Compare commits

...

43 Commits

Author SHA1 Message Date
Eduardo Chiarotti
773a96687d Update custom.md 2024-08-06 15:23:41 -03:00
Eduardo Chiarotti
c315a166aa Update issue templates 2024-08-06 15:23:21 -03:00
Eduardo Chiarotti
498e96a419 Update issue templates (#1067)
* Update issue templates

Add both Bug and Feature templates

* Update feature_request.md
2024-08-06 14:47:00 -03:00
Thiago Moretto
c0c59dc932 Merge pull request #1064 from crewAIInc/thiago/pipeline-fix
Fix flaky test due to suppressed error on `on_llm_start` callback
2024-08-05 16:13:19 -03:00
Thiago Moretto
f3b3d321e5 Fix lint issue 2024-08-05 13:34:03 -03:00
Thiago Moretto
67e4433dc2 Fix flaky test due to suppressed error on on_llm_start callback 2024-08-05 13:29:39 -03:00
Rip&Tear
4a7ae8df71 Update LLM-Connections.md (#1039)
* Minor fixes and updates

* minor fixes across docs

* Updated LLM-Connections.md

---------

Co-authored-by: theCyberTech <mattrapidb@gmail.com>
2024-08-02 15:04:52 -03:00
Rip&Tear
09f92122d5 Docs minor fixes (#1035)
* Minor fixes and updates

* minor fixes across docs

---------

Co-authored-by: theCyberTech <mattrapidb@gmail.com>
2024-08-02 15:01:16 -03:00
Lorenze Jay
8118b7b7d6 Feat/sliding context window (#1042)
* patching for non-gpt model

* removal of json_object tool name assignment

* fixed issue for smaller models due to instructions prompt

* fixing for ollama llama3 models

* WIP: generated summary from documents split, could also create memgpt approach

* WIP: need tests but user inputted summarization strategy implemented - handling context window exceeding errors

* rm extra line

* removed type ignores

* added tests

* handling n to summarize prompt

* code cleanup, using click for cli asker

* rm not used class

* better refactor

* reverted poetry lock

* reverted poetry.locl

* improved context window exceeding exception class
2024-08-01 13:15:50 -07:00
João Moura
c93b85ac53 Preparing for new version 2024-07-30 19:21:18 -04:00
Lorenze Jay
6378f6caec WIP fixed mypy src types (#1036) 2024-07-30 10:59:50 -07:00
Eduardo Chiarotti
d824db82a3 feat: Add execution time to both task and testing feature (#1031)
* feat: Add execution time to both task and testing feature

* feat: Remove unused functions

* feat: change test_crew to evalaute_crew to avoid issues with testing libs

* feat: fix tests
2024-07-29 23:17:07 -03:00
Matt Young
de6b597eff telemetry.py - fix typo in comment. (#1020) 2024-07-29 23:03:51 -03:00
Deepak Tammali
6111d05219 docs: Fix crewai-tools package name typo in getting-started docs (#1026) 2024-07-29 23:03:32 -03:00
Monarch Wadia
f83c91d612 Fixed package name typo in pip install command (#1029)
Changed `pip install crewai-tools` to `pip install crewai-tools`
2024-07-29 23:02:48 -03:00
Mackensie Alvarez
c8f360414e Update Start-a-New-CrewAI-Project-Template-Method.md (#1030) 2024-07-29 23:02:18 -03:00
Brandon Hancock (bhancock_ai)
fa4393d77e Add in missing triple quote and execution time to resume agent functionality. (#1025)
* Add in missing triple quote and execution time to resume agent functionality

* Fixing broken kwargs and other issues causing our tests to fail
2024-07-29 14:39:02 -03:00
Rip&Tear
25c314befc Minor fixes and updates (#1019)
Co-authored-by: theCyberTech <mattrapidb@gmail.com>
2024-07-29 03:24:23 -03:00
Rip&Tear
2fe79e68cd Small 404 error fixes (#1018)
* Updated Docs:  New Getting started section + content update / addition

* fixed indentation issue

* Minor updates to fix typos

* Fixed up 404 error on latest commit

---------

Co-authored-by: theCyberTech <the_t3ch@pm.me>
Co-authored-by: theCyberTech <mattrapidb@gmail.com>
2024-07-28 22:01:04 -03:00
Nuraly
37d05a2365 Update Force-Tool-Ouput-as-Result.md (#964)
I think there is some mistake, because there is no such parameter as force_output_result, and as the code shows, the correct parameter result_as_answer is set during agent creation, not task.
2024-07-28 15:41:56 -03:00
Carine Bruyndoncx
0111d261a4 Update Crews.md - correct result variable to crew_output (#972) 2024-07-28 15:40:36 -03:00
Taleb
0a23e1dc13 Performed spell check across the rest of code base, and enahnced the yaml paraser code a little (#895)
* Performed spell check across the entire documentation

Thank you once again!

* Performed spell check across the most of code base
Folders been checked:
- agents
- cli
- memory
- project
- tasks
- telemetry
- tools
- translations

* Trying to add a max_token for the agents, so they limited by number of tokens.

* Performed spell check across the rest of code base, and enahnced the yaml paraser code a little

* Small change in the main agent doc

* Improve _save_file method to handle both dict and str inputs

- Add check for dict type input
- Use json.dump for dict serialization
- Convert non-dict inputs to string
- Remove type ignore comments

---------

Co-authored-by: João Moura <joaomdmoura@gmail.com>
2024-07-28 15:39:54 -03:00
Henri Wenlin
ef5ff71346 feat: add verbose option for printing in ToolUsage (#990) 2024-07-28 15:12:10 -03:00
Samuel Mallet
1697b4cacb Add docs for new parameters to SerperDevTool (#993) 2024-07-28 15:09:55 -03:00
Taleb
6b4710a8d1 Improve _save_file method to handle both dict and str inputs (#1011)
- Add check for dict type input
- Use json.dump for dict serialization
- Convert non-dict inputs to string
- Remove type ignore comments
2024-07-28 15:03:18 -03:00
Lennex Zinyando
6f2a8f08ba Fixes getting started section links (#1016) 2024-07-28 15:02:41 -03:00
João Moura
4e6abf596d updating test 2024-07-28 13:23:03 -04:00
Rip&Tear
9018e2ab6a Docs update (#1008)
* Updated Docs:  New Getting started section + content update / addition

* fixed indentation issue

* Minor updates to fix typos

---------

Co-authored-by: theCyberTech <the_t3ch@pm.me>
2024-07-28 11:55:09 -03:00
ResearchAI
99d023c5f3 Update reset_memories_command.py (#974) 2024-07-26 14:40:47 -07:00
Brandon Hancock (bhancock_ai)
da7d8256eb Json Task Output Truncation with Escape Characters (#1009)
* Fixed special character issue when converting json to models. Added numerous tests to ensure thigns work properly.

* Fix linting error and cleaned up tests

* Fix customer_converter_cls test failure

* Fixed tests. Thank you lorenze for pointing that out. added a few more to ensure converter creation works properly

* Address lorenze feedback

* Fix linting issues
2024-07-26 17:27:01 -04:00
Brandon Hancock (bhancock_ai)
88bffaa0d0 Merge pull request #1012 from crewAIInc/fix/breaking-test-task-eval
fix test due to asserting instructions model_schema change
2024-07-26 16:55:26 -04:00
Lorenze Jay
1159140d9f fix test due to asserting instructions model_schema change 2024-07-26 13:37:44 -07:00
Lorenze Jay
5ac7050f7a Patch/non gpt model pydantic output (#1003)
* patching for non-gpt model

* removal of json_object tool name assignment

* fixed issue for smaller models due to instructions prompt

* fixing for ollama llama3 models

* closing brackets

* removed not used and fixes
2024-07-26 10:57:56 -07:00
Lorenze Jay
8b513de64c hierarchical process unblocked for async tasks (#995)
* WIP: hierarchical unblock for async tasks

* added better test

* update name change

* added more test and crew manager cleanup

* remove prints

* code cleanup, no need to pass manager
2024-07-26 10:55:51 -07:00
Eduardo Chiarotti
144e6d203f feat: add ability to set LLM for AgentPLanner on Crew (#1001)
* feat: add ability to set LLM for AgentPLanner on Crew

* feat: fixes issue on instantiating the ChatOpenAI on the crew

* docs: add docs for the planning_llm new parameter

* docs: change message to ChatOpenAI llm

* feat: add tests
2024-07-26 14:24:29 -03:00
Eduardo Chiarotti
2d2154ed65 feat: add crew Testing/Evaluating feature (#998)
* feat: add crew Testing/evalauting feature

* feat: add docs and add unit test

* feat: improve testing output table

* feat: add tests

* feat: fix type checking issue

* feat: add raise ValueError when testing if output is not the expected

* docs: add docs for Testing

* feat: improve tests and fix some issue

* feat: back to sync

* feat: change opdeai model

* feat: fix test
2024-07-26 14:23:51 -03:00
Brandon Hancock (bhancock_ai)
2d086ab596 Merge pull request #994 from crewAIInc/fix/getting-started-docs
fixed bullet points for crew yaml annoations
2024-07-23 14:36:45 -04:00
Lorenze Jay
776c67cc0f clearer usage for crewai create command 2024-07-23 11:32:25 -07:00
Lorenze Jay
78ef490646 fixed bullet points for crew yaml annoations 2024-07-23 11:31:09 -07:00
Lorenze Jay
4da5cc9778 Feat yaml config all attributes (#985)
* WIP: yaml proper mapping for agents and agent

* WIP: added output_json and output_pydantic setup

* WIP: core logic added, need cleanup

* code cleanup

* updated docs and example template to use yaml to reference agents within tasks

* cleanup type errors

* Update Start-a-New-CrewAI-Project.md

---------

Co-authored-by: João Moura <joaomdmoura@gmail.com>
2024-07-23 00:21:01 -03:00
Eduardo Chiarotti
6930656897 feat: add crewai test feature (#984)
* feat: add crewai test feature

* fix: remove unused import

* feat: update docstirng

* fix: tests
2024-07-22 17:21:05 -03:00
João Moura
349753a013 prepping new version 2024-07-20 12:26:32 -04:00
Eduardo Chiarotti
f53a3a00e1 fix: planning feature output (#969)
* fix: planning feature output

* fix: add validation for planning result
2024-07-20 11:56:53 -03:00
72 changed files with 438514 additions and 7861 deletions

35
.github/ISSUE_TEMPLATE/bug_report.md vendored Normal file
View File

@@ -0,0 +1,35 @@
---
name: Bug report
about: Create a report to help us improve CrewAI
title: "[BUG]"
labels: bug
assignees: ''
---
**Description**
Provide a clear and concise description of what the bug is.
**Steps to Reproduce**
Provide a step-by-step process to reproduce the behavior:
**Expected behavior**
A clear and concise description of what you expected to happen.
**Screenshots/Code snippets**
If applicable, add screenshots or code snippets to help explain your problem.
**Environment Details:**
- **Operating System**: [e.g., Ubuntu 20.04, macOS Catalina, Windows 10]
- **Python Version**: [e.g., 3.8, 3.9, 3.10]
- **crewAI Version**: [e.g., 0.30.11]
- **crewAI Tools Version**: [e.g., 0.2.6]
**Logs**
Include relevant logs or error messages if applicable.
**Possible Solution**
Have a solution in mind? Please suggest it here, or write "None".
**Additional context**
Add any other context about the problem here.

24
.github/ISSUE_TEMPLATE/custom.md vendored Normal file
View File

@@ -0,0 +1,24 @@
---
name: Custom issue template
about: Describe this issue template's purpose here.
title: "[DOCS]"
labels: documentation
assignees: ''
---
## Documentation Page
<!-- Provide a link to the documentation page that needs improvement -->
## Description
<!-- Describe what needs to be changed or improved in the documentation -->
## Suggested Changes
<!-- If possible, provide specific suggestions for how to improve the documentation -->
## Additional Context
<!-- Add any other context about the documentation issue here -->
## Checklist
- [ ] I have searched the existing issues to make sure this is not a duplicate
- [ ] I have checked the latest version of the documentation to ensure this hasn't been addressed

View File

@@ -254,7 +254,7 @@ pip install dist/*.tar.gz
CrewAI uses anonymous telemetry to collect usage data with the main purpose of helping us improve the library by focusing our efforts on the most used features, integrations and tools.
There is NO data being collected on the prompts, tasks descriptions agents backstories or goals nor tools usage, no API calls, nor responses nor any data that is being processed by the agents, nor any secrets and env vars.
It's pivotal to understand that **NO data is collected** concerning prompts, task descriptions, agents' backstories or goals, usage of tools, API calls, responses, any data processed by the agents, or secrets and environment variables, with the exception of the conditions mentioned. When the `share_crew` feature is enabled, detailed data including task descriptions, agents' backstories or goals, and other specific attributes are collected to provide deeper insights while respecting user privacy. We don't offer a way to disable it now, but we will in the future.
Data collected includes:
@@ -279,7 +279,7 @@ Data collected includes:
- Tools names available
- Understand out of the publically available tools, which ones are being used the most so we can improve them
Users can opt-in sharing the complete telemetry data by setting the `share_crew` attribute to `True` on their Crews.
Users can opt-in to Further Telemetry, sharing the complete telemetry data by setting the `share_crew` attribute to `True` on their Crews. Enabling `share_crew` results in the collection of detailed crew and task execution data, including `goal`, `backstory`, `context`, and `output` of tasks. This enables a deeper insight into usage patterns while respecting the user's choice to share.
## License

View File

@@ -114,7 +114,7 @@ from langchain.agents import load_tools
langchain_tools = load_tools(["google-serper"], llm=llm)
agent1 = CustomAgent(
role="backstory agent",
role="agent role",
goal="who is {input}?",
backstory="agent backstory",
verbose=True,
@@ -127,7 +127,7 @@ task1 = Task(
)
agent2 = Agent(
role="bio agent",
role="agent role",
goal="summarize the short bio for {input} and if needed do more research",
backstory="agent backstory",
verbose=True,

View File

@@ -33,6 +33,7 @@ A crew in crewAI represents a collaborative group of agents working together to
| **Manager Callbacks** _(optional)_ | `manager_callbacks` | `manager_callbacks` takes a list of callback handlers to be executed by the manager agent when a hierarchical process is used. |
| **Prompt File** _(optional)_ | `prompt_file` | Path to the prompt JSON file to be used for the crew. |
| **Planning** *(optional)* | `planning` | Adds planning ability to the Crew. When activated before each Crew iteration, all Crew data is sent to an AgentPlanner that will plan the tasks and this plan will be added to each task description.
| **Planning LLM** *(optional)* | `planning_llm` | The language model used by the AgentPlanner in a planning process. |
!!! note "Crew Max RPM"
The `max_rpm` attribute sets the maximum number of requests per minute the crew can perform to avoid rate limits and will override individual agents' `max_rpm` settings if you set it.
@@ -136,7 +137,7 @@ crew = Crew(
verbose=2
)
result = crew.kickoff()
crew_output = crew.kickoff()
# Accessing the crew output
print(f"Raw Output: {crew_output.raw}")
@@ -220,7 +221,7 @@ These methods provide flexibility in how you manage and execute tasks within you
### Replaying from specific task:
You can now replay from a specific task using our cli command replay.
The replay_from_tasks feature in CrewAI allows you to replay from a specific task using the command-line interface (CLI). By running the command `crewai replay -t <task_id>`, you can specify the `task_id` for the replay process.
The replay feature in CrewAI allows you to replay from a specific task using the command-line interface (CLI). By running the command `crewai replay -t <task_id>`, you can specify the `task_id` for the replay process.
Kickoffs will now save the latest kickoffs returned task outputs locally for you to be able to replay from.

View File

@@ -23,6 +23,25 @@ my_crew = Crew(
From this point on, your crew will have planning enabled, and the tasks will be planned before each iteration.
#### Planning LLM
Now you can define the LLM that will be used to plan the tasks. You can use any ChatOpenAI LLM model available.
```python
from crewai import Crew, Agent, Task, Process
from langchain_openai import ChatOpenAI
# Assemble your crew with planning capabilities and custom LLM
my_crew = Crew(
agents=self.agents,
tasks=self.tasks,
process=Process.sequential,
planning=True,
planning_llm=ChatOpenAI(model="gpt-4o")
)
```
### Example
When running the base case example, you will see something like the following output, which represents the output of the AgentPlanner responsible for creating the step-by-step logic to add to the Agents tasks.

View File

@@ -0,0 +1,41 @@
---
title: crewAI Testing
description: Learn how to test your crewAI Crew and evaluate their performance.
---
## Introduction
Testing is a crucial part of the development process, and it is essential to ensure that your crew is performing as expected. And with crewAI, you can easily test your crew and evaluate its performance using the built-in testing capabilities.
### Using the Testing Feature
We added the CLI command `crewai test` to make it easy to test your crew. This command will run your crew for a specified number of iterations and provide detailed performance metrics.
The parameters are `n_iterations` and `model` which are optional and default to 2 and `gpt-4o-mini` respectively. For now the only provider available is OpenAI.
```bash
crewai test
```
If you want to run more iterations or use a different model, you can specify the parameters like this:
```bash
crewai test --n_iterations 5 --model gpt-4o
```
What happens when you run the `crewai test` command is that the crew will be executed for the specified number of iterations, and the performance metrics will be displayed at the end of the run.
A table of scores at the end will show the performance of the crew in terms of the following metrics:
```
Task Scores
(1-10 Higher is better)
┏━━━━━━━━━━━━┳━━━━━━━┳━━━━━━━┳━━━━━━━━━━━━┓
┃ Tasks/Crew ┃ Run 1 ┃ Run 2 ┃ Avg. Total ┃
┡━━━━━━━━━━━━╇━━━━━━━╇━━━━━━━╇━━━━━━━━━━━━┩
│ Task 1 │ 10.0 │ 9.0 │ 9.5 │
│ Task 2 │ 9.0 │ 9.0 │ 9.0 │
│ Crew │ 9.5 │ 9.0 │ 9.2 │
└────────────┴───────┴───────┴────────────┘
```
The example above shows the test results for two runs of the crew with two tasks, with the average total score for each task and the crew as a whole.

