mirror of
https://github.com/crewAIInc/crewAI.git
synced 2026-05-06 09:42:39 +00:00
Merge branch 'main' into aa
This commit is contained in:
@@ -14,7 +14,7 @@ warnings.filterwarnings(
|
||||
category=UserWarning,
|
||||
module="pydantic.main",
|
||||
)
|
||||
__version__ = "0.100.1"
|
||||
__version__ = "0.102.0"
|
||||
__all__ = [
|
||||
"Agent",
|
||||
"Crew",
|
||||
|
||||
@@ -56,7 +56,8 @@ def test():
|
||||
Test the crew execution and returns the results.
|
||||
"""
|
||||
inputs = {
|
||||
"topic": "AI LLMs"
|
||||
"topic": "AI LLMs",
|
||||
"current_year": str(datetime.now().year)
|
||||
}
|
||||
try:
|
||||
{{crew_name}}().crew().test(n_iterations=int(sys.argv[1]), openai_model_name=sys.argv[2], inputs=inputs)
|
||||
|
||||
@@ -5,7 +5,7 @@ description = "{{name}} using crewAI"
|
||||
authors = [{ name = "Your Name", email = "you@example.com" }]
|
||||
requires-python = ">=3.10,<3.13"
|
||||
dependencies = [
|
||||
"crewai[tools]>=0.100.1,<1.0.0"
|
||||
"crewai[tools]>=0.102.0,<1.0.0"
|
||||
]
|
||||
|
||||
[project.scripts]
|
||||
|
||||
@@ -5,7 +5,7 @@ description = "{{name}} using crewAI"
|
||||
authors = [{ name = "Your Name", email = "you@example.com" }]
|
||||
requires-python = ">=3.10,<3.13"
|
||||
dependencies = [
|
||||
"crewai[tools]>=0.100.1,<1.0.0",
|
||||
"crewai[tools]>=0.102.0,<1.0.0",
|
||||
]
|
||||
|
||||
[project.scripts]
|
||||
|
||||
@@ -5,7 +5,7 @@ description = "Power up your crews with {{folder_name}}"
|
||||
readme = "README.md"
|
||||
requires-python = ">=3.10,<3.13"
|
||||
dependencies = [
|
||||
"crewai[tools]>=0.100.1"
|
||||
"crewai[tools]>=0.102.0"
|
||||
]
|
||||
|
||||
[tool.crewai]
|
||||
|
||||
@@ -1148,19 +1148,24 @@ class Crew(BaseModel):
|
||||
def test(
|
||||
self,
|
||||
n_iterations: int,
|
||||
openai_model_name: Optional[str] = None,
|
||||
eval_llm: Union[str, InstanceOf[LLM]],
|
||||
inputs: Optional[Dict[str, Any]] = None,
|
||||
) -> None:
|
||||
"""Test and evaluate the Crew with the given inputs for n iterations concurrently using concurrent.futures."""
|
||||
test_crew = self.copy()
|
||||
|
||||
eval_llm = create_llm(eval_llm)
|
||||
|
||||
if not eval_llm:
|
||||
raise ValueError("Failed to create LLM instance.")
|
||||
|
||||
self._test_execution_span = test_crew._telemetry.test_execution_span(
|
||||
test_crew,
|
||||
n_iterations,
|
||||
inputs,
|
||||
openai_model_name, # type: ignore[arg-type]
|
||||
eval_llm.model, # type: ignore[arg-type]
|
||||
) # type: ignore[arg-type]
|
||||
evaluator = CrewEvaluator(test_crew, openai_model_name) # type: ignore[arg-type]
|
||||
evaluator = CrewEvaluator(test_crew, eval_llm) # type: ignore[arg-type]
|
||||
|
||||
for i in range(1, n_iterations + 1):
|
||||
evaluator.set_iteration(i)
|
||||
|
||||
@@ -1,4 +1,5 @@
|
||||
import asyncio
|
||||
import copy
|
||||
import inspect
|
||||
import logging
|
||||
from typing import (
|
||||
@@ -394,7 +395,6 @@ class FlowMeta(type):
|
||||
or hasattr(attr_value, "__trigger_methods__")
|
||||
or hasattr(attr_value, "__is_router__")
|
||||
):
|
||||
|
||||
# Register start methods
|
||||
if hasattr(attr_value, "__is_start_method__"):
|
||||
start_methods.append(attr_name)
|
||||
@@ -569,6 +569,9 @@ class Flow(Generic[T], metaclass=FlowMeta):
|
||||
f"Initial state must be dict or BaseModel, got {type(self.initial_state)}"
|
||||
)
|
||||
|
||||
def _copy_state(self) -> T:
|
||||
return copy.deepcopy(self._state)
|
||||
|
||||
@property
|
||||
def state(self) -> T:
|
||||
return self._state
|
||||
@@ -740,6 +743,7 @@ class Flow(Generic[T], metaclass=FlowMeta):
|
||||
event=FlowStartedEvent(
|
||||
type="flow_started",
|
||||
flow_name=self.__class__.__name__,
|
||||
inputs=inputs,
