Open
Description
Description
when I try to specify function_calling_llm in agents.yaml, got following error:
Traceback (most recent call last):
File "/root/code/ai_learning/agent/latest_ai_development/src/latest_ai_development/main.py", line 100, in <module>
File "/root/code/venv/lib/python3.11/site-packages/crewai/project/crew_base.py", line 35, in __init__
self.map_all_agent_variables()
File "/root/code/venv/lib/python3.11/site-packages/crewai/project/crew_base.py", line 147, in map_all_agent_variables
self._map_agent_variables(
File "/root/code/venv/lib/python3.11/site-packages/crewai/project/crew_base.py", line 179, in _map_agent_variables
self.agents_config[agent_name]["function_calling_llm"] = agents[
^^^^^^^
KeyError: 'gpt-4o-mini'The code seems to be looking for function_calling_llm from the registered agent.
Steps to Reproduce
The complete code is as follows
Expected behavior
Ability to use specify function_calling_llm in agents.yaml
Screenshots/Code snippets
config/agents.yaml
log_qurier:
role: >
LOG Qurier.
goal: >
Read the log file of the target node.
backstory: >
You are an experienced operator who is good at log reading.
You can read the specified log file on the specified node and
capture the log information for a specific time period.
tools: [query_log]
function_calling_llm: "gpt-4o-mini"
log_analyst:
role: >
LOG Analyst
goal: >
Find the log with the keyword and extract the relevant part.
backstory: >
You are a meticulous analyst with a keen eye for detail.
You are known for your ability to find relevant logs based on keywords in complex log files,
making it easy for others to understand and act on the information you provide.config/tasks.yaml
research_task:
description: >
Search for logs in '{log_file}' on host '{remote_host}' within the last {capture_time} minutes.now time is {now_time}.
expected_output: >
Extract all relevant log entries from '{log_file}' within the last {capture_time} minutes.
agent: log_qurier
reporting_task:
description: >
Analyze the context data and expand sections relevant to '{keyword}' into comprehensive report sections. Ensure the report is detailed and includes all pertinent information.
expected_output: >
A self-contained report featuring main topics, each with a dedicated section containing all pertinent information, formatted in markdown without code blocks (```).
agent: log_analystmain.py
#!/usr/bin/env python
import os
import warnings
from typing import Type
from crewai import Agent, Crew, Process, Task
from crewai.project import CrewBase, agent, crew, task, tool
from crewai.tools import BaseTool
from pydantic import BaseModel, Field
warnings.filterwarnings("ignore", category=SyntaxWarning, module="pysbd")
os.environ["OPENAI_API_KEY"] = "EMPTY"
os.environ["OPENAI_API_BASE"] = "http://xxx.xxx.xxx.xxx:8000/v1/"
os.environ["OPENAI_MODEL_NAME"] = "openai//data/QwQ-32B-AWQ"
os.environ['OTEL_SDK_DISABLED'] = "true"
class LOGQurierInput(BaseModel):
host: str = Field(..., description="IP address of the target host.")
now_time: str = Field(..., description="Current timestamp for log query.")
capture_time: int = Field(2, description="Time in minutes before the current timestamp to capture log.")
log_file_path: str = Field(..., description="Full path to the log file for query.")
class LOGQurier(BaseTool):
name: str = "Query LOG"
description: str = "Query the designated log during a specific time period on a specified host."
args_schema: Type[BaseModel] = LOGQurierInput
def _run(self, host: str, log_file_path: str, now_time: str, capture_time: int=2) -> str:
print(f"call args: host: {host}, log_file_path: {log_file_path}, now_time: {now_time}, capture_time: {capture_time}")
return """
[2025-02-28 17:19:37] Input parameters: add_route_ipv4 66.77.88.0/24 1.2.3.254
[2025-02-28 17:19:37] add_route_ipv4 66.77.88.0/24 1.2.3.254
[2025-02-28 17:19:37] Command 'sudo ip route add 66.77.88.0/24 via 1.2.3.254 metric 0'
[2025-02-28 17:19:37] Execute command failed: Error: Nexthop has invalid gateway.
"""
@CrewBase
class LatestAiDevelopment():
"""LatestAiDevelopment crew"""
agents_config = 'config/agents.yaml'
tasks_config = 'config/tasks.yaml'
@tool
def query_log(self):
return LOGQurier()
@agent
def log_qurier(self) -> Agent:
return Agent(
config=self.agents_config['log_qurier'],
verbose=True
)
@agent
def log_analyst(self) -> Agent:
return Agent(
config=self.agents_config['log_analyst'],
verbose=True
)
@task
def research_task(self) -> Task:
return Task(
config=self.tasks_config['research_task'],
)
@task
def reporting_task(self) -> Task:
return Task(
config=self.tasks_config['reporting_task']
)
@crew
def crew(self) -> Crew:
"""Creates the LatestAiDevelopment crew"""
return Crew(
agents=self.agents,
tasks=self.tasks,
process=Process.sequential,
verbose=True,
)
inputs = {
'keyword': 'failed',
'log_file': '/home/tecs/sa.log',
'capture_time': 5,
'now_time': '2025-02-28 17:19:00',
'remote_host': '127.0.0.1'
}
LatestAiDevelopment().crew().kickoff(inputs=inputs)Operating System
Ubuntu 20.04
Python Version
3.11
crewAI Version
0.102.0
crewAI Tools Version
N/A
Virtual Environment
Venv
Evidence
Traceback (most recent call last):
File "/root/code/ai_learning/agent/latest_ai_development/src/latest_ai_development/main.py", line 100, in <module>
File "/root/code/venv/lib/python3.11/site-packages/crewai/project/crew_base.py", line 35, in __init__
self.map_all_agent_variables()
File "/root/code/venv/lib/python3.11/site-packages/crewai/project/crew_base.py", line 147, in map_all_agent_variables
self._map_agent_variables(
File "/root/code/venv/lib/python3.11/site-packages/crewai/project/crew_base.py", line 179, in _map_agent_variables
self.agents_config[agent_name]["function_calling_llm"] = agents[
^^^^^^^
KeyError: 'gpt-4o-mini'Possible Solution
I think it should look for function_calling_llm in the registered llms
Additional context
None