View File

@@ -18,4 +18,7 @@ pip install crewai
# Install the main crewAI package and the tools package
# that includes a series of helpful tools for your agents
pip install 'crewai[tools]'
# Alternatively, you can also use:
pip install crewai crewai-tools
```

View File

@@ -0,0 +1,255 @@
---
title: Starting a New CrewAI Project - Using Template
description: A comprehensive guide to starting a new CrewAI project, including the latest updates and project setup methods.
---
# Starting Your CrewAI Project
Welcome to the ultimate guide for starting a new CrewAI project. This document will walk you through the steps to create, customize, and run your CrewAI project, ensuring you have everything you need to get started.
Beforre we start there are a couple of things to note:
1. CrewAI is a Python package and requires Python >=3.10 and <=3.13 to run.
2. The preferred way of setting up CrewAI is using the `crewai create` command.This will create a new project folder and install a skeleton template for you to work on.
## Prerequisites
Before getting started with CrewAI, make sure that you have installed it via pip:
```shell
$ pip install crewai crewai-tools
```
### Virtual Environments
It is highly recommended that you use virtual environments to ensure that your CrewAI project is isolated from other projects and dependencies. Virtual environments provide a clean, separate workspace for each project, preventing conflicts between different versions of packages and libraries. This isolation is crucial for maintaining consistency and reproducibility in your development process. You have multiple options for setting up virtual environments depending on your operating system and Python version:
1. Use venv (Python's built-in virtual environment tool):
venv is included with Python 3.3 and later, making it a convenient choice for many developers. It's lightweight and easy to use, perfect for simple project setups.
To set up virtual environments with venv, refer to the official [Python documentation](https://docs.python.org/3/tutorial/venv.html).
2. Use Conda (A Python virtual environment manager):
Conda is an open-source package manager and environment management system for Python. It's widely used by data scientists, developers, and researchers to manage dependencies and environments in a reproducible way.
To set up virtual environments with Conda, refer to the official [Conda documentation](https://docs.conda.io/projects/conda/en/stable/user-guide/getting-started.html).
3. Use Poetry (A Python package manager and dependency management tool):
Poetry is an open-source Python package manager that simplifies the installation of packages and their dependencies. Poetry offers a convenient way to manage virtual environments and dependencies.
Poetry is CrewAI's prefered tool for package / dependancy management in CrewAI.
### Code IDEs
Most users of CrewAI a Code Editor / Integrated Development Environment (IDE) for building there Crews. You can use any code IDE of your choice. Seee below for some popular options for Code Editors / Integrated Development Environments (IDE):
- [Visual Studio Code](https://code.visualstudio.com/) - Most popular
- [PyCharm](https://www.jetbrains.com/pycharm/)
- [Cursor AI](https://cursor.com)
Pick one that suits your style and needs.
## Creating a New Project
In this example we will be using Venv as our virtual environment manager.
To setup a virtual environment, run the following CLI command:
```shell
$ python3 -m venv <venv-name>
```
Activate your virtual environment by running the following CLI command:
```shell
$ source <venv-name>/bin/activate
```
Now, to create a new CrewAI project, run the following CLI command:
```shell
$ crewai create <project_name>
```
This command will create a new project folder with the following structure:
```shell
my_project/
├── .gitignore
├── pyproject.toml
├── README.md
└── src/
└── my_project/
├── __init__.py
├── main.py
├── crew.py
├── tools/
│ ├── custom_tool.py
│ └── __init__.py
└── config/
├── agents.yaml
└── tasks.yaml
```
You can now start developing your project by editing the files in the `src/my_project` folder. The `main.py` file is the entry point of your project, and the `crew.py` file is where you define your agents and tasks.
## Customizing Your Project
To customize your project, you can:
- Modify `src/my_project/config/agents.yaml` to define your agents.
- Modify `src/my_project/config/tasks.yaml` to define your tasks.
- Modify `src/my_project/crew.py` to add your own logic, tools, and specific arguments.
- Modify `src/my_project/main.py` to add custom inputs for your agents and tasks.
- Add your environment variables into the `.env` file.
### Example: Defining Agents and Tasks
#### agents.yaml
```yaml
researcher:
role: >
Job Candidate Researcher
goal: >
Find potential candidates for the job
backstory: >
You are adept at finding the right candidates by exploring various online
resources. Your skill in identifying suitable candidates ensures the best
match for job positions.
```
#### tasks.yaml
```yaml
research_candidates_task:
description: >
Conduct thorough research to find potential candidates for the specified job.
Utilize various online resources and databases to gather a comprehensive list of potential candidates.
Ensure that the candidates meet the job requirements provided.
Job Requirements:
{job_requirements}
expected_output: >
A list of 10 potential candidates with their contact information and brief profiles highlighting their suitability.
agent: researcher # THIS NEEDS TO MATCH THE AGENT NAME IN THE AGENTS.YAML FILE AND THE AGENT DEFINED IN THE Crew.PY FILE
context: # THESE NEED TO MATCH THE TASK NAMES DEFINED ABOVE AND THE TASKS.YAML FILE AND THE TASK DEFINED IN THE Crew.PY FILE
- researcher
```
### Referencing Variables:
Your defined functions with the same name will be used. For example, you can reference the agent for specific tasks from task.yaml file. Ensure your annotated agent and function name is the same otherwise your task wont recognize the reference properly.
#### Example References
agent.yaml
```yaml
email_summarizer:
role: >
Email Summarizer
goal: >
Summarize emails into a concise and clear summary
backstory: >
You will create a 5 bullet point summary of the report
llm: mixtal_llm
```
task.yaml
```yaml
email_summarizer_task:
description: >
Summarize the email into a 5 bullet point summary
expected_output: >
A 5 bullet point summary of the email
agent: email_summarizer
context:
- reporting_task
- research_task
```
Use the annotations are used to properly reference the agent and task in the crew.py file.
### Annotations include:
* @agent
* @task
* @crew
* @llm
* @tool
* @callback
* @output_json
* @output_pydantic
* @cache_handler
crew.py
```py
...
@llm
def mixtal_llm(self):
return ChatGroq(temperature=0, model_name="mixtral-8x7b-32768")
@agent
def email_summarizer(self) -> Agent:
return Agent(
config=self.agents_config["email_summarizer"],
)
## ...other tasks defined
@task
def email_summarizer_task(self) -> Task:
return Task(
config=self.tasks_config["email_summarizer_task"],
)
...
```
## Installing Dependencies
To install the dependencies for your project, you can use Poetry. First, navigate to your project directory:
```shell
$ cd my_project
$ poetry lock
$ poetry install
```
This will install the dependencies specified in the `pyproject.toml` file.
## Interpolating Variables
Any variable interpolated in your `agents.yaml` and `tasks.yaml` files like `{variable}` will be replaced by the value of the variable in the `main.py` file.
#### agents.yaml
```yaml
research_task:
description: >
Conduct a thorough research about the customer and competitors in the context
of {customer_domain}.
Make sure you find any interesting and relevant information given the
current year is 2024.
expected_output: >
A complete report on the customer and their customers and competitors,
including their demographics, preferences, market positioning and audience engagement.
```
#### main.py
```python
# main.py
def run():
inputs = {
"customer_domain": "crewai.com"
}
MyProjectCrew(inputs).crew().kickoff(inputs=inputs)
```
## Running Your Project
To run your project, use the following command:
```shell
$ poetry run my_project
```
This will initialize your crew of AI agents and begin task execution as defined in your configuration in the `main.py` file.
## Deploying Your Project
The easiest way to deploy your crew is through [CrewAI+](https://www.crewai.com/crewaiplus), where you can deploy your crew in a few clicks.

View File

@@ -7,6 +7,7 @@ description: Comprehensive guide on crafting, using, and managing custom tools w
This guide provides detailed instructions on creating custom tools for the crewAI framework and how to efficiently manage and utilize these tools, incorporating the latest functionalities such as tool delegation, error handling, and dynamic tool calling. It also highlights the importance of collaboration tools, enabling agents to perform a wide range of actions.
### Prerequisites
Before creating your own tools, ensure you have the crewAI extra tools package installed:
```bash
@@ -31,7 +32,7 @@ class MyCustomTool(BaseTool):
### Using the `tool` Decorator
Alternatively, use the `tool` decorator for a direct approach to create tools. This requires specifying attributes and the tool's logic within a function.
Alternatively, you can use the tool decorator `@tool`. This approach allows you to define the tool's attributes and functionality directly within a function, offering a concise and efficient way to create specialized tools tailored to your needs.
```python
from crewai_tools import tool

View File

@@ -1,84 +0,0 @@
---
title: Assembling and Activating Your CrewAI Team
description: A comprehensive guide to creating a dynamic CrewAI team for your projects, with updated functionalities including verbose mode, memory capabilities, asynchronous execution, output customization, language model configuration, code execution, integration with third-party agents, and improved task management.
---
## Introduction
Embark on your CrewAI journey by setting up your environment and initiating your AI crew with the latest features. This guide ensures a smooth start, incorporating all recent updates for an enhanced experience, including code execution capabilities, integration with third-party agents, and advanced task management.
## Step 0: Installation
Install CrewAI and any necessary packages for your project. CrewAI is compatible with Python >=3.10,<=3.13.
```shell
pip install crewai
pip install 'crewai[tools]'
```
## Step 1: Assemble Your Agents
Define your agents with distinct roles, backstories, and enhanced capabilities. The Agent class now supports a wide range of attributes for fine-tuned control over agent behavior and interactions, including code execution and integration with third-party agents.
```python
import os
from langchain.llms import OpenAI
from crewai import Agent
from crewai_tools import SerperDevTool, BrowserbaseLoadTool, EXASearchTool
os.environ["OPENAI_API_KEY"] = "Your OpenAI Key"
os.environ["SERPER_API_KEY"] = "Your Serper Key"
os.environ["BROWSERBASE_API_KEY"] = "Your BrowserBase Key"
os.environ["BROWSERBASE_PROJECT_ID"] = "Your BrowserBase Project Id"
search_tool = SerperDevTool()
browser_tool = BrowserbaseLoadTool()
exa_search_tool = EXASearchTool()
# Creating a senior researcher agent with advanced configurations
researcher = Agent(
role='Senior Researcher',
goal='Uncover groundbreaking technologies in {topic}',
backstory=("Driven by curiosity, you're at the forefront of innovation, "
"eager to explore and share knowledge that could change the world."),
memory=True,
verbose=True,
allow_delegation=False,
tools=[search_tool, browser_tool],
allow_code_execution=False, # New attribute for enabling code execution
max_iter=15, # Maximum number of iterations for task execution
max_rpm=100, # Maximum requests per minute
max_execution_time=3600, # Maximum execution time in seconds
system_template="Your custom system template here", # Custom system template
prompt_template="Your custom prompt template here", # Custom prompt template
response_template="Your custom response template here", # Custom response template
)
# Creating a writer agent with custom tools and specific configurations
writer = Agent(
role='Writer',
goal='Narrate compelling tech stories about {topic}',
backstory=("With a flair for simplifying complex topics, you craft engaging "
"narratives that captivate and educate, bringing new discoveries to light."),
verbose=True,
allow_delegation=False,
memory=True,
tools=[exa_search_tool],
function_calling_llm=OpenAI(model_name="gpt-3.5-turbo"), # Separate LLM for function calling
)
# Setting a specific manager agent
manager = Agent(
role='Manager',
goal='Ensure the smooth operation and coordination of the team',
verbose=True,
backstory=(
"As a seasoned project manager, you excel in organizing "
"tasks, managing timelines, and ensuring the team stays on track."
),
allow_code_execution=True, # Enable code execution for the manager
)
```
### New Agent Attributes and Features
1. `allow_code_execution`: Enable or disable code execution capabilities for the agent (default is False).
2. `max_execution_time`: Set a maximum execution time (in seconds) for the agent to complete a task.
3. `function_calling_llm`: Specify a separate language model for function calling.

View File

@@ -7,7 +7,7 @@ description: Learn how to force tool output as the result in of an Agent's task
In CrewAI, you can force the output of a tool as the result of an agent's task. This feature is useful when you want to ensure that the tool output is captured and returned as the task result, and avoid the agent modifying the output during the task execution.
## Forcing Tool Output as Result
To force the tool output as the result of an agent's task, you can set the `force_tool_output` parameter to `True` when creating the task. This parameter ensures that the tool output is captured and returned as the task result, without any modifications by the agent.
To force the tool output as the result of an agent's task, you can set the `result_as_answer` parameter to `True` when creating the agent. This parameter ensures that the tool output is captured and returned as the task result, without any modifications by the agent.
Here's an example of how to force the tool output as the result of an agent's task:
@@ -16,7 +16,7 @@ Here's an example of how to force the tool output as the result of an agent's ta
# Define a custom tool that returns the result as the answer
coding_agent =Agent(
role="Data Scientist",
goal="Product amazing resports on AI",
goal="Product amazing reports on AI",
backstory="You work with data and AI",
tools=[MyCustomTool(result_as_answer=True)],
)

View File

@@ -6,33 +6,25 @@ description: Comprehensive guide on integrating CrewAI with various Large Langua
## Connect CrewAI to LLMs
!!! note "Default LLM"
By default, CrewAI uses OpenAI's GPT-4 model (specifically, the model specified by the OPENAI_MODEL_NAME environment variable, defaulting to "gpt-4o") for language processing. You can configure your agents to use a different model or API as described in this guide.
By default, CrewAI uses OpenAI's GPT-4o model (specifically, the model specified by the OPENAI_MODEL_NAME environment variable, defaulting to "gpt-4o") for language processing. You can configure your agents to use a different model or API as described in this guide.
By default, CrewAI uses OpenAI's GPT-4 model (specifically, the model specified by the OPENAI_MODEL_NAME environment variable, defaulting to "gpt-4") for language processing. You can configure your agents to use a different model or API as described in this guide.
CrewAI offers flexibility in connecting to various LLMs, including local models via [Ollama](https://ollama.ai) and different APIs like Azure. It's compatible with all [LangChain LLM](https://python.langchain.com/docs/integrations/llms/) components, enabling diverse integrations for tailored AI solutions.
CrewAI provides extensive versatility in integrating with various Language Models (LLMs), including local options through Ollama such as Llama and Mixtral to cloud-based solutions like Azure. Its compatibility extends to all [LangChain LLM components](https://python.langchain.com/v0.2/docs/integrations/llms/), offering a wide range of integration possibilities for customized AI applications.
## CrewAI Agent Overview
The platform supports connections to an array of Generative AI models, including:
The `Agent` class is the cornerstone for implementing AI solutions in CrewAI. Here's a comprehensive overview of the Agent class attributes and methods:
- OpenAI's suite of advanced language models
- Anthropic's cutting-edge AI offerings
- Ollama's diverse range of locally-hosted generative model & embeddings
- LM Studio's diverse range of locally hosted generative models & embeddings
- Groq's Super Fast LLM offerings
- Azures' generative AI offerings
- HuggingFace's generative AI offerings
- **Attributes**:
- `role`: Defines the agent's role within the solution.
- `goal`: Specifies the agent's objective.
- `backstory`: Provides a background story to the agent.
- `cache` *Optional*: Determines whether the agent should use a cache for tool usage. Default is `True`.
- `max_rpm` *Optional*: Maximum number of requests per minute the agent's execution should respect. Optional.
- `verbose` *Optional*: Enables detailed logging of the agent's execution. Default is `False`.
- `allow_delegation` *Optional*: Allows the agent to delegate tasks to other agents, default is `True`.
- `tools`: Specifies the tools available to the agent for task execution. Optional.
- `max_iter` *Optional*: Maximum number of iterations for an agent to execute a task, default is 25.
- `max_execution_time` *Optional*: Maximum execution time for an agent to execute a task. Optional.
- `step_callback` *Optional*: Provides a callback function to be executed after each step. Optional.
- `llm` *Optional*: Indicates the Large Language Model the agent uses. By default, it uses the GPT-4 model defined in the environment variable "OPENAI_MODEL_NAME".
- `function_calling_llm` *Optional* : Will turn the ReAct CrewAI agent into a function-calling agent.
- `callbacks` *Optional*: A list of callback functions from the LangChain library that are triggered during the agent's execution process.
- `system_template` *Optional*: Optional string to define the system format for the agent.
- `prompt_template` *Optional*: Optional string to define the prompt format for the agent.
- `response_template` *Optional*: Optional string to define the response format for the agent.
This broad spectrum of LLM options enables users to select the most suitable model for their specific needs, whether prioritizing local deployment, specialized capabilities, or cloud-based scalability.
## Changing the default LLM
The default LLM is provided through the `langchain openai` package, which is installed by default when you install CrewAI. You can change this default LLM to a different model or API by setting the `OPENAI_MODEL_NAME` environment variable. This straightforward process allows you to harness the power of different OpenAI models, enhancing the flexibility and capabilities of your CrewAI implementation.
```python
# Required
os.environ["OPENAI_MODEL_NAME"]="gpt-4-0125-preview"
@@ -45,30 +37,27 @@ example_agent = Agent(
verbose=True
)
```
## Ollama Local Integration
Ollama is preferred for local LLM integration, offering customization and privacy benefits. To integrate Ollama with CrewAI, you will need the `langchain-ollama` package. You can then set the following environment variables to connect to your Ollama instance running locally on port 11434.
## Ollama Integration
Ollama is preferred for local LLM integration, offering customization and privacy benefits. To integrate Ollama with CrewAI, set the appropriate environment variables as shown below.
### Setting Up Ollama
- **Environment Variables Configuration**: To integrate Ollama, set the following environment variables:
```sh
OPENAI_API_BASE='http://localhost:11434'
OPENAI_MODEL_NAME='llama2' # Adjust based on available model
OPENAI_API_KEY=''
os.environ[OPENAI_API_BASE]='http://localhost:11434'
os.environ[OPENAI_MODEL_NAME]='llama2' # Adjust based on available model
os.environ[OPENAI_API_KEY]='' # No API Key required for Ollama
```
## Ollama Integration (ex. for using Llama 2 locally)
1. [Download Ollama](https://ollama.com/download).
2. After setting up the Ollama, Pull the Llama2 by typing following lines into the terminal ```ollama pull llama2```.
3. Enjoy your free Llama2 model that powered up by excellent agents from crewai.
## Ollama Integration Step by Step (ex. for using Llama 3.1 8B locally)
1. [Download and install Ollama](https://ollama.com/download).
2. After setting up the Ollama, Pull the Llama3.1 8B model by typing following lines into your terminal ```ollama run llama3.1```.
3. Llama3.1 should now be served locally on `http://localhost:11434`
```
from crewai import Agent, Task, Crew
from langchain.llms import Ollama
from langchain_ollama import ChatOllama
import os
os.environ["OPENAI_API_KEY"] = "NA"
llm = Ollama(
model = "llama2",
model = "llama3.1",
base_url = "http://localhost:11434")
general_agent = Agent(role = "Math Professor",
@@ -98,13 +87,14 @@ There are a couple of different ways you can use HuggingFace to host your LLM.
### Your own HuggingFace endpoint
```python
from langchain_community.llms import HuggingFaceEndpoint
from langchain_huggingface import HuggingFaceEndpoint,
llm = HuggingFaceEndpoint(
endpoint_url="<YOUR_ENDPOINT_URL_HERE>",
huggingfacehub_api_token="<HF_TOKEN_HERE>",
repo_id="microsoft/Phi-3-mini-4k-instruct",
task="text-generation",
max_new_tokens=512
max_new_tokens=512,
do_sample=False,
repetition_penalty=1.03,
)
agent = Agent(
@@ -115,66 +105,50 @@ agent = Agent(
)
```
### From HuggingFaceHub endpoint
```python
from langchain_community.llms import HuggingFaceHub
llm = HuggingFaceHub(
repo_id="HuggingFaceH4/zephyr-7b-beta",
huggingfacehub_api_token="<HF_TOKEN_HERE>",
task="text-generation",
)
```
## OpenAI Compatible API Endpoints
Switch between APIs and models seamlessly using environment variables, supporting platforms like FastChat, LM Studio, Groq, and Mistral AI.
### Configuration Examples
#### FastChat
```sh
OPENAI_API_BASE="http://localhost:8001/v1"
OPENAI_MODEL_NAME='oh-2.5m7b-q51'
OPENAI_API_KEY=NA
os.environ[OPENAI_API_BASE]="http://localhost:8001/v1"
os.environ[OPENAI_MODEL_NAME]='oh-2.5m7b-q51'
os.environ[OPENAI_API_KEY]=NA
```
#### LM Studio
Launch [LM Studio](https://lmstudio.ai) and go to the Server tab. Then select a model from the dropdown menu and wait for it to load. Once it's loaded, click the green Start Server button and use the URL, port, and API key that's shown (you can modify them). Below is an example of the default settings as of LM Studio 0.2.19:
```sh
OPENAI_API_BASE="http://localhost:1234/v1"
OPENAI_API_KEY="lm-studio"
os.environ[OPENAI_API_BASE]="http://localhost:1234/v1"
os.environ[OPENAI_API_KEY]="lm-studio"
```
#### Groq API
```sh
OPENAI_API_KEY=your-groq-api-key
OPENAI_MODEL_NAME='llama3-8b-8192'
OPENAI_API_BASE=https://api.groq.com/openai/v1
os.environ[OPENAI_API_KEY]=your-groq-api-key
os.environ[OPENAI_MODEL_NAME]='llama3-8b-8192'
os.environ[OPENAI_API_BASE]=https://api.groq.com/openai/v1
```
#### Mistral API
```sh
OPENAI_API_KEY=your-mistral-api-key
OPENAI_API_BASE=https://api.mistral.ai/v1
OPENAI_MODEL_NAME="mistral-small"
os.environ[OPENAI_API_KEY]=your-mistral-api-key
os.environ[OPENAI_API_BASE]=https://api.mistral.ai/v1
os.environ[OPENAI_MODEL_NAME]="mistral-small"
```
### Solar
```python
```sh
from langchain_community.chat_models.solar import SolarChat
# Initialize language model
os.environ["SOLAR_API_KEY"] = "your-solar-api-key"
llm = SolarChat(max_tokens=1024)
```
```sh
os.environ[SOLAR_API_BASE]="https://api.upstage.ai/v1/solar"
os.environ[SOLAR_API_KEY]="your-solar-api-key"
```
# Free developer API key available here: https://console.upstage.ai/services/solar
# Langchain Example: https://github.com/langchain-ai/langchain/pull/18556
```
### text-gen-web-ui
```sh
OPENAI_API_BASE=http://localhost:5000/v1
OPENAI_MODEL_NAME=NA
OPENAI_API_KEY=NA
```
### Cohere
```python
@@ -190,10 +164,11 @@ llm = ChatCohere()
### Azure Open AI Configuration
For Azure OpenAI API integration, set the following environment variables:
```sh
AZURE_OPENAI_VERSION="2022-12-01"
AZURE_OPENAI_DEPLOYMENT=""
AZURE_OPENAI_ENDPOINT=""
AZURE_OPENAI_KEY=""
os.environ[AZURE_OPENAI_DEPLOYMENT] = "You deployment"
os.environ["OPENAI_API_VERSION"] = "2023-12-01-preview"
os.environ["AZURE_OPENAI_ENDPOINT"] = "Your Endpoint"
os.environ["AZURE_OPENAI_API_KEY"] = "<Your API Key>"
```
### Example Agent with Azure LLM
@@ -216,6 +191,5 @@ azure_agent = Agent(
llm=azure_llm
)
```
## Conclusion
Integrating CrewAI with different LLMs expands the framework's versatility, allowing for customized, efficient AI solutions across various domains and platforms.

