andre/crewAI
1
0
Fork 0
mirror of https://github.com/crewAIInc/crewAI.git synced 2026-01-08 15:48:29 +00:00
Code Issues Actions 4 Packages Projects Releases Wiki Activity

adding manager_llm

Browse Source
This commit is contained in:
João Moura
2024-02-05 20:46:47 -08:00
parent cb09f17fc7
commit d7f77b42bf
8 changed files with 68 additions and 102 deletions
Download Patch File Download Diff File Expand all files Collapse all files

BIN
.cache/plugin/social/e580fe32a1d3f15fc89057d053ae3e52.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 39 KiB

95
docs/getting-started.md
View File

@@ -1,95 +0,0 @@
# Getting Started
To get started with CrewAI, follow these simple steps:
1. **Installation**:
```shell
pip install crewai
```
The example below also uses duckduckgo, so also install that
```shell
pip install duckduckgo-search
```
2. **Setting Up Your Crew**:
```python
import os
from crewai import Agent, Task, Crew, Process
os.environ["OPENAI_API_KEY"] = "YOUR KEY"
# You can choose to use a local model through Ollama for example. See ./docs/llm-connections.md for more information.
# from langchain.llms import Ollama
# ollama_llm = Ollama(model="openhermes")
# Install duckduckgo-search for this example:
# !pip install -U duckduckgo-search
from langchain_community.tools import DuckDuckGoSearchRun
search_tool = DuckDuckGoSearchRun()
# Define your agents with roles and goals
researcher = Agent(
role='Senior Research Analyst',
goal='Uncover cutting-edge developments in AI and data science',
backstory="""You work at a leading tech think tank.
Your expertise lies in identifying emerging trends.
You have a knack for dissecting complex data and presenting
actionable insights.""",
verbose=True,
allow_delegation=False,
tools=[search_tool]
# You can pass an optional llm attribute specifying what mode you wanna use.
# It can be a local model through Ollama / LM Studio or a remote
# model like OpenAI, Mistral, Antrophic of others (https://python.langchain.com/docs/integrations/llms/)
#
# Examples:
# llm=ollama_llm # was defined above in the file
# llm=ChatOpenAI(model_name="gpt-3.5", temperature=0.7)
)
writer = Agent(
role='Tech Content Strategist',
goal='Craft compelling content on tech advancements',
backstory="""You are a renowned Content Strategist, known for
your insightful and engaging articles.
You transform complex concepts into compelling narratives.""",
verbose=True,
allow_delegation=True,
# (optional) llm=ollama_llm
)
# Create tasks for your agents
task1 = Task(
description="""Conduct a comprehensive analysis of the latest advancements in AI in 2024.
Identify key trends, breakthrough technologies, and potential industry impacts.
Your final answer MUST be a full analysis report""",
agent=researcher
)
task2 = Task(
description="""Using the insights provided, develop an engaging blog
post that highlights the most significant AI advancements.
Your post should be informative yet accessible, catering to a tech-savvy audience.
Make it sound cool, avoid complex words so it doesn't sound like AI.
Your final answer MUST be the full blog post of at least 4 paragraphs.""",
agent=writer
)
# Instantiate your crew with a sequential process
crew = Crew(
agents=[researcher, writer],
tasks=[task1, task2],
verbose=2, # You can set it to 1 or 2 to different logging levels
)
# Get your crew to work!
result = crew.kickoff()
print("######################")
print(result)
```
Currently the only supported process is `Process.sequential`, where one task is executed after the other and the outcome of one is passed as extra content into this next.

10
docs/how-to/Creating-a-Crew-and-kick-it-off.md
View File

@@ -4,7 +4,15 @@ description: A step-by-step guide to creating a cohesive CrewAI team for your pr
---
## Introduction
Assembling a crew in CrewAI is akin to casting for a play, where each agent plays a unique role. This guide walks you through creating a crew, assigning roles and tasks, and activating them to work in harmony.
Embarking on your CrewAI journey involves a few straightforward steps to set up your environment and initiate your AI crew. This guide ensures a seamless start.
## Step 0: Installation
Begin by installing CrewAI and any additional packages required for your project. For instance, the `duckduckgo-search` package is used in this example for enhanced search capabilities.
```shell
pip install crewai
pip install duckduckgo-search
```
## Step 1: Assemble Your Agents
Begin by defining your agents with distinct roles and backstories. These elements not only add depth but also guide their task execution and interaction within the crew.

5
docs/how-to/Hierarchical.md
View File

@@ -23,7 +23,11 @@ To utilize the hierarchical process, you must define a crew with a designated ma
!!! note "Tools on the hierarchical process"
For tools when using the hierarchical process, you want to make sure to assign them to the agents instead of the tasks, as the manager will be the one delegating the tasks and the agents will be the ones executing them.
!!! note "Manager LLM"
A manager will be automatically set for the crew, you don't need to define it. You do need to set the `manager_llm` parameter in the crew though.
```python
from langchain_openai import ChatOpenAI
from crewai import Crew, Process, Agent
# Define your agents, no need to define a manager
@@ -42,6 +46,7 @@ writer = Agent(
project_crew = Crew(
tasks=[...], # Tasks that that manager will figure out how to complete
agents=[researcher, writer],
manager_llm=ChatOpenAI(temperature=0, model="gpt-4"), # The manager's LLM that will be used internally
process=Process.hierarchical # Designating the hierarchical approach
)
```