|
||||
),
|
||||
)
|
||||
self._log_flow_event(
|
||||
@@ -803,6 +807,18 @@ class Flow(Generic[T], metaclass=FlowMeta):
|
||||
async def _execute_method(
|
||||
self, method_name: str, method: Callable, *args: Any, **kwargs: Any
|
||||
) -> Any:
|
||||
dumped_params = {f"_{i}": arg for i, arg in enumerate(args)} | (kwargs or {})
|
||||
self.event_emitter.send(
|
||||
self,
|
||||
event=MethodExecutionStartedEvent(
|
||||
type="method_execution_started",
|
||||
method_name=method_name,
|
||||
flow_name=self.__class__.__name__,
|
||||
params=dumped_params,
|
||||
state=self._copy_state(),
|
||||
),
|
||||
)
|
||||
|
||||
result = (
|
||||
await method(*args, **kwargs)
|
||||
if asyncio.iscoroutinefunction(method)
|
||||
@@ -812,6 +828,18 @@ class Flow(Generic[T], metaclass=FlowMeta):
|
||||
self._method_execution_counts[method_name] = (
|
||||
self._method_execution_counts.get(method_name, 0) + 1
|
||||
)
|
||||
|
||||
self.event_emitter.send(
|
||||
self,
|
||||
event=MethodExecutionFinishedEvent(
|
||||
type="method_execution_finished",
|
||||
method_name=method_name,
|
||||
flow_name=self.__class__.__name__,
|
||||
state=self._copy_state(),
|
||||
result=result,
|
||||
),
|
||||
)
|
||||
|
||||
return result
|
||||
|
||||
async def _execute_listeners(self, trigger_method: str, result: Any) -> None:
|
||||
@@ -950,16 +978,6 @@ class Flow(Generic[T], metaclass=FlowMeta):
|
||||
"""
|
||||
try:
|
||||
method = self._methods[listener_name]
|
||||
|
||||
self.event_emitter.send(
|
||||
self,
|
||||
event=MethodExecutionStartedEvent(
|
||||
type="method_execution_started",
|
||||
method_name=listener_name,
|
||||
flow_name=self.__class__.__name__,
|
||||
),
|
||||
)
|
||||
|
||||
sig = inspect.signature(method)
|
||||
params = list(sig.parameters.values())
|
||||
method_params = [p for p in params if p.name != "self"]
|
||||
@@ -971,15 +989,6 @@ class Flow(Generic[T], metaclass=FlowMeta):
|
||||
else:
|
||||
listener_result = await self._execute_method(listener_name, method)
|
||||
|
||||
self.event_emitter.send(
|
||||
self,
|
||||
event=MethodExecutionFinishedEvent(
|
||||
type="method_execution_finished",
|
||||
method_name=listener_name,
|
||||
flow_name=self.__class__.__name__,
|
||||
),
|
||||
)
|
||||
|
||||
# Execute listeners (and possibly routers) of this listener
|
||||
await self._execute_listeners(listener_name, listener_result)
|
||||
|
||||
|
||||
@@ -1,6 +1,8 @@
|
||||
from dataclasses import dataclass, field
|
||||
from datetime import datetime
|
||||
from typing import Any, Optional
|
||||
from typing import Any, Dict, Optional, Union
|
||||
|
||||
from pydantic import BaseModel
|
||||
|
||||
|
||||
@dataclass
|
||||
@@ -15,17 +17,21 @@ class Event:
|
||||
|
||||
@dataclass
|
||||
class FlowStartedEvent(Event):
|
||||
pass
|
||||
inputs: Optional[Dict[str, Any]] = None
|
||||
|
||||
|
||||
@dataclass
|
||||
class MethodExecutionStartedEvent(Event):
|
||||
method_name: str
|
||||
state: Union[Dict[str, Any], BaseModel]
|
||||
params: Optional[Dict[str, Any]] = None
|
||||
|
||||
|
||||
@dataclass
|
||||
class MethodExecutionFinishedEvent(Event):
|
||||
method_name: str
|
||||
state: Union[Dict[str, Any], BaseModel]
|
||||
result: Any = None
|
||||
|
||||
|
||||
@dataclass
|
||||
|
||||
52
src/crewai/flow/state_utils.py
Normal file
52
src/crewai/flow/state_utils.py
Normal file
@@ -0,0 +1,52 @@
|
||||
from datetime import date, datetime
|
||||
from typing import Any
|
||||
|
||||
from pydantic import BaseModel
|
||||
|
||||
from crewai.flow import Flow
|
||||
|
||||
|
||||
def export_state(flow: Flow) -> dict[str, Any]:
|
||||
"""Exports the Flow's internal state as JSON-compatible data structures.