View File

@@ -36,14 +36,14 @@ To replay from a task programmatically, use the following steps:
2. Execute the replay command within a try-except block to handle potential errors.
```python
def replay_from_task():
def replay():
"""
Replay the crew execution from a specific task.
"""
task_id = '<task_id>'
inputs = {"topic": "CrewAI Training"} # this is optional, you can pass in the inputs you want to replay otherwise uses the previous kickoffs inputs
try:
YourCrewName_Crew().crew().replay_from_task(task_id=task_id, inputs=inputs)
YourCrewName_Crew().crew().replay(task_id=task_id, inputs=inputs)
except Exception as e:
raise Exception(f"An error occurred while replaying the crew: {e}")

View File

@@ -1,137 +0,0 @@
---
title: Starting a New CrewAI Project
description: A comprehensive guide to starting a new CrewAI project, including the latest updates and project setup methods.
---
# Starting Your CrewAI Project
Welcome to the ultimate guide for starting a new CrewAI project. This document will walk you through the steps to create, customize, and run your CrewAI project, ensuring you have everything you need to get started.
## Prerequisites
We assume you have already installed CrewAI. If not, please refer to the [installation guide](https://docs.crewai.com/how-to/Installing-CrewAI/) to install CrewAI and its dependencies.
## Creating a New Project
To create a new project, run the following CLI command:
```shell
$ crewai create my_project
```
This command will create a new project folder with the following structure:
```shell
my_project/
├── .gitignore
├── pyproject.toml
├── README.md
└── src/
└── my_project/
├── __init__.py
├── main.py
├── crew.py
├── tools/
│ ├── custom_tool.py
│ └── __init__.py
└── config/
├── agents.yaml
└── tasks.yaml
```
You can now start developing your project by editing the files in the `src/my_project` folder. The `main.py` file is the entry point of your project, and the `crew.py` file is where you define your agents and tasks.
## Customizing Your Project
To customize your project, you can:
- Modify `src/my_project/config/agents.yaml` to define your agents.
- Modify `src/my_project/config/tasks.yaml` to define your tasks.
- Modify `src/my_project/crew.py` to add your own logic, tools, and specific arguments.
- Modify `src/my_project/main.py` to add custom inputs for your agents and tasks.
- Add your environment variables into the `.env` file.
### Example: Defining Agents and Tasks
#### agents.yaml
```yaml
researcher:
role: >
Job Candidate Researcher
goal: >
Find potential candidates for the job
backstory: >
You are adept at finding the right candidates by exploring various online
resources. Your skill in identifying suitable candidates ensures the best
match for job positions.
```
#### tasks.yaml
```yaml
research_candidates_task:
description: >
Conduct thorough research to find potential candidates for the specified job.
Utilize various online resources and databases to gather a comprehensive list of potential candidates.
Ensure that the candidates meet the job requirements provided.
Job Requirements:
{job_requirements}
expected_output: >
A list of 10 potential candidates with their contact information and brief profiles highlighting their suitability.
```
## Installing Dependencies
To install the dependencies for your project, you can use Poetry. First, navigate to your project directory:
```shell
$ cd my_project
$ poetry lock
$ poetry install
```
This will install the dependencies specified in the `pyproject.toml` file.
## Interpolating Variables
Any variable interpolated in your `agents.yaml` and `tasks.yaml` files like `{variable}` will be replaced by the value of the variable in the `main.py` file.
#### agents.yaml
```yaml
research_task:
description: >
Conduct a thorough research about the customer and competitors in the context
of {customer_domain}.
Make sure you find any interesting and relevant information given the
current year is 2024.
expected_output: >
A complete report on the customer and their customers and competitors,
including their demographics, preferences, market positioning and audience engagement.
```
#### main.py
```python
# main.py
def run():
inputs = {
"customer_domain": "crewai.com"
}
MyProjectCrew(inputs).crew().kickoff(inputs=inputs)
```
## Running Your Project
To run your project, use the following command:
```shell
$ poetry run my_project
```
This will initialize your crew of AI agents and begin task execution as defined in your configuration in the `main.py` file.
## Deploying Your Project
The easiest way to deploy your crew is through [CrewAI+](https://www.crewai.com/crewaiplus), where you can deploy your crew in a few clicks.

View File

@@ -5,6 +5,19 @@
Cutting-edge framework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks.
<div style="display:flex; margin:0 auto; justify-content: center;">
<div style="width:25%">
<h2>Getting Started</h2>
<ul>
<li><a href='./getting-started/Installing-CrewAI'>
Installing CrewAI
</a>
</li>
<li><a href='./getting-started/Start-a-New-CrewAI-Project-Template-Method'>
Start a New CrewAI Project: Template Method
</a>
</li>
</ul>
</div>
<div style="width:25%">
<h2>Core Concepts</h2>
<ul>
@@ -53,21 +66,6 @@ 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/Start-a-New-CrewAI-Project">
Starting Your crewAI Project
</a>
</li>
<li>
<a href="./how-to/Installing-CrewAI">
Installing crewAI
</a>
</li>
<li>
<a href="./how-to/Creating-a-Crew-and-kick-it-off">
Getting Started
</a>
</li>
<li>
<a href="./how-to/Create-Custom-Tools">
Create Custom Tools

View File

@@ -5,7 +5,7 @@ description: Understanding the telemetry data collected by CrewAI and how it con
## Telemetry
CrewAI utilizes anonymous telemetry to gather usage statistics with the primary goal of enhancing the library. Our focus is on improving and developing the features, integrations, and tools most utilized by our users.
CrewAI utilizes anonymous telemetry to gather usage statistics with the primary goal of enhancing the library. Our focus is on improving and developing the features, integrations, and tools most utilized by our users. We don't offer a way to disable it now, but we will in the future.
It's pivotal to understand that **NO data is collected** concerning prompts, task descriptions, agents' backstories or goals, usage of tools, API calls, responses, any data processed by the agents, or secrets and environment variables, with the exception of the conditions mentioned. When the `share_crew` feature is enabled, detailed data including task descriptions, agents' backstories or goals, and other specific attributes are collected to provide deeper insights while respecting user privacy.
@@ -22,7 +22,7 @@ It's pivotal to understand that **NO data is collected** concerning prompts, tas
- **Tool Usage**: Identifying which tools are most frequently used allows us to prioritize improvements in those areas.
### Opt-In Further Telemetry Sharing
Users can choose to share their complete telemetry data by enabling the `share_crew` attribute to `True` in their crew configurations. This opt-in approach respects user privacy and aligns with data protection standards by ensuring users have control over their data sharing preferences. Enabling `share_crew` results in the collection of detailed crew and task execution data, including `goal`, `backstory`, `context`, and `output` of tasks. This enables a deeper insight into usage patterns while respecting the user's choice to share.
Users can choose to share their complete telemetry data by enabling the `share_crew` attribute to `True` in their crew configurations. Enabling `share_crew` results in the collection of detailed crew and task execution data, including `goal`, `backstory`, `context`, and `output` of tasks. This enables a deeper insight into usage patterns while respecting the user's choice to share.
### Updates and Revisions
We are committed to maintaining the accuracy and transparency of our documentation. Regular reviews and updates are performed to ensure our documentation accurately reflects the latest developments of our codebase and telemetry practices. Users are encouraged to review this section for the most current information on our data collection practices and how they contribute to the improvement of CrewAI.

View File

@@ -1,9 +1,9 @@
# CodeInterpreterTool
## Description
This tool is used to give the Agent the ability to run code (Python3) from the code generated by the Agent itself. The code is executed in a sandboxed environment, so it is safe to run any code.
This tool enables the Agent to execute Python 3 code that it has generated autonomously. The code is run in a secure, isolated environment, ensuring safety regardless of the content.
It is incredible useful since it allows the Agent to generate code, run it in the same environment, get the result and use it to make decisions.
This functionality is particularly valuable as it allows the Agent to create code, execute it within the same ecosystem, obtain the results, and utilize that information to inform subsequent decisions and actions.
## Requirements

View File

@@ -2,7 +2,7 @@
## Description
This tools is a wrapper around the composio toolset and gives your agent access to a wide variety of tools from the composio SDK.
This tools is a wrapper around the composio set of tools and gives your agent access to a wide variety of tools from the composio SDK.
## Installation
@@ -19,7 +19,7 @@ after the installation is complete, either run `composio login` or export your c
The following example demonstrates how to initialize the tool and execute a github action:
1. Initialize toolset
1. Initialize Composio tools
```python
from composio import App

View File

@@ -29,5 +29,69 @@ To effectively use the `SerperDevTool`, follow these steps:
2. **API Key Acquisition**: Acquire a `serper.dev` API key by registering for a free account at `serper.dev`.
3. **Environment Configuration**: Store your obtained API key in an environment variable named `SERPER_API_KEY` to facilitate its use by the tool.
## Parameters
The `SerperDevTool` comes with several parameters that will be passed to the API :
- **search_url**: The URL endpoint for the search API. (Default is `https://google.serper.dev/search`)
- **country**: Optional. Specify the country for the search results.
- **location**: Optional. Specify the location for the search results.
- **locale**: Optional. Specify the locale for the search results.
- **n_results**: Number of search results to return. Default is `10`.
The values for `country`, `location`, `locale` and `search_url` can be found on the [Serper Playground](https://serper.dev/playground).
## Example with Parameters
Here is an example demonstrating how to use the tool with additional parameters:
```python
from crewai_tools import SerperDevTool
tool = SerperDevTool(
search_url="https://google.serper.dev/scholar",
n_results=2,
)
print(tool.run(search_query="ChatGPT"))
# Using Tool: Search the internet
# Search results: Title: Role of chat gpt in public health
# Link: https://link.springer.com/article/10.1007/s10439-023-03172-7
# Snippet: … ChatGPT in public health. In this overview, we will examine the potential uses of ChatGPT in
# ---
# Title: Potential use of chat gpt in global warming
# Link: https://link.springer.com/article/10.1007/s10439-023-03171-8
# Snippet: … as ChatGPT, have the potential to play a critical role in advancing our understanding of climate
# ---
```
```python
from crewai_tools import SerperDevTool
tool = SerperDevTool(
country="fr",
locale="fr",
location="Paris, Paris, Ile-de-France, France",
n_results=2,
)
print(tool.run(search_query="Jeux Olympiques"))
# Using Tool: Search the internet
# Search results: Title: Jeux Olympiques de Paris 2024 - Actualités, calendriers, résultats
# Link: https://olympics.com/fr/paris-2024
# Snippet: Quels sont les sports présents aux Jeux Olympiques de Paris 2024 ? · Athlétisme · Aviron · Badminton · Basketball · Basketball 3x3 · Boxe · Breaking · Canoë ...
# ---
# Title: Billetterie Officielle de Paris 2024 - Jeux Olympiques et Paralympiques
# Link: https://tickets.paris2024.org/
# Snippet: Achetez vos billets exclusivement sur le site officiel de la billetterie de Paris 2024 pour participer au plus grand événement sportif au monde.
# ---
```
## Conclusion
By integrating the `SerperDevTool` into Python projects, users gain the ability to conduct real-time, relevant searches across the internet directly from their applications. By adhering to the setup and usage guidelines provided, incorporating this tool into projects is streamlined and straightforward.
By integrating the `SerperDevTool` into Python projects, users gain the ability to conduct real-time, relevant searches across the internet directly from their applications. The updated parameters allow for more customized and localized search results. By adhering to the setup and usage guidelines provided, incorporating this tool into projects is streamlined and straightforward.

View File

@@ -119,6 +119,9 @@ theme:
nav:
- Home: '/'
- Getting Started:
- Installing CrewAI: 'getting-started/Installing-CrewAI.md'
- Starting a new CrewAI project: 'getting-started/Start-a-New-CrewAI-Project-Template-Method.md'
- Core Concepts:
- Agents: 'core-concepts/Agents.md'
- Tasks: 'core-concepts/Tasks.md'
@@ -129,6 +132,7 @@ nav:
- Training: 'core-concepts/Training-Crew.md'
- Memory: 'core-concepts/Memory.md'
- Planning: 'core-concepts/Planning.md'
- Testing: 'core-concepts/Testing.md'
- Using LangChain Tools: 'core-concepts/Using-LangChain-Tools.md'
- Using LlamaIndex Tools: 'core-concepts/Using-LlamaIndex-Tools.md'
- How to Guides:

774
poetry.lock generated

File diff suppressed because it is too large Load Diff

View File

@@ -1,6 +1,6 @@
[tool.poetry]
name = "crewai"
version = "0.41.0"
version = "0.46.0"
description = "Cutting-edge framework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks."
authors = ["Joao Moura <joao@crewai.com>"]
readme = "README.md"

View File

@@ -55,8 +55,6 @@ class Agent(BaseAgent):
tools: Tools at agents disposal
step_callback: Callback to be executed after each step of the agent execution.
callbacks: A list of callback functions from the langchain library that are triggered during the agent's execution process
allow_code_execution: Enable code execution for the agent.
max_retry_limit: Maximum number of retries for an agent to execute a task when an error occurs.
"""
_times_executed: int = PrivateAttr(default=0)
@@ -262,6 +260,7 @@ class Agent(BaseAgent):
"tools_handler": self.tools_handler,
"function_calling_llm": self.function_calling_llm,
"callbacks": self.callbacks,
"max_tokens": self.max_tokens,
}
if self._rpm_controller:

View File

@@ -45,6 +45,7 @@ class BaseAgent(ABC, BaseModel):
i18n (I18N): Internationalization settings.
cache_handler (InstanceOf[CacheHandler]): An instance of the CacheHandler class.
tools_handler (InstanceOf[ToolsHandler]): An instance of the ToolsHandler class.
max_tokens: Maximum number of tokens for the agent to generate in a response.
Methods:
@@ -118,6 +119,9 @@ class BaseAgent(ABC, BaseModel):
tools_handler: InstanceOf[ToolsHandler] = Field(
default=None, description="An instance of the ToolsHandler class."
)
max_tokens: Optional[int] = Field(
default=None, description="Maximum number of tokens for the agent's execution."
)
_original_role: str | None = None
_original_goal: str | None = None

View File

@@ -3,7 +3,6 @@ from typing import TYPE_CHECKING, Optional
from crewai.memory.entity.entity_memory_item import EntityMemoryItem
from crewai.memory.long_term.long_term_memory_item import LongTermMemoryItem
from crewai.memory.short_term.short_term_memory_item import ShortTermMemoryItem
from crewai.utilities.converter import ConverterError
from crewai.utilities.evaluators.task_evaluator import TaskEvaluator
from crewai.utilities import I18N
@@ -39,18 +38,17 @@ class CrewAgentExecutorMixin:
and "Action: Delegate work to coworker" not in output.log
):
try:
memory = ShortTermMemoryItem(
data=output.log,
agent=self.crew_agent.role,
metadata={
"observation": self.task.description,
},
)
if (
hasattr(self.crew, "_short_term_memory")
and self.crew._short_term_memory
):
self.crew._short_term_memory.save(memory)
self.crew._short_term_memory.save(
value=output.log,
metadata={
"observation": self.task.description,
},
agent=self.crew_agent.role,
)
except Exception as e:
print(f"Failed to add to short term memory: {e}")
pass

