agno 1.8.2__py3-none-any.whl → 2.0.0__py3-none-any.whl

agno/os/routers/evals/utils.py

@@ -0,0 +1,161 @@
+from typing import Optional
+
+from fastapi import HTTPException
+
+from agno.agent.agent import Agent
+from agno.db.base import BaseDb
+from agno.eval.accuracy import AccuracyEval
+from agno.eval.performance import PerformanceEval
+from agno.eval.reliability import ReliabilityEval
+from agno.models.base import Model
+from agno.os.routers.evals.schemas import EvalRunInput, EvalSchema
+from agno.team.team import Team
+
+
+async def run_accuracy_eval(
+    eval_run_input: EvalRunInput,
+    db: BaseDb,
+    agent: Optional[Agent] = None,
+    team: Optional[Team] = None,
+    default_model: Optional[Model] = None,
+) -> EvalSchema:
+    """Run an Accuracy evaluation for the given agent or team"""
+    if not eval_run_input.expected_output:
+        raise HTTPException(status_code=400, detail="expected_output is required for accuracy evaluation")
+
+    accuracy_eval = AccuracyEval(
+        db=db,
+        agent=agent,
+        team=team,
+        input=eval_run_input.input,
+        expected_output=eval_run_input.expected_output,
+        additional_guidelines=eval_run_input.additional_guidelines,
+        additional_context=eval_run_input.additional_context,
+        num_iterations=eval_run_input.num_iterations or 1,
+        name=eval_run_input.name,
+    )
+
+    result = accuracy_eval.run(print_results=False, print_summary=False)
+    if not result:
+        raise HTTPException(status_code=500, detail="Failed to run accuracy evaluation")
+
+    eval_run = EvalSchema.from_accuracy_eval(accuracy_eval=accuracy_eval, result=result)
+
+    if default_model is not None:
+        if agent is not None:
+            agent.model = default_model
+        elif team is not None:
+            team.model = default_model
+
+    return eval_run
+
+
+async def run_performance_eval(
+    eval_run_input: EvalRunInput,
+    db: BaseDb,
+    agent: Optional[Agent] = None,
+    team: Optional[Team] = None,
+    default_model: Optional[Model] = None,
+) -> EvalSchema:
+    """Run a performance evaluation for the given agent or team"""
+    if agent:
+
+        def run_component():  # type: ignore
+            return agent.run(eval_run_input.input)
+
+        model_id = agent.model.id if agent and agent.model else None
+        model_provider = agent.model.provider if agent and agent.model else None
+
+    elif team:
+
+        def run_component():
+            return team.run(eval_run_input.input)
+
+        model_id = team.model.id if team and team.model else None
+        model_provider = team.model.provider if team and team.model else None
+
+    performance_eval = PerformanceEval(
+        db=db,
+        name=eval_run_input.name,
+        func=run_component,
+        num_iterations=eval_run_input.num_iterations or 10,
+        warmup_runs=eval_run_input.warmup_runs,
+        agent_id=agent.id if agent else None,
+        team_id=team.id if team else None,
+        model_id=model_id,
+        model_provider=model_provider,
+    )
+    result = performance_eval.run(print_results=False, print_summary=False)
+    if not result:
+        raise HTTPException(status_code=500, detail="Failed to run performance evaluation")
+
+    eval_run = EvalSchema.from_performance_eval(
+        performance_eval=performance_eval,
+        result=result,
+        agent_id=agent.id if agent else None,
+        team_id=team.id if team else None,
+        model_id=model_id,
+        model_provider=model_provider,
+    )
+
+    if default_model is not None:
+        if agent is not None:
+            agent.model = default_model
+        elif team is not None:
+            team.model = default_model
+
+    return eval_run
+
+
+async def run_reliability_eval(
+    eval_run_input: EvalRunInput,
+    db: BaseDb,
+    agent: Optional[Agent] = None,
+    team: Optional[Team] = None,
+    default_model: Optional[Model] = None,
+) -> EvalSchema:
+    """Run a reliability evaluation for the given agent or team"""
+    if not eval_run_input.expected_tool_calls:
+        raise HTTPException(status_code=400, detail="expected_tool_calls is required for reliability evaluations")
+
+    if agent:
+        agent_response = agent.run(eval_run_input.input)
+        reliability_eval = ReliabilityEval(
+            db=db,
+            name=eval_run_input.name,
+            agent_response=agent_response,
+            expected_tool_calls=eval_run_input.expected_tool_calls,
+        )
+        model_id = agent.model.id if agent and agent.model else None
+        model_provider = agent.model.provider if agent and agent.model else None
+
+    elif team:
+        team_response = team.run(eval_run_input.input)
+        reliability_eval = ReliabilityEval(
+            db=db,
+            name=eval_run_input.name,
+            team_response=team_response,
+            expected_tool_calls=eval_run_input.expected_tool_calls,
+        )
+        model_id = team.model.id if team and team.model else None
+        model_provider = team.model.provider if team and team.model else None
+
+    result = reliability_eval.run(print_results=False)
+    if not result:
+        raise HTTPException(status_code=500, detail="Failed to run reliability evaluation")
+
+    eval_run = EvalSchema.from_reliability_eval(
+        reliability_eval=reliability_eval,
+        result=result,
+        agent_id=agent.id if agent else None,
+        model_id=model_id,
+        model_provider=model_provider,
+    )
+
+    if default_model is not None:
+        if agent is not None:
+            agent.model = default_model
+        elif team is not None:
+            team.model = default_model
+
+    return eval_run

agno/os/routers/knowledge/__init__.py

@@ -0,0 +1,3 @@
+from agno.os.routers.knowledge.knowledge import get_knowledge_router
+
+__all__ = ["get_knowledge_router"]