[BUG] Hierarchical example from documentation doesn't work

[ernestp]

Description

I'm following this how to documentation to start on hierarchical process https://docs.crewai.com/how-to/hierarchical-process#hierarchical-process

I had to add kickoff call to make it run

from langchain_openai import ChatOpenAI
from crewai import Crew, Process, Agent

# Agents are defined with attributes for backstory, cache, and verbose mode
researcher = Agent(
    role='Researcher',
    goal='Conduct in-depth analysis',
    backstory='Experienced data analyst with a knack for uncovering hidden trends.',
    cache=True,
    verbose=True,
    # tools=[]  # This can be optionally specified; defaults to an empty list
    use_system_prompt=True,  # Enable or disable system prompts for this agent
    max_rpm=30,  # Limit on the number of requests per minute
    max_iter=5  # Maximum number of iterations for a final answer
)
writer = Agent(
    role='Writer',
    goal='Create engaging content',
    backstory='Creative writer passionate about storytelling in technical domains.',
    cache=True,
    verbose=True,
    # tools=[]  # Optionally specify tools; defaults to an empty list
    use_system_prompt=True,  # Enable or disable system prompts for this agent
    max_rpm=30,  # Limit on the number of requests per minute
    max_iter=5  # Maximum number of iterations for a final answer
)

# Establishing the crew with a hierarchical process and additional configurations
project_crew = Crew(
    tasks=[],  # Tasks to be delegated and executed under the manager's supervision
    agents=[researcher, writer],
    manager_llm=ChatOpenAI(temperature=0, model="gpt-4o-mini"),  # Mandatory if manager_agent is not set
    process=Process.hierarchical,  # Specifies the hierarchical management approach
    #respect_context_window=True,  # Enable respect of the context window for tasks
    memory=True,  # Enable memory usage for enhanced task execution
    manager_agent=None,  # Optional: explicitly set a specific agent as manager instead of the manager_llm
    planning=True,  # Enable planning feature for pre-execution strategy
    verbose=True,
)

# Kick off the crew's work
results = project_crew.kickoff()

# Print the results
print("Crew Work Results:")
print(results)

Steps to Reproduce

python test.py

Expected behavior

Crew run and produce result

Screenshots/Code snippets

See evidence

Operating System

macOS Sonoma

Python Version

3.11

crewAI Version

0.74.2

crewAI Tools Version

0.13.2

Virtual Environment

Conda

Evidence

[2024-10-20 20:25:21][INFO]: Planning the crew execution
Traceback (most recent call last):
  File "/Users/user/Documents/GitHub/crewai-updated-tutorial-hierarchical/test.py", line 42, in <module>
    results = project_crew.kickoff()
              ^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/miniconda3/envs/crewai-311/lib/python3.11/site-packages/crewai/crew.py", line 495, in kickoff
    result = self._run_hierarchical_process()
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/miniconda3/envs/crewai-311/lib/python3.11/site-packages/crewai/crew.py", line 602, in _run_hierarchical_process
    return self._execute_tasks(self.tasks)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/miniconda3/envs/crewai-311/lib/python3.11/site-packages/crewai/crew.py", line 707, in _execute_tasks
    return self._create_crew_output(task_outputs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/miniconda3/envs/crewai-311/lib/python3.11/site-packages/crewai/crew.py", line 811, in _create_crew_output
    raise ValueError(
ValueError: Something went wrong. Kickoff should return only one task output.

Possible Solution

None

Additional context

None

[ernestp]

I found that task is missing in this example and added following task:

task = Task(
    description='Write engaging blog post on CrewAI',
    agent=writer,
    expected_output='Markdown formatted text',
)

However in this case execution is fall into endless loop with error:

## Tool Output: 
Error: the Action Input is not a valid key, value dictionary.

[lorenzejay]

@tonykipkemboi can we have a working example on the docs ?

[lorenzejay]

You can define those tasks without an agent defined. @ernestp Yes. just add a couple of tasks there. the manager agent will then decide which agent to execute the task with.