View File

@@ -1,6 +1,8 @@
import threading
import time
from typing import Any, Dict, Iterator, List, Optional, Tuple, Union
from typing import Any, Dict, Iterator, List, Literal, Optional, Tuple, Union
import click
from langchain.agents import AgentExecutor
from langchain.agents.agent import ExceptionTool
@@ -11,12 +13,21 @@ from langchain_core.tools import BaseTool
from langchain_core.utils.input import get_color_mapping
from pydantic import InstanceOf
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.chains.summarize import load_summarize_chain
from crewai.agents.agent_builder.base_agent_executor_mixin import CrewAgentExecutorMixin
from crewai.agents.tools_handler import ToolsHandler
from crewai.tools.tool_usage import ToolUsage, ToolUsageErrorException
from crewai.utilities import I18N
from crewai.utilities.constants import TRAINING_DATA_FILE
from crewai.utilities.exceptions.context_window_exceeding_exception import (
LLMContextLengthExceededException,
)
from crewai.utilities.training_handler import CrewTrainingHandler
from crewai.utilities.logger import Logger
class CrewAgentExecutor(AgentExecutor, CrewAgentExecutorMixin):
@@ -40,6 +51,8 @@ class CrewAgentExecutor(AgentExecutor, CrewAgentExecutorMixin):
system_template: Optional[str] = None
prompt_template: Optional[str] = None
response_template: Optional[str] = None
_logger: Logger = Logger(verbose_level=2)
_fit_context_window_strategy: Optional[Literal["summarize"]] = "summarize"
def _call(
self,
@@ -131,7 +144,7 @@ class CrewAgentExecutor(AgentExecutor, CrewAgentExecutorMixin):
intermediate_steps = self._prepare_intermediate_steps(intermediate_steps)
# Call the LLM to see what to do.
output = self.agent.plan( # type: ignore # Incompatible types in assignment (expression has type "AgentAction | AgentFinish | list[AgentAction]", variable has type "AgentAction")
output = self.agent.plan(
intermediate_steps,
callbacks=run_manager.get_child() if run_manager else None,
**inputs,
@@ -185,6 +198,27 @@ class CrewAgentExecutor(AgentExecutor, CrewAgentExecutorMixin):
yield AgentStep(action=output, observation=observation)
return
except Exception as e:
if LLMContextLengthExceededException(str(e))._is_context_limit_error(
str(e)
):
output = self._handle_context_length_error(
intermediate_steps, run_manager, inputs
)
if isinstance(output, AgentFinish):
yield output
elif isinstance(output, list):
for step in output:
yield step
return
yield AgentStep(
action=AgentAction("_Exception", str(e), str(e)),
observation=str(e),
)
return
# If the tool chosen is the finishing tool, then we end and return.
if isinstance(output, AgentFinish):
if self.should_ask_for_human_input:
@@ -235,6 +269,7 @@ class CrewAgentExecutor(AgentExecutor, CrewAgentExecutorMixin):
agent=self.crew_agent,
action=agent_action,
)
tool_calling = tool_usage.parse(agent_action.log)
if isinstance(tool_calling, ToolUsageErrorException):
@@ -280,3 +315,91 @@ class CrewAgentExecutor(AgentExecutor, CrewAgentExecutorMixin):
CrewTrainingHandler(TRAINING_DATA_FILE).append(
self.crew._train_iteration, agent_id, training_data
)
def _handle_context_length(
self, intermediate_steps: List[Tuple[AgentAction, str]]
) -> List[Tuple[AgentAction, str]]:
text = intermediate_steps[0][1]
original_action = intermediate_steps[0][0]
text_splitter = RecursiveCharacterTextSplitter(
separators=["\n\n", "\n"],
chunk_size=8000,
chunk_overlap=500,
)
if self._fit_context_window_strategy == "summarize":
docs = text_splitter.create_documents([text])
self._logger.log(
"debug",
"Summarizing Content, it is recommended to use a RAG tool",
color="bold_blue",
)
summarize_chain = load_summarize_chain(
self.llm, chain_type="map_reduce", verbose=True
)
summarized_docs = []
for doc in docs:
summary = summarize_chain.invoke(
{"input_documents": [doc]}, return_only_outputs=True
)
summarized_docs.append(summary["output_text"])
formatted_results = "\n\n".join(summarized_docs)
summary_step = AgentStep(
action=AgentAction(
tool=original_action.tool,
tool_input=original_action.tool_input,
log=original_action.log,
),
observation=formatted_results,
)
summary_tuple = (summary_step.action, summary_step.observation)
return [summary_tuple]
return intermediate_steps
def _handle_context_length_error(
self,
intermediate_steps: List[Tuple[AgentAction, str]],
run_manager: Optional[CallbackManagerForChainRun],
inputs: Dict[str, str],
) -> Union[AgentFinish, List[AgentStep]]:
self._logger.log(
"debug",
"Context length exceeded. Asking user if they want to use summarize prompt to fit, this will reduce context length.",
color="yellow",
)
user_choice = click.confirm(
"Context length exceeded. Do you want to summarize the text to fit models context window?"
)
if user_choice:
self._logger.log(
"debug",
"Context length exceeded. Using summarize prompt to fit, this will reduce context length.",
color="bold_blue",
)
intermediate_steps = self._handle_context_length(intermediate_steps)
output = self.agent.plan(
intermediate_steps,
callbacks=run_manager.get_child() if run_manager else None,
**inputs,
)
if isinstance(output, AgentFinish):
return output
elif isinstance(output, AgentAction):
return [AgentStep(action=output, observation=None)]
else:
return [AgentStep(action=action, observation=None) for action in output]
else:
self._logger.log(
"debug",
"Context length exceeded. Consider using smaller text or RAG tools from crewai_tools.",
color="red",
)
raise SystemExit(
"Context length exceeded and user opted not to summarize. Consider using smaller text or RAG tools from crewai_tools."
)

View File

@@ -5,11 +5,11 @@ from crewai.memory.storage.kickoff_task_outputs_storage import (
KickoffTaskOutputsSQLiteStorage,
)
from .create_crew import create_crew
from .train_crew import train_crew
from .evaluate_crew import evaluate_crew
from .replay_from_task import replay_task_command
from .reset_memories_command import reset_memories_command
from .train_crew import train_crew
@click.group()
@@ -126,5 +126,26 @@ def reset_memories(long, short, entities, kickoff_outputs, all):
click.echo(f"An error occurred while resetting memories: {e}", err=True)
@crewai.command()
@click.option(
"-n",
"--n_iterations",
type=int,
default=3,
help="Number of iterations to Test the crew",
)
@click.option(
"-m",
"--model",
type=str,
default="gpt-4o-mini",
help="LLM Model to run the tests on the Crew. For now only accepting only OpenAI models.",
)
def test(n_iterations: int, model: str):
"""Test the crew and evaluate the results."""
click.echo(f"Testing the crew for {n_iterations} iterations with model {model}")
evaluate_crew(n_iterations, model)
if __name__ == "__main__":
crewai()

View File

@@ -0,0 +1,30 @@
import subprocess
import click
def evaluate_crew(n_iterations: int, model: str) -> None:
"""
Test and Evaluate the crew by running a command in the Poetry environment.
Args:
n_iterations (int): The number of iterations to test the crew.
model (str): The model to test the crew with.
"""
command = ["poetry", "run", "test", str(n_iterations), model]
try:
if n_iterations <= 0:
raise ValueError("The number of iterations must be a positive integer.")
result = subprocess.run(command, capture_output=False, text=True, check=True)
if result.stderr:
click.echo(result.stderr, err=True)
except subprocess.CalledProcessError as e:
click.echo(f"An error occurred while testing the crew: {e}", err=True)
click.echo(e.output, err=True)
except Exception as e:
click.echo(f"An unexpected error occurred: {e}", err=True)

View File

@@ -9,10 +9,14 @@ from crewai.utilities.task_output_storage_handler import TaskOutputStorageHandle
def reset_memories_command(long, short, entity, kickoff_outputs, all) -> None:
"""
Replay the crew execution from a specific task.
Reset the crew memories.
Args:
task_id (str): The ID of the task to replay from.
long (bool): Whether to reset the long-term memory.
short (bool): Whether to reset the short-term memory.
entity (bool): Whether to reset the entity memory.
kickoff_outputs (bool): Whether to reset the latest kickoff task outputs.
all (bool): Whether to reset all memories.
"""
try:

View File

@@ -5,6 +5,7 @@ research_task:
the current year is 2024.
expected_output: >
A list with 10 bullet points of the most relevant information about {topic}
agent: researcher
reporting_task:
description: >
@@ -13,3 +14,4 @@ reporting_task:
expected_output: >
A fully fledge reports with the mains topics, each with a full section of information.
Formatted as markdown without '```'
agent: reporting_analyst

View File

@@ -32,14 +32,12 @@ class {{crew_name}}Crew():
def research_task(self) -> Task:
return Task(
config=self.tasks_config['research_task'],
agent=self.researcher()
)
@task
def reporting_task(self) -> Task:
return Task(
config=self.tasks_config['reporting_task'],
agent=self.reporting_analyst(),
output_file='report.md'
)

View File

@@ -39,3 +39,16 @@ def replay():
except Exception as e:
raise Exception(f"An error occurred while replaying the crew: {e}")
def test():
"""
Test the crew execution and returns the results.
"""
inputs = {
"topic": "AI LLMs"
}
try:
{{crew_name}}Crew().crew().test(n_iterations=int(sys.argv[1]), openai_model_name=sys.argv[2], inputs=inputs)
except Exception as e:
raise Exception(f"An error occurred while replaying the crew: {e}")

View File

@@ -6,12 +6,13 @@ authors = ["Your Name <you@example.com>"]
[tool.poetry.dependencies]
python = ">=3.10,<=3.13"
crewai = { extras = ["tools"], version = "^0.41.0" }
crewai = { extras = ["tools"], version = "^0.46.0" }
[tool.poetry.scripts]
{{folder_name}} = "{{folder_name}}.main:run"
train = "{{folder_name}}.main:train"
replay = "{{folder_name}}.main:replay"
test = "{{folder_name}}.main:test"
[build-system]
requires = ["poetry-core"]

View File

@@ -37,6 +37,7 @@ from crewai.utilities.constants import (
TRAINED_AGENTS_DATA_FILE,
TRAINING_DATA_FILE,
)
from crewai.utilities.evaluators.crew_evaluator_handler import CrewEvaluator
from crewai.utilities.evaluators.task_evaluator import TaskEvaluator
from crewai.utilities.formatter import (
aggregate_raw_outputs_from_task_outputs,
@@ -154,6 +155,10 @@ class Crew(BaseModel):
default=False,
description="Plan the crew execution and add the plan to the crew.",
)
planning_llm: Optional[Any] = Field(
default=None,
description="Language model that will run the AgentPlanner if planning is True.",
)
task_execution_output_json_files: Optional[List[str]] = Field(
default=None,
description="List of file paths for task execution JSON files.",
@@ -266,20 +271,6 @@ class Crew(BaseModel):
return self
@model_validator(mode="after")
def check_tasks_in_hierarchical_process_not_async(self):
"""Validates that the tasks in hierarchical process are not flagged with async_execution."""
if self.process == Process.hierarchical:
for task in self.tasks:
if task.async_execution:
raise PydanticCustomError(
"async_execution_in_hierarchical_process",
"Hierarchical process error: Tasks cannot be flagged with async_execution.",
{},
)
return self
@model_validator(mode="after")
def validate_end_with_at_most_one_async_task(self):
"""Validates that the crew ends with at most one asynchronous task."""
@@ -559,7 +550,9 @@ class Crew(BaseModel):
def _handle_crew_planning(self):
"""Handles the Crew planning."""
self._logger.log("info", "Planning the crew execution")
result = CrewPlanner(self.tasks)._handle_crew_planning()
result = CrewPlanner(
tasks=self.tasks, planning_agent_llm=self.planning_llm
)._handle_crew_planning()
for task, step_plan in zip(self.tasks, result.list_of_plans_per_task):
task.description += step_plan
@@ -600,7 +593,7 @@ class Crew(BaseModel):
def _run_hierarchical_process(self) -> CrewOutput:
"""Creates and assigns a manager agent to make sure the crew completes the tasks."""
self._create_manager_agent()
return self._execute_tasks(self.tasks, self.manager_agent)
return self._execute_tasks(self.tasks)
def _create_manager_agent(self):
i18n = I18N(prompt_file=self.prompt_file)
@@ -624,7 +617,6 @@ class Crew(BaseModel):
def _execute_tasks(
self,
tasks: List[Task],
manager: Optional[BaseAgent] = None,
start_index: Optional[int] = 0,
was_replayed: bool = False,
) -> CrewOutput:
@@ -652,13 +644,13 @@ class Crew(BaseModel):
last_sync_output = task.output
continue
agent_to_use = self._get_agent_to_use(task, manager)
agent_to_use = self._get_agent_to_use(task)
if agent_to_use is None:
raise ValueError(
f"No agent available for task: {task.description}. Ensure that either the task has an assigned agent or a manager agent is provided."
)
self._prepare_agent_tools(task, manager)
self._prepare_agent_tools(task)
self._log_task_start(task, agent_to_use.role)
if isinstance(task, ConditionalTask):
@@ -724,20 +716,18 @@ class Crew(BaseModel):
return skipped_task_output
return None
def _prepare_agent_tools(self, task: Task, manager: Optional[BaseAgent]):
def _prepare_agent_tools(self, task: Task):
if self.process == Process.hierarchical:
if manager:
self._update_manager_tools(task, manager)
if self.manager_agent:
self._update_manager_tools(task)
else:
raise ValueError("Manager agent is required for hierarchical process.")
elif task.agent and task.agent.allow_delegation:
self._add_delegation_tools(task)
def _get_agent_to_use(
self, task: Task, manager: Optional[BaseAgent]
) -> Optional[BaseAgent]:
def _get_agent_to_use(self, task: Task) -> Optional[BaseAgent]:
if self.process == Process.hierarchical:
return manager
return self.manager_agent
return task.agent
def _add_delegation_tools(self, task: Task):
@@ -773,11 +763,14 @@ class Crew(BaseModel):
if self.output_log_file:
self._file_handler.log(agent=role, task=task.description, status="started")
def _update_manager_tools(self, task: Task, manager: BaseAgent):
if task.agent:
manager.tools = task.agent.get_delegation_tools([task.agent])
else:
manager.tools = manager.get_delegation_tools(self.agents)
def _update_manager_tools(self, task: Task):
if self.manager_agent:
if task.agent:
self.manager_agent.tools = task.agent.get_delegation_tools([task.agent])
else:
self.manager_agent.tools = self.manager_agent.get_delegation_tools(
self.agents
)
def _get_context(self, task: Task, task_outputs: List[TaskOutput]):
context = (
@@ -876,7 +869,7 @@ class Crew(BaseModel):
self.tasks[i].output = task_output
self._logging_color = "bold_blue"
result = self._execute_tasks(self.tasks, self.manager_agent, start_index, True)
result = self._execute_tasks(self.tasks, start_index, True)
return result
def copy(self):
@@ -961,5 +954,20 @@ class Crew(BaseModel):
return total_usage_metrics
def test(
self,
n_iterations: int,
openai_model_name: str,
inputs: Optional[Dict[str, Any]] = None,
) -> None:
"""Test and evaluate the Crew with the given inputs for n iterations."""
evaluator = CrewEvaluator(self, openai_model_name)
for i in range(1, n_iterations + 1):
evaluator.set_iteration(i)
self.kickoff(inputs=inputs)
evaluator.print_crew_evaluation_result()
def __repr__(self):
return f"Crew(id={self.id}, process={self.process}, number_of_agents={len(self.agents)}, number_of_tasks={len(self.tasks)})"

View File

@@ -1,3 +1,4 @@
from typing import Any, Dict, Optional
from crewai.memory.memory import Memory
from crewai.memory.short_term.short_term_memory_item import ShortTermMemoryItem
from crewai.memory.storage.rag_storage import RAGStorage
@@ -18,8 +19,15 @@ class ShortTermMemory(Memory):
)
super().__init__(storage)
def save(self, item: ShortTermMemoryItem) -> None:
super().save(item.data, item.metadata, item.agent)
def save(
self,
value: Any,
metadata: Optional[Dict[str, Any]] = None,
agent: Optional[str] = None,
) -> None:
item = ShortTermMemoryItem(data=value, metadata=metadata, agent=agent)
super().save(value=item.data, metadata=item.metadata, agent=item.agent)
def search(self, query: str, score_threshold: float = 0.35):
return self.storage.search(query=query, score_threshold=score_threshold) # type: ignore # BUG? The reference is to the parent class, but the parent class does not have this parameters

View File

@@ -3,7 +3,10 @@ from typing import Any, Dict, Optional
class ShortTermMemoryItem:
def __init__(
self, data: Any, agent: str, metadata: Optional[Dict[str, Any]] = None
self,
data: Any,
agent: Optional[str] = None,
metadata: Optional[Dict[str, Any]] = None,
):
self.data = data
self.agent = agent

View File

@@ -4,7 +4,7 @@ from typing import Any, Dict
class Storage:
"""Abstract base class defining the storage interface"""
def save(self, key: str, value: Any, metadata: Dict[str, Any]) -> None:
def save(self, value: Any, metadata: Dict[str, Any]) -> None:
pass
def search(self, key: str) -> Dict[str, Any]: # type: ignore

View File

@@ -1,2 +1,25 @@
from .annotations import agent, crew, task
from .annotations import (
agent,
crew,
task,
output_json,
output_pydantic,
tool,
callback,
llm,
cache_handler,
)
from .crew_base import CrewBase
__all__ = [
"agent",
"crew",
"task",
"output_json",
"output_pydantic",
"tool",
"callback",
"CrewBase",
"llm",
"cache_handler",
]

View File

@@ -30,6 +30,37 @@ def agent(func):
return func
def llm(func):
func.is_llm = True
func = memoize(func)
return func
def output_json(cls):
cls.is_output_json = True
return cls
def output_pydantic(cls):
cls.is_output_pydantic = True
return cls
def tool(func):
func.is_tool = True
return memoize(func)
def callback(func):
func.is_callback = True
return memoize(func)
def cache_handler(func):
func.is_cache_handler = True
return memoize(func)
def crew(func):
def wrapper(self, *args, **kwargs):
instantiated_tasks = []

View File

@@ -1,6 +1,7 @@
import inspect
import os
from pathlib import Path
from typing import Any, Callable, Dict
import yaml
from dotenv import load_dotenv
@@ -20,11 +21,6 @@ def CrewBase(cls):
base_directory = Path(frame_info.filename).parent.resolve()
break
if base_directory is None:
raise Exception(
"Unable to dynamically determine the project's base directory, you must run it from the project's root directory."
)
original_agents_config_path = getattr(
cls, "agents_config", "config/agents.yaml"
)
@@ -32,12 +28,20 @@ def CrewBase(cls):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
if self.base_directory is None:
raise Exception(
"Unable to dynamically determine the project's base directory, you must run it from the project's root directory."
)
self.agents_config = self.load_yaml(
os.path.join(self.base_directory, self.original_agents_config_path)
)
self.tasks_config = self.load_yaml(
os.path.join(self.base_directory, self.original_tasks_config_path)
)
self.map_all_agent_variables()
self.map_all_task_variables()
@staticmethod
def load_yaml(config_path: str):
@@ -45,4 +49,138 @@ def CrewBase(cls):
# parsedContent = YamlParser.parse(file) # type: ignore # Argument 1 to "parse" has incompatible type "TextIOWrapper"; expected "YamlParser"
return yaml.safe_load(file)
def _get_all_functions(self):
return {
name: getattr(self, name)
for name in dir(self)
if callable(getattr(self, name))
}
def _filter_functions(
self, functions: Dict[str, Callable], attribute: str
) -> Dict[str, Callable]:
return {
name: func
for name, func in functions.items()
if hasattr(func, attribute)
}
def map_all_agent_variables(self) -> None:
all_functions = self._get_all_functions()
llms = self._filter_functions(all_functions, "is_llm")
tool_functions = self._filter_functions(all_functions, "is_tool")
cache_handler_functions = self._filter_functions(
all_functions, "is_cache_handler"
)
callbacks = self._filter_functions(all_functions, "is_callback")
agents = self._filter_functions(all_functions, "is_agent")
for agent_name, agent_info in self.agents_config.items():
self._map_agent_variables(
agent_name,
agent_info,
agents,
llms,
tool_functions,
cache_handler_functions,
callbacks,
)
def _map_agent_variables(
self,
agent_name: str,
agent_info: Dict[str, Any],
agents: Dict[str, Callable],
llms: Dict[str, Callable],
tool_functions: Dict[str, Callable],
cache_handler_functions: Dict[str, Callable],
callbacks: Dict[str, Callable],
) -> None:
if llm := agent_info.get("llm"):
self.agents_config[agent_name]["llm"] = llms[llm]()
if tools := agent_info.get("tools"):
self.agents_config[agent_name]["tools"] = [
tool_functions[tool]() for tool in tools
]
if function_calling_llm := agent_info.get("function_calling_llm"):
self.agents_config[agent_name]["function_calling_llm"] = agents[
function_calling_llm
]()
if step_callback := agent_info.get("step_callback"):
self.agents_config[agent_name]["step_callback"] = callbacks[
step_callback
]()
if cache_handler := agent_info.get("cache_handler"):
self.agents_config[agent_name]["cache_handler"] = (
cache_handler_functions[cache_handler]()
)
def map_all_task_variables(self) -> None:
all_functions = self._get_all_functions()
agents = self._filter_functions(all_functions, "is_agent")
tasks = self._filter_functions(all_functions, "is_task")
output_json_functions = self._filter_functions(
all_functions, "is_output_json"
)
tool_functions = self._filter_functions(all_functions, "is_tool")
callback_functions = self._filter_functions(all_functions, "is_callback")
output_pydantic_functions = self._filter_functions(
all_functions, "is_output_pydantic"
)
for task_name, task_info in self.tasks_config.items():
self._map_task_variables(
task_name,
task_info,
agents,
tasks,
output_json_functions,
tool_functions,
callback_functions,
output_pydantic_functions,
)
def _map_task_variables(
self,
task_name: str,
task_info: Dict[str, Any],
agents: Dict[str, Callable],
tasks: Dict[str, Callable],
output_json_functions: Dict[str, Callable],
tool_functions: Dict[str, Callable],
callback_functions: Dict[str, Callable],
output_pydantic_functions: Dict[str, Callable],
) -> None:
if context_list := task_info.get("context"):
self.tasks_config[task_name]["context"] = [
tasks[context_task_name]() for context_task_name in context_list
]
if tools := task_info.get("tools"):
self.tasks_config[task_name]["tools"] = [
tool_functions[tool]() for tool in tools
]
if agent_name := task_info.get("agent"):
self.tasks_config[task_name]["agent"] = agents[agent_name]()
if output_json := task_info.get("output_json"):
self.tasks_config[task_name]["output_json"] = output_json_functions[
output_json
]
if output_pydantic := task_info.get("output_pydantic"):
self.tasks_config[task_name]["output_pydantic"] = (
output_pydantic_functions[output_pydantic]
)
if callbacks := task_info.get("callbacks"):
self.tasks_config[task_name]["callbacks"] = [
callback_functions[callback]() for callback in callbacks
]
return WrappedClass