|
||||
|
||||
Performs a one-way transformation of a Flow's state into basic Python types
|
||||
that can be safely serialized to JSON. To prevent infinite recursion with
|
||||
circular references, the conversion is limited to a depth of 5 levels.
|
||||
|
||||
Args:
|
||||
flow: The Flow object whose state needs to be exported
|
||||
|
||||
Returns:
|
||||
dict[str, Any]: The transformed state using JSON-compatible Python
|
||||
types.
|
||||
"""
|
||||
return _to_serializable(flow._state)
|
||||
|
||||
|
||||
def _to_serializable(obj: Any, max_depth: int = 5, _current_depth: int = 0) -> Any:
|
||||
if _current_depth >= max_depth:
|
||||
return repr(obj)
|
||||
|
||||
if isinstance(obj, (str, int, float, bool, type(None))):
|
||||
return obj
|
||||
elif isinstance(obj, (date, datetime)):
|
||||
return obj.isoformat()
|
||||
elif isinstance(obj, (list, tuple, set)):
|
||||
return [_to_serializable(item, max_depth, _current_depth + 1) for item in obj]
|
||||
elif isinstance(obj, dict):
|
||||
return {
|
||||
_to_serializable_key(key): _to_serializable(
|
||||
value, max_depth, _current_depth + 1
|
||||
)
|
||||
for key, value in obj.items()
|
||||
}
|
||||
elif isinstance(obj, BaseModel):
|
||||
return _to_serializable(obj.model_dump(), max_depth, _current_depth + 1)
|
||||
else:
|
||||
return repr(obj)
|
||||
|
||||
|
||||
def _to_serializable_key(key: Any) -> str:
|
||||
if isinstance(key, (str, int)):
|
||||
return str(key)
|
||||
return f"key_{id(key)}_{repr(key)}"
|
||||
@@ -1,11 +1,12 @@
|
||||
from collections import defaultdict
|
||||
|
||||
from pydantic import BaseModel, Field
|
||||
from pydantic import BaseModel, Field, InstanceOf
|
||||
from rich.box import HEAVY_EDGE
|
||||
from rich.console import Console
|
||||
from rich.table import Table
|
||||
|
||||
from crewai.agent import Agent
|
||||
from crewai.llm import LLM
|
||||
from crewai.task import Task
|
||||
from crewai.tasks.task_output import TaskOutput
|
||||
from crewai.telemetry import Telemetry
|
||||
@@ -23,7 +24,7 @@ class CrewEvaluator:
|
||||
|
||||
Attributes:
|
||||
crew (Crew): The crew of agents to evaluate.
|
||||
openai_model_name (str): The model to use for evaluating the performance of the agents (for now ONLY OpenAI accepted).
|
||||
eval_llm (LLM): Language model instance to use for evaluations
|
||||
tasks_scores (defaultdict): A dictionary to store the scores of the agents for each task.
|
||||
iteration (int): The current iteration of the evaluation.
|
||||
"""
|
||||
@@ -32,9 +33,9 @@ class CrewEvaluator:
|
||||
run_execution_times: defaultdict = defaultdict(list)
|
||||
iteration: int = 0
|
||||
|
||||
def __init__(self, crew, openai_model_name: str):
|
||||
def __init__(self, crew, eval_llm: InstanceOf[LLM]):
|
||||
self.crew = crew
|
||||
self.openai_model_name = openai_model_name
|
||||
self.llm = eval_llm
|
||||
self._telemetry = Telemetry()
|
||||
self._setup_for_evaluating()
|
||||
|
||||
@@ -51,7 +52,7 @@ class CrewEvaluator:
|
||||
),
|
||||
backstory="Evaluator agent for crew evaluation with precise capabilities to evaluate the performance of the agents in the crew based on the tasks they have performed",
|
||||
verbose=False,
|
||||
llm=self.openai_model_name,
|
||||
llm=self.llm,
|
||||
)
|
||||
|
||||
def _evaluation_task(
|
||||
@@ -181,7 +182,7 @@ class CrewEvaluator:
|
||||
self.crew,
|
||||
evaluation_result.pydantic.quality,
|
||||
current_task.execution_duration,
|
||||
self.openai_model_name,
|
||||
self.llm.model,
|
||||
)
|
||||
self.tasks_scores[self.iteration].append(evaluation_result.pydantic.quality)
|
||||
self.run_execution_times[self.iteration].append(
|
||||
|
||||
Reference in New Issue
Block a user