisor"]
+
 [build-system]
 requires = ["hatchling"]
 build-backend = "hatchling.build"