View File

@@ -1,6 +1,6 @@
import datetime
import json
import os
import re
import threading
import uuid
from concurrent.futures import Future
@@ -8,7 +8,6 @@ from copy import copy
from hashlib import md5
from typing import Any, Dict, List, Optional, Tuple, Type, Union
from langchain_openai import ChatOpenAI
from opentelemetry.trace import Span
from pydantic import UUID4, BaseModel, Field, field_validator, model_validator
from pydantic_core import PydanticCustomError
@@ -17,10 +16,8 @@ from crewai.agents.agent_builder.base_agent import BaseAgent
from crewai.tasks.output_format import OutputFormat
from crewai.tasks.task_output import TaskOutput
from crewai.telemetry.telemetry import Telemetry
from crewai.utilities.converter import Converter, ConverterError
from crewai.utilities.converter import Converter, convert_to_model
from crewai.utilities.i18n import I18N
from crewai.utilities.printer import Printer
from crewai.utilities.pydantic_schema_parser import PydanticSchemaParser
class Task(BaseModel):
@@ -111,6 +108,7 @@ class Task(BaseModel):
_original_description: str | None = None
_original_expected_output: str | None = None
_thread: threading.Thread | None = None
_execution_time: float | None = None
def __init__(__pydantic_self__, **data):
config = data.pop("config", {})
@@ -124,9 +122,15 @@ class Task(BaseModel):
"may_not_set_field", "This field is not to be set by the user.", {}
)
def _set_start_execution_time(self) -> float:
return datetime.datetime.now().timestamp()
def _set_end_execution_time(self, start_time: float) -> None:
self._execution_time = datetime.datetime.now().timestamp() - start_time
@field_validator("output_file")
@classmethod
def output_file_validattion(cls, value: str) -> str:
def output_file_validation(cls, value: str) -> str:
"""Validate the output file path by removing the / from the beginning of the path."""
if value.startswith("/"):
return value[1:]
@@ -220,6 +224,7 @@ class Task(BaseModel):
f"The task '{self.description}' has no agent assigned, therefore it can't be executed directly and should be executed in a Crew using a specific process that support that, like hierarchical."
)
start_time = self._set_start_execution_time()
self._execution_span = self._telemetry.task_started(crew=agent.crew, task=self)
self.prompt_context = context
@@ -243,6 +248,7 @@ class Task(BaseModel):
)
self.output = task_output
self._set_end_execution_time(start_time)
if self.callback:
self.callback(self.output)
@@ -326,18 +332,6 @@ class Task(BaseModel):
return copied_task
def _create_converter(self, *args, **kwargs) -> Converter:
"""Create a converter instance."""
if self.agent and not self.converter_cls:
converter = self.agent.get_output_converter(*args, **kwargs)
elif self.converter_cls:
converter = self.converter_cls(*args, **kwargs)
if not converter:
raise Exception("No output converter found or set.")
return converter
def _export_output(
self, result: str
) -> Tuple[Optional[BaseModel], Optional[Dict[str, Any]]]:
@@ -345,75 +339,26 @@ class Task(BaseModel):
json_output: Optional[Dict[str, Any]] = None
if self.output_pydantic or self.output_json:
model_output = self._convert_to_model(result)
pydantic_output = (
model_output if isinstance(model_output, BaseModel) else None
model_output = convert_to_model(
result,
self.output_pydantic,
self.output_json,
self.agent,
self.converter_cls,
)
if isinstance(model_output, str):
if isinstance(model_output, BaseModel):
pydantic_output = model_output
elif isinstance(model_output, dict):
json_output = model_output
elif isinstance(model_output, str):
try:
json_output = json.loads(model_output)
except json.JSONDecodeError:
json_output = None
else:
json_output = model_output if isinstance(model_output, dict) else None
return pydantic_output, json_output
def _convert_to_model(self, result: str) -> Union[dict, BaseModel, str]:
model = self.output_pydantic or self.output_json
if model is None:
return result
try:
return self._validate_model(result, model)
except Exception:
return self._handle_partial_json(result, model)
def _validate_model(
self, result: str, model: Type[BaseModel]
) -> Union[dict, BaseModel]:
exported_result = model.model_validate_json(result)
if self.output_json:
return exported_result.model_dump()
return exported_result
def _handle_partial_json(
self, result: str, model: Type[BaseModel]
) -> Union[dict, BaseModel, str]:
match = re.search(r"({.*})", result, re.DOTALL)
if match:
try:
exported_result = model.model_validate_json(match.group(0))
if self.output_json:
return exported_result.model_dump()
return exported_result
except Exception:
pass
return self._convert_with_instructions(result, model)
def _convert_with_instructions(
self, result: str, model: Type[BaseModel]
) -> Union[dict, BaseModel, str]:
llm = self.agent.function_calling_llm or self.agent.llm # type: ignore # Item "None" of "BaseAgent | None" has no attribute "function_calling_llm"
instructions = self._get_conversion_instructions(model, llm)
converter = self._create_converter(
llm=llm, text=result, model=model, instructions=instructions
)
exported_result = (
converter.to_pydantic() if self.output_pydantic else converter.to_json()
)
if isinstance(exported_result, ConverterError):
Printer().print(
content=f"{exported_result.message} Using raw output instead.",
color="red",
)
return result
return exported_result
def _get_output_format(self) -> OutputFormat:
if self.output_json:
return OutputFormat.JSON
@@ -421,34 +366,22 @@ class Task(BaseModel):
return OutputFormat.PYDANTIC
return OutputFormat.RAW
def _get_conversion_instructions(self, model: Type[BaseModel], llm: Any) -> str:
instructions = "I'm gonna convert this raw text into valid JSON."
if not self._is_gpt(llm):
model_schema = PydanticSchemaParser(model=model).get_schema()
instructions = f"{instructions}\n\nThe json should have the following structure, with the following keys:\n{model_schema}"
return instructions
def _save_output(self, content: str) -> None:
if not self.output_file:
raise Exception("Output file path is not set.")
directory = os.path.dirname(self.output_file)
if directory and not os.path.exists(directory):
os.makedirs(directory)
with open(self.output_file, "w", encoding="utf-8") as file:
file.write(content)
def _is_gpt(self, llm) -> bool:
return isinstance(llm, ChatOpenAI) and llm.openai_api_base is None
def _save_file(self, result: Any) -> None:
if self.output_file is None:
raise ValueError("output_file is not set.")
directory = os.path.dirname(self.output_file) # type: ignore # Value of type variable "AnyOrLiteralStr" of "dirname" cannot be "str | None"
if directory and not os.path.exists(directory):
os.makedirs(directory)
with open(self.output_file, "w", encoding="utf-8") as file: # type: ignore # Argument 1 to "open" has incompatible type "str | None"; expected "int | str | bytes | PathLike[str] | PathLike[bytes]"
file.write(result)
with open(self.output_file, "w", encoding="utf-8") as file:
if isinstance(result, dict):
import json
json.dump(result, file, ensure_ascii=False, indent=2)
else:
file.write(str(result))
return None
def __repr__(self):

View File

@@ -40,7 +40,7 @@ class Telemetry:
- Roles of agents in a crew
- Tools names available
Users can opt-in to sharing more complete data suing the `share_crew`
Users can opt-in to sharing more complete data using the `share_crew`
attribute in the Crew class.
"""

View File

@@ -16,7 +16,7 @@ try:
except ImportError:
agentops = None
OPENAI_BIGGER_MODELS = ["gpt-4"]
OPENAI_BIGGER_MODELS = ["gpt-4o"]
class ToolUsageErrorException(Exception):
@@ -86,7 +86,8 @@ class ToolUsage:
) -> str:
if isinstance(calling, ToolUsageErrorException):
error = calling.message
self._printer.print(content=f"\n\n{error}\n", color="red")
if self.agent.verbose:
self._printer.print(content=f"\n\n{error}\n", color="red")
self.task.increment_tools_errors()
return error
@@ -96,7 +97,8 @@ class ToolUsage:
except Exception as e:
error = getattr(e, "message", str(e))
self.task.increment_tools_errors()
self._printer.print(content=f"\n\n{error}\n", color="red")
if self.agent.verbose:
self._printer.print(content=f"\n\n{error}\n", color="red")
return error
return f"{self._use(tool_string=tool_string, tool=tool, calling=calling)}" # type: ignore # BUG?: "_use" of "ToolUsage" does not return a value (it only ever returns None)
@@ -112,7 +114,8 @@ class ToolUsage:
result = self._i18n.errors("task_repeated_usage").format(
tool_names=self.tools_names
)
self._printer.print(content=f"\n\n{result}\n", color="purple")
if self.agent.verbose:
self._printer.print(content=f"\n\n{result}\n", color="purple")
self._telemetry.tool_repeated_usage(
llm=self.function_calling_llm,
tool_name=tool.name,
@@ -168,7 +171,10 @@ class ToolUsage:
f'\n{error_message}.\nMoving on then. {self._i18n.slice("format").format(tool_names=self.tools_names)}'
).message
self.task.increment_tools_errors()
self._printer.print(content=f"\n\n{error_message}\n", color="red")
if self.agent.verbose:
self._printer.print(
content=f"\n\n{error_message}\n", color="red"
)
return error # type: ignore # No return value expected
self.task.increment_tools_errors()
@@ -192,7 +198,8 @@ class ToolUsage:
calling=calling, output=result, should_cache=should_cache
)
self._printer.print(content=f"\n\n{result}\n", color="purple")
if self.agent.verbose:
self._printer.print(content=f"\n\n{result}\n", color="purple")
if agentops:
agentops.record(tool_event)
self._telemetry.tool_usage(
@@ -346,7 +353,8 @@ class ToolUsage:
if self._run_attempts > self._max_parsing_attempts:
self._telemetry.tool_usage_error(llm=self.function_calling_llm)
self.task.increment_tools_errors()
self._printer.print(content=f"\n\n{e}\n", color="red")
if self.agent.verbose:
self._printer.print(content=f"\n\n{e}\n", color="red")
return ToolUsageErrorException( # type: ignore # Incompatible return value type (got "ToolUsageErrorException", expected "ToolCalling | InstructorToolCalling")
f'{self._i18n.errors("tool_usage_error").format(error=e)}\nMoving on then. {self._i18n.slice("format").format(tool_names=self.tools_names)}'
)

View File

@@ -7,6 +7,9 @@ from .parser import YamlParser
from .printer import Printer
from .prompts import Prompts
from .rpm_controller import RPMController
from .exceptions.context_window_exceeding_exception import (
LLMContextLengthExceededException,
)
__all__ = [
"Converter",
@@ -19,4 +22,5 @@ __all__ = [
"Prompts",
"RPMController",
"YamlParser",
"LLMContextLengthExceededException",
]

View File

@@ -1,9 +1,14 @@
import json
import re
from typing import Any, Optional, Type, Union
from langchain.schema import HumanMessage, SystemMessage
from langchain_openai import ChatOpenAI
from pydantic import BaseModel, ValidationError
from crewai.agents.agent_builder.utilities.base_output_converter import OutputConverter
from crewai.utilities.printer import Printer
from crewai.utilities.pydantic_schema_parser import PydanticSchemaParser
class ConverterError(Exception):
@@ -72,3 +77,153 @@ class Converter(OutputConverter):
def is_gpt(self) -> bool:
"""Return if llm provided is of gpt from openai."""
return isinstance(self.llm, ChatOpenAI) and self.llm.openai_api_base is None
def convert_to_model(
result: str,
output_pydantic: Optional[Type[BaseModel]],
output_json: Optional[Type[BaseModel]],
agent: Any,
converter_cls: Optional[Type[Converter]] = None,
) -> Union[dict, BaseModel, str]:
model = output_pydantic or output_json
if model is None:
return result
try:
escaped_result = json.dumps(json.loads(result, strict=False))
return validate_model(escaped_result, model, bool(output_json))
except json.JSONDecodeError as e:
Printer().print(
content=f"Error parsing JSON: {e}. Attempting to handle partial JSON.",
color="yellow",
)
return handle_partial_json(
result, model, bool(output_json), agent, converter_cls
)
except ValidationError as e:
Printer().print(
content=f"Pydantic validation error: {e}. Attempting to handle partial JSON.",
color="yellow",
)
return handle_partial_json(
result, model, bool(output_json), agent, converter_cls
)
except Exception as e:
Printer().print(
content=f"Unexpected error during model conversion: {type(e).__name__}: {e}. Returning original result.",
color="red",
)
return result
def validate_model(
result: str, model: Type[BaseModel], is_json_output: bool
) -> Union[dict, BaseModel]:
exported_result = model.model_validate_json(result)
if is_json_output:
return exported_result.model_dump()
return exported_result
def handle_partial_json(
result: str,
model: Type[BaseModel],
is_json_output: bool,
agent: Any,
converter_cls: Optional[Type[Converter]] = None,
) -> Union[dict, BaseModel, str]:
match = re.search(r"({.*})", result, re.DOTALL)
if match:
try:
exported_result = model.model_validate_json(match.group(0))
if is_json_output:
return exported_result.model_dump()
return exported_result
except json.JSONDecodeError as e:
Printer().print(
content=f"Error parsing JSON: {e}. The extracted JSON-like string is not valid JSON. Attempting alternative conversion method.",
color="yellow",
)
except ValidationError as e:
Printer().print(
content=f"Pydantic validation error: {e}. The JSON structure doesn't match the expected model. Attempting alternative conversion method.",
color="yellow",
)
except Exception as e:
Printer().print(
content=f"Unexpected error during partial JSON handling: {type(e).__name__}: {e}. Attempting alternative conversion method.",
color="red",
)
return convert_with_instructions(
result, model, is_json_output, agent, converter_cls
)
def convert_with_instructions(
result: str,
model: Type[BaseModel],
is_json_output: bool,
agent: Any,
converter_cls: Optional[Type[Converter]] = None,
) -> Union[dict, BaseModel, str]:
llm = agent.function_calling_llm or agent.llm
instructions = get_conversion_instructions(model, llm)
converter = create_converter(
agent=agent,
converter_cls=converter_cls,
llm=llm,
text=result,
model=model,
instructions=instructions,
)
exported_result = (
converter.to_pydantic() if not is_json_output else converter.to_json()
)
if isinstance(exported_result, ConverterError):
Printer().print(
content=f"{exported_result.message} Using raw output instead.",
color="red",
)
return result
return exported_result
def get_conversion_instructions(model: Type[BaseModel], llm: Any) -> str:
instructions = "I'm gonna convert this raw text into valid JSON."
if not is_gpt(llm):
model_schema = PydanticSchemaParser(model=model).get_schema()
instructions = f"{instructions}\n\nThe json should have the following structure, with the following keys:\n{model_schema}"
return instructions
def is_gpt(llm: Any) -> bool:
from langchain_openai import ChatOpenAI
return isinstance(llm, ChatOpenAI) and llm.openai_api_base is None
def create_converter(
agent: Optional[Any] = None,
converter_cls: Optional[Type[Converter]] = None,
*args,
**kwargs,
) -> Converter:
if agent and not converter_cls:
if hasattr(agent, "get_output_converter"):
converter = agent.get_output_converter(*args, **kwargs)
else:
raise AttributeError("Agent does not have a 'get_output_converter' method")
elif converter_cls:
converter = converter_cls(*args, **kwargs)
else:
raise ValueError("Either agent or converter_cls must be provided")
if not converter:
raise Exception("No output converter found or set.")
return converter

View File

@@ -1,5 +1,5 @@
import json
from typing import Any, List, Type, Union
from typing import Any, List, Type
import regex
from langchain.output_parsers import PydanticOutputParser
@@ -7,29 +7,24 @@ from langchain_core.exceptions import OutputParserException
from langchain_core.outputs import Generation
from langchain_core.pydantic_v1 import ValidationError
from pydantic import BaseModel
from pydantic.v1 import BaseModel as V1BaseModel
class CrewPydanticOutputParser(PydanticOutputParser):
"""Parses the text into pydantic models"""
pydantic_object: Union[Type[BaseModel], Type[V1BaseModel]]
pydantic_object: Type[BaseModel]
def parse_result(self, result: List[Generation], *, partial: bool = False) -> Any:
def parse_result(self, result: List[Generation]) -> Any:
result[0].text = self._transform_in_valid_json(result[0].text)
# Treating edge case of function calling llm returning the name instead of tool_name
json_object = json.loads(result[0].text)
json_object["tool_name"] = (
json_object["name"]
if "tool_name" not in json_object
else json_object["tool_name"]
)
if "tool_name" not in json_object:
json_object["tool_name"] = json_object.get("name", "")
result[0].text = json.dumps(json_object)
json_object = super().parse_result(result)
try:
return self.pydantic_object.parse_obj(json_object)
return self.pydantic_object.model_validate(json_object)
except ValidationError as e:
name = self.pydantic_object.__name__
msg = f"Failed to parse {name} from completion {json_object}. Got: {e}"