25
docs/index.md
View File

@@ -33,6 +33,21 @@ Cutting-edge framework for orchestrating role-playing, autonomous AI agents. By
<div style="width:30%">
<h2>How-To Guides</h2>
<ul>
<li>
<a href="./how-to/Creating-a-Crew-and-kick-it-off">
Getting Started
</a>
</li>
<li>
<a href="./how-to/how-to/Sequential">
Using Sequential Process
</a>
</li>
<li>
<a href="./how-to/Hierarchical">
Using Hierarchical Process
</a>
</li>
<li>
<a href="./how-to/LLM-Connections">
Connecting to LLMs
@@ -43,11 +58,6 @@ Cutting-edge framework for orchestrating role-playing, autonomous AI agents. By
Customizing Agents
</a>
</li>
<li>
<a href="./how-to/Creating-a-Crew-and-kick-it-off">
Creating a Crew and kick it off
</a>
</li>
<li>
<a href="./how-to/Human-Input-on-Execution">
Human Input on Execution
@@ -58,6 +68,11 @@ Cutting-edge framework for orchestrating role-playing, autonomous AI agents. By
<div style="width:30%">
<h2>Examples</h2>
<ul>
<li>
<a target='_blank' href="https://github.com/joaomdmoura/crewAI-examples/tree/main/prep-for-a-meeting">
Prepare for meetings
</a>
</li>
<li>
<a target='_blank' href="https://github.com/joaomdmoura/crewAI-examples/tree/main/trip_planner">
Trip Planner Crew

3
mkdocs.yml
View File

@@ -126,7 +126,7 @@ nav:
- Processes: 'core-concepts/Processes.md'
- Collaboration: 'core-concepts/Collaboration.md'
- How to Guides:
- Creating a Crew Automation: 'how-to/Creating-a-Crew-and-kick-it-off.md'
- Getting Started: 'how-to/Creating-a-Crew-and-kick-it-off.md'
- Using Sequential Process: 'how-to/Sequential.md'
- Using Hierarchical Process: 'how-to/Hierarchical.md'
- Connecting to any LLM: 'how-to/LLM-Connections.md'
@@ -139,6 +139,7 @@ nav:
- Game Generator: https://github.com/joaomdmoura/crewAI-examples/tree/main/game-builder-crew"
- Drafting emails with LangGraph: https://github.com/joaomdmoura/crewAI-examples/tree/main/CrewAI-LangGraph"
- Landing Page Generator: https://github.com/joaomdmoura/crewAI-examples/tree/main/landing_page_generator"
- Prepare for meetings: https://github.com/joaomdmoura/crewAI-examples/tree/main/prep-for-a-meeting"
extra_css:
- stylesheets/output.css
- stylesheets/extra.css

15
src/crewai/crew.py
View File

@@ -30,6 +30,7 @@ class Crew(BaseModel):
Attributes:
tasks: List of tasks assigned to the crew.
agents: List of agents part of this crew.
manager_llm: The language model that will run manager agent.
process: The process flow that the crew will follow (e.g., sequential).
verbose: Indicates the verbosity level for logging during execution.
config: Configuration settings for the crew.
@@ -47,6 +48,9 @@ class Crew(BaseModel):
agents: List[Agent] = Field(default_factory=list)
process: Process = Field(default=Process.sequential)
verbose: Union[int, bool] = Field(default=0)
manager_llm: Optional[Any] = Field(
description="Language model that will run the agent.", default=None
)
config: Optional[Union[Json, Dict[str, Any]]] = Field(default=None)
id: UUID4 = Field(default_factory=uuid.uuid4, frozen=True)
max_rpm: Optional[int] = Field(
@@ -90,6 +94,17 @@ class Crew(BaseModel):
self._rpm_controller = RPMController(max_rpm=self.max_rpm, logger=self._logger)
return self
@model_validator(mode="after")
def check_manager_llm(self):
"""Validates that the language model is set when using hierarchical process."""
if self.process == Process.hierarchical and not self.manager_llm:
raise PydanticCustomError(
"missing_manager_llm",
"Attribute `manager_llm` is required when using hierarchical process.",
{},
)
return self
@model_validator(mode="after")
def check_config(self):
"""Validates that the crew is properly configured with agents and tasks."""

17
tests/crew_test.py
View File

@@ -2,6 +2,7 @@
import json
import pydantic_core
import pytest
from crewai.agent import Agent
@@ -144,6 +145,8 @@ def test_crew_creation():
@pytest.mark.vcr(filter_headers=["authorization"])
def test_hierarchical_process():
from langchain_openai import ChatOpenAI
task = Task(
description="Come up with a list of 5 interesting ideas to explore for an article, then write one amazing paragraph highlight for each idea that showcases how good an article about this topic could be. Return the list of ideas with their paragraph and your notes.",
)
@@ -151,6 +154,7 @@ def test_hierarchical_process():
crew = Crew(
agents=[researcher, writer],
process=Process.hierarchical,
manager_llm=ChatOpenAI(temperature=0, model="gpt-4"),
tasks=[task],
)
@@ -175,6 +179,19 @@ def test_hierarchical_process():
)
def test_manager_llm_requirement_for_hierarchical_process():
task = Task(
description="Come up with a list of 5 interesting ideas to explore for an article, then write one amazing paragraph highlight for each idea that showcases how good an article about this topic could be. Return the list of ideas with their paragraph and your notes.",
)
with pytest.raises(pydantic_core._pydantic_core.ValidationError):
Crew(
agents=[researcher, writer],
process=Process.hierarchical,
tasks=[task],
)
@pytest.mark.vcr(filter_headers=["authorization"])
def test_crew_with_delegating_agents():
tasks = [