View File

@@ -0,0 +1,163 @@
from collections import defaultdict
from langchain_openai import ChatOpenAI
from pydantic import BaseModel, Field
from rich.console import Console
from rich.table import Table
from crewai.agent import Agent
from crewai.task import Task
from crewai.tasks.task_output import TaskOutput
class TaskEvaluationPydanticOutput(BaseModel):
quality: float = Field(
description="A score from 1 to 10 evaluating on completion, quality, and overall performance from the task_description and task_expected_output to the actual Task Output."
)
class CrewEvaluator:
"""
A class to evaluate the performance of the agents in the crew based on the tasks they have performed.
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).
tasks_scores (defaultdict): A dictionary to store the scores of the agents for each task.
iteration (int): The current iteration of the evaluation.
"""
tasks_scores: defaultdict = defaultdict(list)
run_execution_times: defaultdict = defaultdict(list)
iteration: int = 0
def __init__(self, crew, openai_model_name: str):
self.crew = crew
self.openai_model_name = openai_model_name
self._setup_for_evaluating()
def _setup_for_evaluating(self) -> None:
"""Sets up the crew for evaluating."""
for task in self.crew.tasks:
task.callback = self.evaluate
def _evaluator_agent(self):
return Agent(
role="Task Execution Evaluator",
goal=(
"Your goal is to evaluate the performance of the agents in the crew based on the tasks they have performed using score from 1 to 10 evaluating on completion, quality, and overall performance."
),
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=ChatOpenAI(model=self.openai_model_name),
)
def _evaluation_task(
self, evaluator_agent: Agent, task_to_evaluate: Task, task_output: str
) -> Task:
return Task(
description=(
"Based on the task description and the expected output, compare and evaluate the performance of the agents in the crew based on the Task Output they have performed using score from 1 to 10 evaluating on completion, quality, and overall performance."
f"task_description: {task_to_evaluate.description} "
f"task_expected_output: {task_to_evaluate.expected_output} "
f"agent: {task_to_evaluate.agent.role if task_to_evaluate.agent else None} "
f"agent_goal: {task_to_evaluate.agent.goal if task_to_evaluate.agent else None} "
f"Task Output: {task_output}"
),
expected_output="Evaluation Score from 1 to 10 based on the performance of the agents on the tasks",
agent=evaluator_agent,
output_pydantic=TaskEvaluationPydanticOutput,
)
def set_iteration(self, iteration: int) -> None:
self.iteration = iteration
def print_crew_evaluation_result(self) -> None:
"""
Prints the evaluation result of the crew in a table.
A Crew with 2 tasks using the command crewai test -n 2
will output the following table:
Task Scores
(1-10 Higher is better)
┏━━━━━━━━━━━━┳━━━━━━━┳━━━━━━━┳━━━━━━━━━━━━┓
┃ Tasks/Crew ┃ Run 1 ┃ Run 2 ┃ Avg. Total ┃
┡━━━━━━━━━━━━╇━━━━━━━╇━━━━━━━╇━━━━━━━━━━━━┩
│ Task 1 │ 10.0 │ 9.0 │ 9.5 │
│ Task 2 │ 9.0 │ 9.0 │ 9.0 │
│ Crew │ 9.5 │ 9.0 │ 9.2 │
└────────────┴───────┴───────┴────────────┘
"""
task_averages = [
sum(scores) / len(scores) for scores in zip(*self.tasks_scores.values())
]
crew_average = sum(task_averages) / len(task_averages)
# Create a table
table = Table(title="Tasks Scores \n (1-10 Higher is better)")
# Add columns for the table
table.add_column("Tasks/Crew")
for run in range(1, len(self.tasks_scores) + 1):
table.add_column(f"Run {run}")
table.add_column("Avg. Total")
# Add rows for each task
for task_index in range(len(task_averages)):
task_scores = [
self.tasks_scores[run][task_index]
for run in range(1, len(self.tasks_scores) + 1)
]
avg_score = task_averages[task_index]
table.add_row(
f"Task {task_index + 1}", *map(str, task_scores), f"{avg_score:.1f}"
)
# Add a row for the crew average
crew_scores = [
sum(self.tasks_scores[run]) / len(self.tasks_scores[run])
for run in range(1, len(self.tasks_scores) + 1)
]
table.add_row("Crew", *map(str, crew_scores), f"{crew_average:.1f}")
run_exec_times = [
int(sum(tasks_exec_times))
for _, tasks_exec_times in self.run_execution_times.items()
]
execution_time_avg = int(sum(run_exec_times) / len(run_exec_times))
table.add_row(
"Execution Time (s)",
*map(str, run_exec_times),
f"{execution_time_avg}",
)
# Display the table in the terminal
console = Console()
console.print(table)
def evaluate(self, task_output: TaskOutput):
"""Evaluates the performance of the agents in the crew based on the tasks they have performed."""
current_task = None
for task in self.crew.tasks:
if task.description == task_output.description:
current_task = task
break
if not current_task or not task_output:
raise ValueError(
"Task to evaluate and task output are required for evaluation"
)
evaluator_agent = self._evaluator_agent()
evaluation_task = self._evaluation_task(
evaluator_agent, current_task, task_output.raw
)
evaluation_result = evaluation_task.execute_sync()
if isinstance(evaluation_result.pydantic, TaskEvaluationPydanticOutput):
self.tasks_scores[self.iteration].append(evaluation_result.pydantic.quality)
self.run_execution_times[self.iteration].append(
current_task._execution_time
)
else:
raise ValueError("Evaluation result is not in the expected format")

View File

@@ -54,23 +54,23 @@ class TaskEvaluator:
def __init__(self, original_agent):
self.llm = original_agent.llm
def evaluate(self, task, ouput) -> TaskEvaluation:
def evaluate(self, task, output) -> TaskEvaluation:
evaluation_query = (
f"Assess the quality of the task completed based on the description, expected output, and actual results.\n\n"
f"Task Description:\n{task.description}\n\n"
f"Expected Output:\n{task.expected_output}\n\n"
f"Actual Output:\n{ouput}\n\n"
f"Actual Output:\n{output}\n\n"
"Please provide:\n"
"- Bullet points suggestions to improve future similar tasks\n"
"- A score from 0 to 10 evaluating on completion, quality, and overall performance"
"- Entities extracted from the task output, if any, their type, description, and relationships"
)
instructions = "I'm gonna convert this raw text into valid JSON."
instructions = "Convert all responses into valid JSON output."
if not self._is_gpt(self.llm):
model_schema = PydanticSchemaParser(model=TaskEvaluation).get_schema()
instructions = f"{instructions}\n\nThe json should have the following structure, with the following keys:\n{model_schema}"
instructions = f"{instructions}\n\nReturn only valid JSON with the following schema:\n```json\n{model_schema}\n```"
converter = Converter(
llm=self.llm,

View File

@@ -0,0 +1,26 @@
class LLMContextLengthExceededException(Exception):
CONTEXT_LIMIT_ERRORS = [
"maximum context length",
"context length exceeded",
"context_length_exceeded",
"context window full",
"too many tokens",
"input is too long",
"exceeds token limit",
]
def __init__(self, error_message: str):
self.original_error_message = error_message
super().__init__(self._get_error_message(error_message))
def _is_context_limit_error(self, error_message: str) -> bool:
return any(
phrase.lower() in error_message.lower()
for phrase in self.CONTEXT_LIMIT_ERRORS
)
def _get_error_message(self, error_message: str):
return (
f"LLM context length exceeded. Original error: {error_message}\n"
"Consider using a smaller input or implementing a text splitting strategy."
)

View File

@@ -1,17 +1,28 @@
import re
class YamlParser:
@staticmethod
def parse(file):
"""
Parses a YAML file, modifies specific patterns, and checks for unsupported 'context' usage.
Args:
file (file object): The YAML file to parse.
Returns:
str: The modified content of the YAML file.
Raises:
ValueError: If 'context:' is used incorrectly.
"""
content = file.read()
# Replace single { and } with doubled ones, while leaving already doubled ones intact and the other special characters {# and {%
modified_content = re.sub(r"(?<!\{){(?!\{)(?!\#)(?!\%)", "{{", content)
modified_content = re.sub(
r"(?<!\})(?<!\%)(?<!\#)\}(?!})", "}}", modified_content
)
modified_content = re.sub(r"(?<!\})(?<!\%)(?<!\#)\}(?!})", "}}", modified_content)
# Check for 'context:' not followed by '[' and raise an error
if re.search(r"context:(?!\s*\[)", modified_content):
raise ValueError(
"Context is currently only supported in code when creating a task. Please use the 'context' key in the task configuration."
"Context is currently only supported in code when creating a task. "
"Please use the 'context' key in the task configuration."
)
return modified_content

View File

@@ -1,5 +1,6 @@
from typing import List
from typing import Any, List, Optional
from langchain_openai import ChatOpenAI
from pydantic import BaseModel
from crewai.agent import Agent
@@ -11,17 +12,27 @@ class PlannerTaskPydanticOutput(BaseModel):
class CrewPlanner:
def __init__(self, tasks: List[Task]):
def __init__(self, tasks: List[Task], planning_agent_llm: Optional[Any] = None):
self.tasks = tasks
def _handle_crew_planning(self):
if planning_agent_llm is None:
self.planning_agent_llm = ChatOpenAI(model="gpt-4o-mini")
else:
self.planning_agent_llm = planning_agent_llm
def _handle_crew_planning(self) -> PlannerTaskPydanticOutput:
"""Handles the Crew planning by creating detailed step-by-step plans for each task."""
planning_agent = self._create_planning_agent()
tasks_summary = self._create_tasks_summary()
planner_task = self._create_planner_task(planning_agent, tasks_summary)
return planner_task.execute_sync()
result = planner_task.execute_sync()
if isinstance(result.pydantic, PlannerTaskPydanticOutput):
return result.pydantic
raise ValueError("Failed to get the Planning output")
def _create_planning_agent(self) -> Agent:
"""Creates the planning agent for the crew planning."""
@@ -32,6 +43,7 @@ class CrewPlanner:
"available to each agent so that they can perform the tasks in an exemplary manner"
),
backstory="Planner agent for crew planning",
llm=self.planning_agent_llm,
)
def _create_planner_task(self, planning_agent: Agent, tasks_summary: str) -> Task:

View File

@@ -16,11 +16,13 @@ class PydanticSchemaParser(BaseModel):
return self._get_model_schema(self.model)
def _get_model_schema(self, model, depth=0) -> str:
lines = []
indent = " " * depth
lines = [f"{indent}{{"]
for field_name, field in model.model_fields.items():
field_type_str = self._get_field_type(field, depth + 1)
lines.append(f"{' ' * 4 * depth}- {field_name}: {field_type_str}")
lines.append(f"{indent} {field_name}: {field_type_str},")
lines[-1] = lines[-1].rstrip(",") # Remove trailing comma from last item
lines.append(f"{indent}}}")
return "\n".join(lines)
def _get_field_type(self, field, depth) -> str:
@@ -35,6 +37,6 @@ class PydanticSchemaParser(BaseModel):
else:
return f"List[{list_item_type.__name__}]"
elif issubclass(field_type, BaseModel):
return f"\n{self._get_model_schema(field_type, depth)}"
return self._get_model_schema(field_type, depth)
else:
return field_type.__name__

View File

@@ -10,24 +10,24 @@ from crewai.agents.agent_builder.utilities.base_token_process import TokenProces
class TokenCalcHandler(BaseCallbackHandler):
model_name: str = ""
token_cost_process: TokenProcess
encoding: tiktoken.Encoding
def __init__(self, model_name, token_cost_process):
self.model_name = model_name
self.token_cost_process = token_cost_process
try:
self.encoding = tiktoken.encoding_for_model(self.model_name)
except KeyError:
self.encoding = tiktoken.get_encoding("cl100k_base")
def on_llm_start(
self, serialized: Dict[str, Any], prompts: List[str], **kwargs: Any
) -> None:
try:
encoding = tiktoken.encoding_for_model(self.model_name)
except KeyError:
encoding = tiktoken.get_encoding("cl100k_base")
if self.token_cost_process is None:
return
for prompt in prompts:
self.token_cost_process.sum_prompt_tokens(len(encoding.encode(prompt)))
self.token_cost_process.sum_prompt_tokens(len(self.encoding.encode(prompt)))
async def on_llm_new_token(self, token: str, **kwargs) -> None:
self.token_cost_process.sum_completion_tokens(1)

View File

@@ -7,6 +7,7 @@ import pytest
from langchain.tools import tool
from langchain_core.exceptions import OutputParserException
from langchain_openai import ChatOpenAI
from langchain.schema import AgentAction
from crewai import Agent, Crew, Task
from crewai.agents.cache import CacheHandler
@@ -397,7 +398,7 @@ def test_agent_moved_on_after_max_iterations():
)
task = Task(
description="The final answer is 42. But don't give it yet, instead keep using the `get_final_answer` tool over and over until you're told you can give yout final answer.",
description="The final answer is 42. But don't give it yet, instead keep using the `get_final_answer` tool over and over until you're told you can give your final answer.",
expected_output="The final answer",
)
output = agent.execute_task(
@@ -948,7 +949,7 @@ def test_agent_use_trained_data(crew_training_handler):
crew_training_handler().load.return_value = {
agent.role: {
"suggestions": [
"The result of the math operatio must be right.",
"The result of the math operation must be right.",
"Result must be better than 1.",
]
}
@@ -958,7 +959,7 @@ def test_agent_use_trained_data(crew_training_handler):
assert (
result == "What is 1 + 1?You MUST follow these feedbacks: \n "
"The result of the math operatio must be right.\n - Result must be better than 1."
"The result of the math operation must be right.\n - Result must be better than 1."
)
crew_training_handler.assert_has_calls(
[mock.call(), mock.call("trained_agents_data.pkl"), mock.call().load()]
@@ -1014,3 +1015,75 @@ def test_agent_max_retry_limit():
),
]
)
@pytest.mark.vcr(filter_headers=["authorization"])
def test_handle_context_length_exceeds_limit():
agent = Agent(
role="test role",
goal="test goal",
backstory="test backstory",
)
original_action = AgentAction(
tool="test_tool", tool_input="test_input", log="test_log"
)
with patch.object(
CrewAgentExecutor, "_iter_next_step", wraps=agent.agent_executor._iter_next_step
) as private_mock:
task = Task(
description="The final answer is 42. But don't give it yet, instead keep using the `get_final_answer` tool.",
expected_output="The final answer",
)
agent.execute_task(
task=task,
)
private_mock.assert_called_once()
with patch("crewai.agents.executor.click") as mock_prompt:
mock_prompt.return_value = "y"
with patch.object(
CrewAgentExecutor, "_handle_context_length"
) as mock_handle_context:
mock_handle_context.side_effect = ValueError(
"Context length limit exceeded"
)
long_input = "This is a very long input. " * 10000
# Attempt to handle context length, expecting the mocked error
with pytest.raises(ValueError) as excinfo:
agent.agent_executor._handle_context_length(
[(original_action, long_input)]
)
assert "Context length limit exceeded" in str(excinfo.value)
mock_handle_context.assert_called_once()
@pytest.mark.vcr(filter_headers=["authorization"])
def test_handle_context_length_exceeds_limit_cli_no():
agent = Agent(
role="test role",
goal="test goal",
backstory="test backstory",
)
task = Task(description="test task", agent=agent, expected_output="test output")
with patch.object(
CrewAgentExecutor, "_iter_next_step", wraps=agent.agent_executor._iter_next_step
) as private_mock:
task = Task(
description="The final answer is 42. But don't give it yet, instead keep using the `get_final_answer` tool.",
expected_output="The final answer",
)
agent.execute_task(
task=task,
)
private_mock.assert_called_once()
with patch("crewai.agents.executor.click") as mock_prompt:
mock_prompt.return_value = "n"
pytest.raises(SystemExit)
with patch.object(
CrewAgentExecutor, "_handle_context_length"
) as mock_handle_context:
mock_handle_context.assert_not_called()

View File

@@ -0,0 +1,181 @@
interactions:
- request:
body: '{"messages": [{"content": "You are test role. test backstory\nYour personal
goal is: test goalTo give my best complete final answer to the task use the
exact following format:\n\nThought: I now can give a great answer\nFinal Answer:
my best complete final answer to the task.\nYour final answer must be the great
and the most complete as possible, it must be outcome described.\n\nI MUST use
these formats, my job depends on it!\nCurrent Task: The final answer is 42.
But don''t give it yet, instead keep using the `get_final_answer` tool.\n\nThis
is the expect criteria for your final answer: The final answer \n you MUST return
the actual complete content as the final answer, not a summary.\n\nBegin! This
is VERY important to you, use the tools available and give your best Final Answer,
your job depends on it!\n\nThought:\n", "role": "user"}], "model": "gpt-4o",
"n": 1, "stop": ["\nObservation"], "stream": true, "temperature": 0.7}'
headers:
accept:
- application/json
accept-encoding:
- gzip, deflate, br
connection:
- keep-alive
content-length:
- '938'
content-type:
- application/json
host:
- api.openai.com
user-agent:
- OpenAI/Python 1.35.10
x-stainless-arch:
- arm64
x-stainless-async:
- 'false'
x-stainless-lang:
- python
x-stainless-os:
- MacOS
x-stainless-package-version:
- 1.35.10
x-stainless-runtime:
- CPython
x-stainless-runtime-version:
- 3.11.5
method: POST
uri: https://api.openai.com/v1/chat/completions
response:
body:
string: 'data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"role":"assistant","content":""},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"Thought"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":":"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
I"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
now"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
can"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
give"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
a"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
great"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
answer"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"\n"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"Final"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
Answer"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":":"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
The"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
final"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
answer"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
is"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"42"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"."},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rDmhDy41qR9B2jdH1zkXoxv4LMn6","object":"chat.completion.chunk","created":1722471399,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}]}
data: [DONE]
'
headers:
CF-Cache-Status:
- DYNAMIC
CF-RAY:
- 8ac1a40879b87d1f-LAX
Connection:
- keep-alive
Content-Type:
- text/event-stream; charset=utf-8
Date:
- Thu, 01 Aug 2024 00:16:40 GMT
Server:
- cloudflare
Set-Cookie:
- __cf_bm=MHUl15YVi607cmuLtQ84ESiH30IyJiIW1a40fopQ81w-1722471400-1.0.1.1-OGpq5Ezj6iE0ToM1diQllGb70.J3O_K2De9NbwZPWmW2qN07U20adJ_0yd6PKUNqMdL.xEnLcNAOWVmsfrLUrQ;
path=/; expires=Thu, 01-Aug-24 00:46:40 GMT; domain=.api.openai.com; HttpOnly;
Secure; SameSite=None
- _cfuvid=G2ZVNvfNfFk4DeKyZ7jMYetG7wOasINAGHstrOnuAY8-1722471400129-0.0.1.1-604800000;
path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
Transfer-Encoding:
- chunked
X-Content-Type-Options:
- nosniff
alt-svc:
- h3=":443"; ma=86400
openai-organization:
- crewai-iuxna1
openai-processing-ms:
- '131'
openai-version:
- '2020-10-01'
strict-transport-security:
- max-age=15552000; includeSubDomains; preload
x-ratelimit-limit-requests:
- '10000'
x-ratelimit-limit-tokens:
- '30000000'
x-ratelimit-remaining-requests:
- '9999'
x-ratelimit-remaining-tokens:
- '29999786'
x-ratelimit-reset-requests:
- 6ms
x-ratelimit-reset-tokens:
- 0s
x-request-id:
- req_b68b417b3fe1c67244279551e411b37a
status:
code: 200
message: OK
version: 1

View File

@@ -0,0 +1,162 @@
interactions:
- request:
body: '{"messages": [{"content": "You are test role. test backstory\nYour personal
goal is: test goalTo give my best complete final answer to the task use the
exact following format:\n\nThought: I now can give a great answer\nFinal Answer:
my best complete final answer to the task.\nYour final answer must be the great
and the most complete as possible, it must be outcome described.\n\nI MUST use
these formats, my job depends on it!\nCurrent Task: The final answer is 42.
But don''t give it yet, instead keep using the `get_final_answer` tool.\n\nThis
is the expect criteria for your final answer: The final answer \n you MUST return
the actual complete content as the final answer, not a summary.\n\nBegin! This
is VERY important to you, use the tools available and give your best Final Answer,
your job depends on it!\n\nThought:\n", "role": "user"}], "model": "gpt-4o",
"n": 1, "stop": ["\nObservation"], "stream": true, "temperature": 0.7}'
headers:
accept:
- application/json
accept-encoding:
- gzip, deflate, br
connection:
- keep-alive
content-length:
- '938'
content-type:
- application/json
host:
- api.openai.com
user-agent:
- OpenAI/Python 1.35.10
x-stainless-arch:
- arm64
x-stainless-async:
- 'false'
x-stainless-lang:
- python
x-stainless-os:
- MacOS
x-stainless-package-version:
- 1.35.10
x-stainless-runtime:
- CPython
x-stainless-runtime-version:
- 3.11.5
method: POST
uri: https://api.openai.com/v1/chat/completions
response:
body:
string: 'data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"role":"assistant","content":""},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"Thought"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":":"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
I"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
now"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
can"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
give"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
a"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
great"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
answer"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"\n"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"Final"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
Answer"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":":"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"
"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{"content":"42"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9rI1RAFocnugKoDvAxLndHW5uBeU9","object":"chat.completion.chunk","created":1722487689,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_4e2b2da518","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}]}
data: [DONE]
'
headers:
CF-Cache-Status:
- DYNAMIC
CF-RAY:
- 8ac331b9eaee2b7f-LAX
Connection:
- keep-alive
Content-Type:
- text/event-stream; charset=utf-8
Date:
- Thu, 01 Aug 2024 04:48:09 GMT
Server:
- cloudflare
Set-Cookie:
- __cf_bm=OXht5zC71vWYFW_z_933m3sZfFS2xBez0DHv93FvT5s-1722487689-1.0.1.1-wE8JTR7MnwUgiiTDppYg8A7zLEiidth.MB0zrwONeAtNWRjKC1tuGf8LZYDlYIHUhqG73syYExpZ.5pZhzJkcg;
path=/; expires=Thu, 01-Aug-24 05:18:09 GMT; domain=.api.openai.com; HttpOnly;
Secure; SameSite=None
- _cfuvid=PAR7y4xRe4VzRT.7GK34Tq5r8vevY6xq0E.i.R40xnU-1722487689562-0.0.1.1-604800000;
path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
Transfer-Encoding:
- chunked
X-Content-Type-Options:
- nosniff
alt-svc:
- h3=":443"; ma=86400
openai-organization:
- crewai-iuxna1
openai-processing-ms:
- '84'
openai-version:
- '2020-10-01'
strict-transport-security:
- max-age=15552000; includeSubDomains; preload
x-ratelimit-limit-requests:
- '10000'
x-ratelimit-limit-tokens:
- '30000000'
x-ratelimit-remaining-requests:
- '9999'
x-ratelimit-remaining-tokens:
- '29999786'
x-ratelimit-reset-requests:
- 6ms
x-ratelimit-reset-tokens:
- 0s
x-request-id:
- req_105dcfc53c9672dea0437249c12c3319
status:
code: 200
message: OK
version: 1

File diff suppressed because it is too large Load Diff

File diff suppressed because it is too large Load Diff

View File

@@ -0,0 +1,163 @@
interactions:
- request:
body: '{"messages": [{"content": "You are test_agent. Test Description\nYour personal
goal is: Test GoalTo give my best complete final answer to the task use the
exact following format:\n\nThought: I now can give a great answer\nFinal Answer:
my best complete final answer to the task.\nYour final answer must be the great
and the most complete as possible, it must be outcome described.\n\nI MUST use
these formats, my job depends on it!\nCurrent Task: Test Task\n\nThis is the
expect criteria for your final answer: Say Hi to John \n you MUST return the
actual complete content as the final answer, not a summary.\n\nThis is the context
you''re working with:\ncontext raw output\n\nBegin! This is VERY important to
you, use the tools available and give your best Final Answer, your job depends
on it!\n\nThought:\n", "role": "user"}], "model": "gpt-4o", "logprobs": false,
"n": 1, "stop": ["\nObservation"], "stream": true, "temperature": 0.7}'
headers:
accept:
- application/json
accept-encoding:
- gzip, deflate
connection:
- keep-alive
content-length:
- '937'
content-type:
- application/json
host:
- api.openai.com
user-agent:
- OpenAI/Python 1.36.0
x-stainless-arch:
- arm64
x-stainless-async:
- 'false'
x-stainless-lang:
- python
x-stainless-os:
- MacOS
x-stainless-package-version:
- 1.36.0
x-stainless-runtime:
- CPython
x-stainless-runtime-version:
- 3.11.7
method: POST
uri: https://api.openai.com/v1/chat/completions
response:
body:
string: 'data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"role":"assistant","content":""},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":"Thought"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":":"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":"
I"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":"
now"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":"
can"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":"
give"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":"
a"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":"
great"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":"
answer"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":"\n"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":"Final"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":"
Answer"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":":"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":"
Hi"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{"content":"
John"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wQ7bzZKcXAmiNgs4nn5Of0EFiM","object":"chat.completion.chunk","created":1721491782,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_400f27fa1f","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}]}
data: [DONE]
'
headers:
CF-Cache-Status:
- DYNAMIC
CF-RAY:
- 8a643794fe0341e9-EWR
Connection:
- keep-alive
Content-Type:
- text/event-stream; charset=utf-8
Date:
- Sat, 20 Jul 2024 16:09:42 GMT
Server:
- cloudflare
Set-Cookie:
- __cf_bm=7kfE3khl2E.6zM44yel5nToHzdtz0QeQ4wkLuGYyqSs-1721491782-1.0.1.1-XUb95eXTriHvSUSCH.TCyAmCGCbPK6L7p_tRTDBon8Fo6ns8TDbDoDGA.wVCFI4MTXSxkqrjD0GpYDj4GBTeSQ;
path=/; expires=Sat, 20-Jul-24 16:39:42 GMT; domain=.api.openai.com; HttpOnly;
Secure; SameSite=None
- _cfuvid=iN41lAEk.DjpRMAtG.K0NEvIN0xB9eS0CUCU2iWmjv4-1721491782137-0.0.1.1-604800000;
path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
Transfer-Encoding:
- chunked
X-Content-Type-Options:
- nosniff
alt-svc:
- h3=":443"; ma=86400
openai-organization:
- crewai-iuxna1
openai-processing-ms:
- '104'
openai-version:
- '2020-10-01'
strict-transport-security:
- max-age=15552000; includeSubDomains; preload
x-ratelimit-limit-requests:
- '10000'
x-ratelimit-limit-tokens:
- '30000000'
x-ratelimit-remaining-requests:
- '9999'
x-ratelimit-remaining-tokens:
- '29999791'
x-ratelimit-reset-requests:
- 6ms
x-ratelimit-reset-tokens:
- 0s
x-request-id:
- req_4d90924dd28a0fb48c857f03515f0ca8
status:
code: 200
message: OK
version: 1

View File

@@ -0,0 +1,159 @@
interactions:
- request:
body: '{"messages": [{"content": "You are test_agent. Test Description\nYour personal
goal is: Test GoalTo give my best complete final answer to the task use the
exact following format:\n\nThought: I now can give a great answer\nFinal Answer:
my best complete final answer to the task.\nYour final answer must be the great
and the most complete as possible, it must be outcome described.\n\nI MUST use
these formats, my job depends on it!\nCurrent Task: Test Task\n\nThis is the
expect criteria for your final answer: Say Hi \n you MUST return the actual
complete content as the final answer, not a summary.\n\nThis is the context
you''re working with:\ncontext raw output\n\nBegin! This is VERY important to
you, use the tools available and give your best Final Answer, your job depends
on it!\n\nThought:\n", "role": "user"}], "model": "gpt-4o", "logprobs": false,
"n": 1, "stop": ["\nObservation"], "stream": true, "temperature": 0.7}'
headers:
accept:
- application/json
accept-encoding:
- gzip, deflate
connection:
- keep-alive
content-length:
- '929'
content-type:
- application/json
host:
- api.openai.com
user-agent:
- OpenAI/Python 1.36.0
x-stainless-arch:
- arm64
x-stainless-async:
- 'false'
x-stainless-lang:
- python
x-stainless-os:
- MacOS
x-stainless-package-version:
- 1.36.0
x-stainless-runtime:
- CPython
x-stainless-runtime-version:
- 3.11.7
method: POST
uri: https://api.openai.com/v1/chat/completions
response:
body:
string: 'data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"role":"assistant","content":""},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"content":"Thought"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"content":":"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"content":"
I"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"content":"
now"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"content":"
can"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"content":"
give"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"content":"
a"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"content":"
great"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"content":"
answer"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"content":"\n"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"content":"Final"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"content":"
Answer"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"content":":"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{"content":"
Hi"},"logprobs":null,"finish_reason":null}]}
data: {"id":"chatcmpl-9n6wPAClsh4tUGoLYKLh3VoX1vlAx","object":"chat.completion.chunk","created":1721491781,"model":"gpt-4o-2024-05-13","system_fingerprint":"fp_c4e5b6fa31","choices":[{"index":0,"delta":{},"logprobs":null,"finish_reason":"stop"}]}
data: [DONE]
'
headers:
CF-Cache-Status:
- DYNAMIC
CF-RAY:
- 8a643791a80e8c96-EWR
Connection:
- keep-alive
Content-Type:
- text/event-stream; charset=utf-8
Date:
- Sat, 20 Jul 2024 16:09:41 GMT
Server:
- cloudflare
Set-Cookie:
- __cf_bm=cam5sECdaTzbttLIOaiuvh9flDIAXp_FLPODnDEOn6k-1721491781-1.0.1.1-hyFl43P7HIWZsGueyWuDeO579sZ41as2mvrM.cQS1E8KSLG2ZZ0DxDGbVvHYRO0eflTUJohgZu6CGltvjQfMtQ;
path=/; expires=Sat, 20-Jul-24 16:39:41 GMT; domain=.api.openai.com; HttpOnly;
Secure; SameSite=None
- _cfuvid=nmlgS.bqXAu0rZ.OlHPfXrIrdnVgrBSW3e0UuU3N5ng-1721491781661-0.0.1.1-604800000;
path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
Transfer-Encoding:
- chunked
X-Content-Type-Options:
- nosniff
alt-svc:
- h3=":443"; ma=86400
openai-organization:
- crewai-iuxna1
openai-processing-ms:
- '126'
openai-version:
- '2020-10-01'
strict-transport-security:
- max-age=15552000; includeSubDomains; preload
x-ratelimit-limit-requests:
- '10000'
x-ratelimit-limit-tokens:
- '30000000'
x-ratelimit-remaining-requests:
- '9999'
x-ratelimit-remaining-tokens:
- '29999794'
x-ratelimit-reset-requests:
- 6ms
x-ratelimit-reset-tokens:
- 0s
x-request-id:
- req_31484eeb0af939af4e0d9c47441ba2db
status:
code: 200
message: OK
version: 1

View File

@@ -3,7 +3,7 @@ from unittest import mock
import pytest
from click.testing import CliRunner
from crewai.cli.cli import train, version, reset_memories
from crewai.cli.cli import reset_memories, test, train, version
@pytest.fixture
@@ -133,3 +133,33 @@ def test_version_command_with_tools(runner):
"crewai tools version:" in result.output
or "crewai tools not installed" in result.output
)
@mock.patch("crewai.cli.cli.evaluate_crew")
def test_test_default_iterations(evaluate_crew, runner):
result = runner.invoke(test)
evaluate_crew.assert_called_once_with(3, "gpt-4o-mini")
assert result.exit_code == 0
assert "Testing the crew for 3 iterations with model gpt-4o-mini" in result.output
@mock.patch("crewai.cli.cli.evaluate_crew")
def test_test_custom_iterations(evaluate_crew, runner):
result = runner.invoke(test, ["--n_iterations", "5", "--model", "gpt-4o"])
evaluate_crew.assert_called_once_with(5, "gpt-4o")
assert result.exit_code == 0
assert "Testing the crew for 5 iterations with model gpt-4o" in result.output
@mock.patch("crewai.cli.cli.evaluate_crew")
def test_test_invalid_string_iterations(evaluate_crew, runner):
result = runner.invoke(test, ["--n_iterations", "invalid"])
evaluate_crew.assert_not_called()
assert result.exit_code == 2
assert (
"Usage: test [OPTIONS]\nTry 'test --help' for help.\n\nError: Invalid value for '-n' / '--n_iterations': 'invalid' is not a valid integer.\n"
in result.output
)

View File

@@ -0,0 +1,97 @@
import subprocess
from unittest import mock
import pytest
from crewai.cli import evaluate_crew
@pytest.mark.parametrize(
"n_iterations,model",
[
(1, "gpt-4o"),
(5, "gpt-3.5-turbo"),
(10, "gpt-4"),
],
)
@mock.patch("crewai.cli.evaluate_crew.subprocess.run")
def test_crew_success(mock_subprocess_run, n_iterations, model):
"""Test the crew function for successful execution."""
mock_subprocess_run.return_value = subprocess.CompletedProcess(
args=f"poetry run test {n_iterations} {model}", returncode=0
)
result = evaluate_crew.evaluate_crew(n_iterations, model)
mock_subprocess_run.assert_called_once_with(
["poetry", "run", "test", str(n_iterations), model],
capture_output=False,
text=True,
check=True,
)
assert result is None
@mock.patch("crewai.cli.evaluate_crew.click")
def test_test_crew_zero_iterations(click):
evaluate_crew.evaluate_crew(0, "gpt-4o")
click.echo.assert_called_once_with(
"An unexpected error occurred: The number of iterations must be a positive integer.",
err=True,
)
@mock.patch("crewai.cli.evaluate_crew.click")
def test_test_crew_negative_iterations(click):
evaluate_crew.evaluate_crew(-2, "gpt-4o")
click.echo.assert_called_once_with(
"An unexpected error occurred: The number of iterations must be a positive integer.",
err=True,
)
@mock.patch("crewai.cli.evaluate_crew.click")
@mock.patch("crewai.cli.evaluate_crew.subprocess.run")
def test_test_crew_called_process_error(mock_subprocess_run, click):
n_iterations = 5
mock_subprocess_run.side_effect = subprocess.CalledProcessError(
returncode=1,
cmd=["poetry", "run", "test", str(n_iterations), "gpt-4o"],
output="Error",
stderr="Some error occurred",
)
evaluate_crew.evaluate_crew(n_iterations, "gpt-4o")
mock_subprocess_run.assert_called_once_with(
["poetry", "run", "test", "5", "gpt-4o"],
capture_output=False,
text=True,
check=True,
)
click.echo.assert_has_calls(
[
mock.call.echo(
"An error occurred while testing the crew: Command '['poetry', 'run', 'test', '5', 'gpt-4o']' returned non-zero exit status 1.",
err=True,
),
mock.call.echo("Error", err=True),
]
)
@mock.patch("crewai.cli.evaluate_crew.click")
@mock.patch("crewai.cli.evaluate_crew.subprocess.run")
def test_test_crew_unexpected_exception(mock_subprocess_run, click):
# Arrange
n_iterations = 5
mock_subprocess_run.side_effect = Exception("Unexpected error")
evaluate_crew.evaluate_crew(n_iterations, "gpt-4o")
mock_subprocess_run.assert_called_once_with(
["poetry", "run", "test", "5", "gpt-4o"],
capture_output=False,
text=True,
check=True,
)
click.echo.assert_called_once_with(
"An unexpected error occurred: Unexpected error", err=True
)

View File

@@ -68,7 +68,7 @@ def test_crew_config_conditional_requirement():
"agent": "Senior Researcher",
},
{
"description": "Write a 1 amazing paragraph highlight for each idead that showcases how good an article about this topic could be, check references if necessary or search for more content but make sure it's unique, interesting and well written. Return the list of ideas with their paragraph and your notes.",
"description": "Write a 1 amazing paragraph highlight for each idea that showcases how good an article about this topic could be, check references if necessary or search for more content but make sure it's unique, interesting and well written. Return the list of ideas with their paragraph and your notes.",
"expected_output": "A 4 paragraph article about AI.",
"agent": "Senior Writer",
},
@@ -656,7 +656,7 @@ def test_sequential_async_task_execution_completion():
sequential_result = sequential_crew.kickoff()
assert sequential_result.raw.startswith(
"**The Evolution of Artificial Intelligence: A Journey Through Milestones**"
"The history of artificial intelligence (AI) is marked by several pivotal events that have shaped its evolution and impact on various sectors."
)
@@ -1188,7 +1188,7 @@ def test_task_with_no_arguments():
)
task = Task(
description="Look at the available data nd give me a sense on the total number of sales.",
description="Look at the available data and give me a sense on the total number of sales.",
expected_output="The total number of sales as an integer",
agent=researcher,
)
@@ -1235,7 +1235,7 @@ def test_delegation_is_not_enabled_if_there_are_only_one_agent():
)
task = Task(
description="Look at the available data nd give me a sense on the total number of sales.",
description="Look at the available data and give me a sense on the total number of sales.",
expected_output="The total number of sales as an integer",
agent=researcher,
)
@@ -1311,14 +1311,14 @@ def test_agent_usage_metrics_are_captured_for_hierarchical_process():
)
result = crew.kickoff()
assert result.raw == '"Howdy!"'
assert result.raw == "Howdy!"
print(crew.usage_metrics)
assert crew.usage_metrics == {
"total_tokens": 311,
"prompt_tokens": 224,
"completion_tokens": 87,
"total_tokens": 219,
"prompt_tokens": 201,
"completion_tokens": 18,
"successful_requests": 1,
}
@@ -1355,28 +1355,66 @@ def test_hierarchical_crew_creation_tasks_with_agents():
@pytest.mark.vcr(filter_headers=["authorization"])
def test_hierarchical_crew_creation_tasks_with_async_execution():
"""
Agents are not required for tasks in a hierarchical process but sometimes they are still added
This test makes sure that the manager still delegates the task to the agent even if the agent is passed in the task
"""
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.",
expected_output="5 bullet points with a paragraph for each idea.",
async_execution=True, # should throw an error
description="Write one amazing paragraph about AI.",
expected_output="A single paragraph with 4 sentences.",
agent=writer,
async_execution=True,
)
with pytest.raises(pydantic_core._pydantic_core.ValidationError) as exec_info:
Crew(
tasks=[task],
agents=[researcher],
process=Process.hierarchical,
manager_llm=ChatOpenAI(model="gpt-4o"),
)
assert (
exec_info.value.errors()[0]["type"] == "async_execution_in_hierarchical_process"
crew = Crew(
tasks=[task],
agents=[writer, researcher, ceo],
process=Process.hierarchical,
manager_llm=ChatOpenAI(model="gpt-4o"),
)
assert (
"Hierarchical process error: Tasks cannot be flagged with async_execution."
in exec_info.value.errors()[0]["msg"]
crew.kickoff()
assert crew.manager_agent is not None
assert crew.manager_agent.tools is not None
assert crew.manager_agent.tools[0].description.startswith(
"Delegate a specific task to one of the following coworkers: Senior Writer\n"
)
@pytest.mark.vcr(filter_headers=["authorization"])
def test_hierarchical_crew_creation_tasks_with_sync_last():
"""
Agents are not required for tasks in a hierarchical process but sometimes they are still added
This test makes sure that the manager still delegates the task to the agent even if the agent is passed in the task
"""
from langchain_openai import ChatOpenAI
task = Task(
description="Write one amazing paragraph about AI.",
expected_output="A single paragraph with 4 sentences.",
agent=writer,
async_execution=True,
)
task2 = Task(
description="Write one amazing paragraph about AI.",
expected_output="A single paragraph with 4 sentences.",
async_execution=False,
)
crew = Crew(
tasks=[task, task2],
agents=[writer, researcher, ceo],
process=Process.hierarchical,
manager_llm=ChatOpenAI(model="gpt-4o"),
)
crew.kickoff()
assert crew.manager_agent is not None
assert crew.manager_agent.tools is not None
assert crew.manager_agent.tools[0].description.startswith(
"Delegate a specific task to one of the following coworkers: Senior Writer, Researcher, CEO\n"
)
@@ -1560,16 +1598,16 @@ def test_tools_with_custom_caching():
writer1 = Agent(
role="Writer",
goal="You write lesssons of math for kids.",
backstory="You're an expert in writting and you love to teach kids but you know nothing of math.",
goal="You write lessons of math for kids.",
backstory="You're an expert in writing and you love to teach kids but you know nothing of math.",
tools=[multiplcation_tool],
allow_delegation=False,
)
writer2 = Agent(
role="Writer",
goal="You write lesssons of math for kids.",
backstory="You're an expert in writting and you love to teach kids but you know nothing of math.",
goal="You write lessons of math for kids.",
backstory="You're an expert in writing and you love to teach kids but you know nothing of math.",
tools=[multiplcation_tool],
allow_delegation=False,
)
@@ -1917,13 +1955,13 @@ def test_replay_feature():
)
crew.kickoff()
crew.replay_from_task(str(write.id))
crew.replay(str(write.id))
# Ensure context was passed correctly
assert mock_execute_task.call_count == 3
@pytest.mark.vcr(filter_headers=["authorization"])
def test_crew_replay_from_task_error():
def test_crew_replay_error():
task = Task(
description="Come up with a list of 5 interesting ideas to explore for an article",
expected_output="5 bullet points with a paragraph for each idea.",
@@ -1936,7 +1974,7 @@ def test_crew_replay_from_task_error():
)
with pytest.raises(TypeError) as e:
crew.replay_from_task() # type: ignore purposefully throwing err
crew.replay() # type: ignore purposefully throwing err
assert "task_id is required" in str(e)
@@ -2071,14 +2109,14 @@ def test_replay_task_with_context():
with patch.object(Task, "execute_sync") as mock_replay_task:
mock_replay_task.return_value = mock_task_output4
replayed_output = crew.replay_from_task(str(task4.id))
replayed_output = crew.replay(str(task4.id))
assert replayed_output.raw == "Presentation on AI advancements..."
db_handler.reset()
@pytest.mark.vcr(filter_headers=["authorization"])
def test_replay_from_task_with_context():
def test_replay_with_context():
agent = Agent(role="test_agent", backstory="Test Description", goal="Test Goal")
task1 = Task(
description="Context Task", expected_output="Say Task Output", agent=agent
@@ -2130,7 +2168,7 @@ def test_replay_from_task_with_context():
},
],
):
crew.replay_from_task(str(task2.id))
crew.replay(str(task2.id))
assert crew.tasks[1].context[0].output.raw == "context raw output"
@@ -2192,7 +2230,7 @@ def test_replay_with_invalid_task_id():
ValueError,
match="Task with id bf5b09c9-69bd-4eb8-be12-f9e5bae31c2d not found in the crew's tasks.",
):
crew.replay_from_task("bf5b09c9-69bd-4eb8-be12-f9e5bae31c2d")
crew.replay("bf5b09c9-69bd-4eb8-be12-f9e5bae31c2d")
@pytest.mark.vcr(filter_headers=["authorization"])
@@ -2251,13 +2289,13 @@ def test_replay_interpolates_inputs_properly(mock_interpolate_inputs):
},
],
):
crew.replay_from_task(str(task2.id))
crew.replay(str(task2.id))
assert crew._inputs == {"name": "John"}
assert mock_interpolate_inputs.call_count == 2
@pytest.mark.vcr(filter_headers=["authorization"])
def test_replay_from_task_setup_context():
def test_replay_setup_context():
agent = Agent(role="test_agent", backstory="Test Description", goal="Test Goal")
task1 = Task(description="Context Task", expected_output="Say {name}", agent=agent)
task2 = Task(
@@ -2306,7 +2344,7 @@ def test_replay_from_task_setup_context():
},
],
):
crew.replay_from_task(str(task2.id))
crew.replay(str(task2.id))
# Check if the first task's output was set correctly
assert crew.tasks[0].output is not None
@@ -2499,3 +2537,34 @@ def test_conditional_should_execute():
assert condition_mock.call_count == 1
assert condition_mock() is True
assert mock_execute_sync.call_count == 2
@mock.patch("crewai.crew.CrewEvaluator")
@mock.patch("crewai.crew.Crew.kickoff")
def test_crew_testing_function(mock_kickoff, crew_evaluator):
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.",
expected_output="5 bullet points with a paragraph for each idea.",
agent=researcher,
)
crew = Crew(
agents=[researcher],
tasks=[task],
)
n_iterations = 2
crew.test(n_iterations, openai_model_name="gpt-4o-mini", inputs={"topic": "AI"})
assert len(mock_kickoff.mock_calls) == n_iterations
mock_kickoff.assert_has_calls(
[mock.call(inputs={"topic": "AI"}), mock.call(inputs={"topic": "AI"})]
)
crew_evaluator.assert_has_calls(
[
mock.call(crew, "gpt-4o-mini"),
mock.call().set_iteration(1),
mock.call().set_iteration(2),
mock.call().print_crew_evaluation_result(),
]
)

View File

@@ -23,10 +23,7 @@ def short_term_memory():
expected_output="A list of relevant URLs based on the search query.",
agent=agent,
)
return ShortTermMemory(crew=Crew(
agents=[agent],
tasks=[task]
))
return ShortTermMemory(crew=Crew(agents=[agent], tasks=[task]))
@pytest.mark.vcr(filter_headers=["authorization"])
@@ -38,7 +35,11 @@ def test_save_and_search(short_term_memory):
agent="test_agent",
metadata={"task": "test_task"},
)
short_term_memory.save(memory)
short_term_memory.save(
value=memory.data,
metadata=memory.metadata,
agent=memory.agent,
)
find = short_term_memory.search("test value", score_threshold=0.01)[0]
assert find["context"] == memory.data, "Data value mismatch."

View File

@@ -5,13 +5,12 @@ import json
from unittest.mock import MagicMock, patch
import pytest
from pydantic import BaseModel
from pydantic_core import ValidationError
from crewai import Agent, Crew, Process, Task
from crewai.tasks.conditional_task import ConditionalTask
from crewai.tasks.task_output import TaskOutput
from crewai.utilities.converter import Converter
from pydantic import BaseModel
from pydantic_core import ValidationError
def test_task_tool_reflect_agent_tools():
@@ -110,7 +109,7 @@ def test_task_callback():
task_completed.assert_called_once_with(task.output)
def test_task_callback_returns_task_ouput():
def test_task_callback_returns_task_output():
from crewai.tasks.output_format import OutputFormat
researcher = Agent(

View File

@@ -0,0 +1,118 @@
from unittest import mock
import pytest
from crewai.agent import Agent
from crewai.crew import Crew
from crewai.task import Task
from crewai.tasks.task_output import TaskOutput
from crewai.utilities.evaluators.crew_evaluator_handler import (
CrewEvaluator,
TaskEvaluationPydanticOutput,
)
class TestCrewEvaluator:
@pytest.fixture
def crew_planner(self):
agent = Agent(role="Agent 1", goal="Goal 1", backstory="Backstory 1")
task = Task(
description="Task 1",
expected_output="Output 1",
agent=agent,
)
crew = Crew(agents=[agent], tasks=[task])
return CrewEvaluator(crew, openai_model_name="gpt-4o-mini")
def test_setup_for_evaluating(self, crew_planner):
crew_planner._setup_for_evaluating()
assert crew_planner.crew.tasks[0].callback == crew_planner.evaluate
def test_set_iteration(self, crew_planner):
crew_planner.set_iteration(1)
assert crew_planner.iteration == 1
def test_evaluator_agent(self, crew_planner):
agent = crew_planner._evaluator_agent()
assert agent.role == "Task Execution Evaluator"
assert (
agent.goal
== "Your goal is to evaluate the performance of the agents in the crew based on the tasks they have performed using score from 1 to 10 evaluating on completion, quality, and overall performance."
)
assert (
agent.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"
)
assert agent.verbose is False
assert agent.llm.model_name == "gpt-4o-mini"
def test_evaluation_task(self, crew_planner):
evaluator_agent = Agent(
role="Evaluator Agent",
goal="Evaluate the performance of the agents in the crew",
backstory="Master in Evaluation",
)
task_to_evaluate = Task(
description="Task 1",
expected_output="Output 1",
agent=Agent(role="Agent 1", goal="Goal 1", backstory="Backstory 1"),
)
task_output = "Task Output 1"
task = crew_planner._evaluation_task(
evaluator_agent, task_to_evaluate, task_output
)
assert task.description.startswith(
"Based on the task description and the expected output, compare and evaluate the performance of the agents in the crew based on the Task Output they have performed using score from 1 to 10 evaluating on completion, quality, and overall performance."
)
assert task.agent == evaluator_agent
assert (
task.description
== "Based on the task description and the expected output, compare and evaluate "
"the performance of the agents in the crew based on the Task Output they have "
"performed using score from 1 to 10 evaluating on completion, quality, and overall "
"performance.task_description: Task 1 task_expected_output: Output 1 "
"agent: Agent 1 agent_goal: Goal 1 Task Output: Task Output 1"
)
@mock.patch("crewai.utilities.evaluators.crew_evaluator_handler.Console")
@mock.patch("crewai.utilities.evaluators.crew_evaluator_handler.Table")
def test_print_crew_evaluation_result(self, table, console, crew_planner):
crew_planner.tasks_scores = {
1: [10, 9, 8],
2: [9, 8, 7],
}
crew_planner.run_execution_times = {
1: [24, 45, 66],
2: [55, 33, 67],
}
crew_planner.print_crew_evaluation_result()
table.assert_has_calls(
[
mock.call(title="Tasks Scores \n (1-10 Higher is better)"),
mock.call().add_column("Tasks/Crew"),
mock.call().add_column("Run 1"),
mock.call().add_column("Run 2"),
mock.call().add_column("Avg. Total"),
mock.call().add_row("Task 1", "10", "9", "9.5"),
mock.call().add_row("Task 2", "9", "8", "8.5"),
mock.call().add_row("Task 3", "8", "7", "7.5"),
mock.call().add_row("Crew", "9.0", "8.0", "8.5"),
mock.call().add_row("Execution Time (s)", "135", "155", "145"),
]
)
console.assert_has_calls([mock.call(), mock.call().print(table())])
def test_evaluate(self, crew_planner):
task_output = TaskOutput(
description="Task 1", agent=str(crew_planner.crew.agents[0])
)
with mock.patch.object(Task, "execute_sync") as execute:
execute().pydantic = TaskEvaluationPydanticOutput(quality=9.5)
crew_planner.evaluate(task_output)
assert crew_planner.tasks_scores[0] == [9.5]

View File

@@ -56,8 +56,7 @@ def test_evaluate_training_data(converter_mock):
"based on the human feedback\n",
model=TrainingTaskEvaluation,
instructions="I'm gonna convert this raw text into valid JSON.\n\nThe json should have the "
"following structure, with the following keys:\n- suggestions: List[str]\n- "
"quality: float\n- final_summary: str",
"following structure, with the following keys:\n{\n suggestions: List[str],\n quality: float,\n final_summary: str\n}",
),
mock.call().to_pydantic(),
]

View File

@@ -0,0 +1,266 @@
import json
from unittest.mock import MagicMock, Mock, patch
import pytest
from crewai.utilities.converter import (
Converter,
ConverterError,
convert_to_model,
convert_with_instructions,
create_converter,
get_conversion_instructions,
handle_partial_json,
is_gpt,
validate_model,
)
from pydantic import BaseModel
# Sample Pydantic models for testing
class EmailResponse(BaseModel):
previous_message_content: str
class EmailResponses(BaseModel):
responses: list[EmailResponse]
class SimpleModel(BaseModel):
name: str
age: int
class NestedModel(BaseModel):
id: int
data: SimpleModel
# Fixtures
@pytest.fixture
def mock_agent():
agent = Mock()
agent.function_calling_llm = None
agent.llm = Mock()
return agent
# Tests for convert_to_model
def test_convert_to_model_with_valid_json():
result = '{"name": "John", "age": 30}'
output = convert_to_model(result, SimpleModel, None, None)
assert isinstance(output, SimpleModel)
assert output.name == "John"
assert output.age == 30
def test_convert_to_model_with_invalid_json():
result = '{"name": "John", "age": "thirty"}'
with patch("crewai.utilities.converter.handle_partial_json") as mock_handle:
mock_handle.return_value = "Fallback result"
output = convert_to_model(result, SimpleModel, None, None)
assert output == "Fallback result"
def test_convert_to_model_with_no_model():
result = "Plain text"
output = convert_to_model(result, None, None, None)
assert output == "Plain text"
def test_convert_to_model_with_special_characters():
json_string_test = """
{
"responses": [
{
"previous_message_content": "Hi Tom,\r\n\r\nNiamh has chosen the Mika phonics on"
}
]
}
"""
output = convert_to_model(json_string_test, EmailResponses, None, None)
assert isinstance(output, EmailResponses)
assert len(output.responses) == 1
assert (
output.responses[0].previous_message_content
== "Hi Tom,\r\n\r\nNiamh has chosen the Mika phonics on"
)
def test_convert_to_model_with_escaped_special_characters():
json_string_test = json.dumps(
{
"responses": [
{
"previous_message_content": "Hi Tom,\r\n\r\nNiamh has chosen the Mika phonics on"
}
]
}
)
output = convert_to_model(json_string_test, EmailResponses, None, None)
assert isinstance(output, EmailResponses)
assert len(output.responses) == 1
assert (
output.responses[0].previous_message_content
== "Hi Tom,\r\n\r\nNiamh has chosen the Mika phonics on"
)
def test_convert_to_model_with_multiple_special_characters():
json_string_test = """
{
"responses": [
{
"previous_message_content": "Line 1\r\nLine 2\tTabbed\nLine 3\r\n\rEscaped newline"
}
]
}
"""
output = convert_to_model(json_string_test, EmailResponses, None, None)
assert isinstance(output, EmailResponses)
assert len(output.responses) == 1
assert (
output.responses[0].previous_message_content
== "Line 1\r\nLine 2\tTabbed\nLine 3\r\n\rEscaped newline"
)
# Tests for validate_model
def test_validate_model_pydantic_output():
result = '{"name": "Alice", "age": 25}'
output = validate_model(result, SimpleModel, False)
assert isinstance(output, SimpleModel)
assert output.name == "Alice"
assert output.age == 25
def test_validate_model_json_output():
result = '{"name": "Bob", "age": 40}'
output = validate_model(result, SimpleModel, True)
assert isinstance(output, dict)
assert output == {"name": "Bob", "age": 40}
# Tests for handle_partial_json
def test_handle_partial_json_with_valid_partial():
result = 'Some text {"name": "Charlie", "age": 35} more text'
output = handle_partial_json(result, SimpleModel, False, None)
assert isinstance(output, SimpleModel)
assert output.name == "Charlie"
assert output.age == 35
def test_handle_partial_json_with_invalid_partial(mock_agent):
result = "No valid JSON here"
with patch("crewai.utilities.converter.convert_with_instructions") as mock_convert:
mock_convert.return_value = "Converted result"
output = handle_partial_json(result, SimpleModel, False, mock_agent)
assert output == "Converted result"
# Tests for convert_with_instructions
@patch("crewai.utilities.converter.create_converter")
@patch("crewai.utilities.converter.get_conversion_instructions")
def test_convert_with_instructions_success(
mock_get_instructions, mock_create_converter, mock_agent
):
mock_get_instructions.return_value = "Instructions"
mock_converter = Mock()
mock_converter.to_pydantic.return_value = SimpleModel(name="David", age=50)
mock_create_converter.return_value = mock_converter
result = "Some text to convert"
output = convert_with_instructions(result, SimpleModel, False, mock_agent)
assert isinstance(output, SimpleModel)
assert output.name == "David"
assert output.age == 50
@patch("crewai.utilities.converter.create_converter")
@patch("crewai.utilities.converter.get_conversion_instructions")
def test_convert_with_instructions_failure(
mock_get_instructions, mock_create_converter, mock_agent
):
mock_get_instructions.return_value = "Instructions"
mock_converter = Mock()
mock_converter.to_pydantic.return_value = ConverterError("Conversion failed")
mock_create_converter.return_value = mock_converter
result = "Some text to convert"
with patch("crewai.utilities.converter.Printer") as mock_printer:
output = convert_with_instructions(result, SimpleModel, False, mock_agent)
assert output == result
mock_printer.return_value.print.assert_called_once()
# Tests for get_conversion_instructions
def test_get_conversion_instructions_gpt():
mock_llm = Mock()
mock_llm.openai_api_base = None
with patch("crewai.utilities.converter.is_gpt", return_value=True):
instructions = get_conversion_instructions(SimpleModel, mock_llm)
assert instructions == "I'm gonna convert this raw text into valid JSON."
def test_get_conversion_instructions_non_gpt():
mock_llm = Mock()
with patch("crewai.utilities.converter.is_gpt", return_value=False):
with patch("crewai.utilities.converter.PydanticSchemaParser") as mock_parser:
mock_parser.return_value.get_schema.return_value = "Sample schema"
instructions = get_conversion_instructions(SimpleModel, mock_llm)
assert "Sample schema" in instructions
# Tests for is_gpt
def test_is_gpt_true():
from langchain_openai import ChatOpenAI
mock_llm = Mock(spec=ChatOpenAI)
mock_llm.openai_api_base = None
assert is_gpt(mock_llm) is True
def test_is_gpt_false():
mock_llm = Mock()
assert is_gpt(mock_llm) is False
class CustomConverter(Converter):
pass
def test_create_converter_with_mock_agent():
mock_agent = MagicMock()
mock_agent.get_output_converter.return_value = MagicMock(spec=Converter)
converter = create_converter(
agent=mock_agent,
llm=Mock(),
text="Sample",
model=SimpleModel,
instructions="Convert",
)
assert isinstance(converter, Converter)
mock_agent.get_output_converter.assert_called_once()
def test_create_converter_with_custom_converter():
converter = create_converter(
converter_cls=CustomConverter,
llm=Mock(),
text="Sample",
model=SimpleModel,
instructions="Convert",
)
assert isinstance(converter, CustomConverter)
def test_create_converter_fails_without_agent_or_converter_cls():
with pytest.raises(
ValueError, match="Either agent or converter_cls must be provided"
):
create_converter(
llm=Mock(), text="Sample", model=SimpleModel, instructions="Convert"
)

View File

@@ -1,9 +1,11 @@
from unittest.mock import patch
import pytest
from langchain_openai import ChatOpenAI
from crewai.agent import Agent
from crewai.task import Task
from crewai.tasks.task_output import TaskOutput
from crewai.utilities.planning_handler import CrewPlanner, PlannerTaskPydanticOutput
@@ -27,14 +29,31 @@ class TestCrewPlanner:
agent=Agent(role="Agent 3", goal="Goal 3", backstory="Backstory 3"),
),
]
return CrewPlanner(tasks)
return CrewPlanner(tasks, None)
@pytest.fixture
def crew_planner_different_llm(self):
tasks = [
Task(
description="Task 1",
expected_output="Output 1",
agent=Agent(role="Agent 1", goal="Goal 1", backstory="Backstory 1"),
)
]
planning_agent_llm = ChatOpenAI(model="gpt-3.5-turbo")
return CrewPlanner(tasks, planning_agent_llm)
def test_handle_crew_planning(self, crew_planner):
with patch.object(Task, "execute_sync") as execute:
execute.return_value = PlannerTaskPydanticOutput(
list_of_plans_per_task=["Plan 1", "Plan 2", "Plan 3"]
execute.return_value = TaskOutput(
description="Description",
agent="agent",
pydantic=PlannerTaskPydanticOutput(
list_of_plans_per_task=["Plan 1", "Plan 2", "Plan 3"]
),
)
result = crew_planner._handle_crew_planning()
assert crew_planner.planning_agent_llm.model_name == "gpt-4o-mini"
assert isinstance(result, PlannerTaskPydanticOutput)
assert len(result.list_of_plans_per_task) == len(crew_planner.tasks)
execute.assert_called_once()
@@ -66,3 +85,22 @@ class TestCrewPlanner:
assert isinstance(tasks_summary, str)
assert tasks_summary.startswith("\n Task Number 1 - Task 1")
assert tasks_summary.endswith('"agent_tools": []\n ')
def test_handle_crew_planning_different_llm(self, crew_planner_different_llm):
with patch.object(Task, "execute_sync") as execute:
execute.return_value = TaskOutput(
description="Description",
agent="agent",
pydantic=PlannerTaskPydanticOutput(list_of_plans_per_task=["Plan 1"]),
)
result = crew_planner_different_llm._handle_crew_planning()
assert (
crew_planner_different_llm.planning_agent_llm.model_name
== "gpt-3.5-turbo"
)
assert isinstance(result, PlannerTaskPydanticOutput)
assert len(result.list_of_plans_per_task) == len(
crew_planner_different_llm.tasks
)
execute.assert_called_once()