Autonoumous multiagent frame system able to execute code in local computer with memory and tool & function usage

Using https://github.com/frdel/agent-zero?tab=readme-ov-file

PRE WORK after you have retrieved code from github:

• optionally install ollama to your local computer and start it in wsl , this enables you to be able to use local ollama models with this framework

• set your .env file (set apicodes you want to use (if no API codes leave empty (in case you using only Ollama )

API_KEY_OPENAI=YOURAPICODEHERE
API_KEY_ANTHROPIC=YOURAPICODEHERE
API_KEY_GROQ=YOURAPICODEHERE
API_KEY_PERPLEXITY=YOURAPICODEHERE
API_KEY_GOOGLE=YOURAPICODEHERE

TOKENIZERS_PARALLELISM=true
PYDEVD_DISABLE_FILE_VALIDATION=1         

• select your model used by uncommenting/ commenting following part (in main.py)

def initialize():
    
    # main chat model used by agents (smarter, more accurate)

    # chat_llm = models.get_groq_llama70b(temperature=0.2)
    # chat_llm = models.get_groq_llama70b_json(temperature=0.2)
    # chat_llm = models.get_groq_llama8b(temperature=0.2)
    # chat_llm = models.get_openai_gpt35(temperature=0)
    # chat_llm = models.get_openai_gpt4o(temperature=0)
    chat_llm = models.get_openai_chat(temperature=0)
    # chat_llm = models.get_anthropic_opus(temperature=0)
    # chat_llm = models.get_anthropic_sonnet(temperature=0)
    # chat_llm = models.get_anthropic_sonnet_35(temperature=0)
    # chat_llm = models.get_anthropic_haiku(temperature=0)
    # chat_llm = models.get_ollama_dolphin()
    # chat_llm = models.get_ollama(model_name="gemma2:27b")
    # chat_llm = models.get_ollama(model_name="llama3:8b-text-fp16")
    # chat_llm = models.get_ollama(model_name="gemma2:latest")
    # chat_llm = models.get_ollama(model_name="qwen:14b")
    # chat_llm = models.get_google_chat()
        

• Check you have python3 and docker and venv in use

Here’s a script designed to be run on a Linux (Ubuntu) system. This script will create a virtual environment, activate it, install the required Python packages from requirements.txt, and start Docker. I use wsl -d Ubuntu in my windows 11

Script: setup_and_start_docker.sh

#!/bin/bash

# Function to check if a command exists
command_exists() {
    command -v "$1" &> /dev/null
}

# Check if Python3 is installed
if ! command_exists python3; then
    echo "Python3 is not installed. Please install Python3 and try again."
    exit 1
fi

# Check if pip is installed
if ! command_exists pip; then
    echo "pip is not installed. Please install pip and try again."
    exit 1
fi

# Check if Docker is installed
if ! command_exists docker; then
    echo "Docker is not installed. Please install Docker and try again."
    exit 1
fi

# Create Virtual Environment
echo "Creating virtual environment..."
python3 -m venv venv

# Activate Virtual Environment and Install Python Packages
echo "Activating virtual environment and installing packages..."
source venv/bin/activate

if [ -f requirements.txt ]; then
    pip install -r requirements.txt
else
    echo "requirements.txt not found"
fi

# Start Docker
echo "Starting Docker..."
sudo service docker start

# Check Docker Status
echo "Checking Docker status..."
sudo service docker status

echo "Setup complete."
        

Instructions to Use the Script

1. Save the Script: Save the script to a file, e.g., setup_and_start_docker.sh.
2. Make the Script Executable: Make the script executable by running:
3. Run the Script: Execute the script:

Important Notes

• sudo: The script uses sudo to start Docker, so you may be prompted to enter your password.
• requirements.txt: Make sure the requirements.txt file is in the same directory where you are running the script. Alternatively, you can modify the script to specify a different path.

Troubleshooting

If you encounter any issues:

• Ensure that Python3, pip, and Docker are installed and properly configured.
• Verify that you have the necessary permissions to run Docker and install Python packages.
• Check the permissions and ensure the script has executable rights.

This script should help automate the process of setting up the Python environment and starting Docker within an Ubuntu system.

Executing framework

# Navigate to the project folder
cd /home/user/myproject
# Activate the virtual environment
source venv/bin/activate
# Ensure Docker is running
sudo service docker start
# Run the main.py script using python3
python3 main.py        

Important Notes

• Virtual Environment Activation: Ensure that the virtual environment is activated each time you open a new terminal session before running main.py.
• Docker: If Docker is not set to start automatically with your system, you will need to start it manually each time before running your script.
• Permissions: If you encounter permission issues with Docker, ensure that your user is added to the Docker group, or use sudo to run Docker-related commands.

By following these instructions, you should be able to navigate to your project folder and start main.py with the necessary environment and services running using python3.

Example how code is created and executed: in folder ./work_dir

I start to discuss with system and at some point paused... at this point the system creates "memoryes" of my actions (I already asked it to tellme current weather in Finland Helsinki and Sweden Stockholm, it failed since it did not have API code for the service it selected to use by its own, requesting the user to provide it)

Agent 0: Mid messages cleanup summary:

{

    "system_info": "Messages have been summarized to save space.",

    "messages_summary": [

        "User confirmed the provided time method works for both Helsinki and Madrid.",

        "User requested to compare the current weather in Helsinki and Stockholm.",

        "AI suggested using various weather APIs, including OpenWeatherMap, Meteosource, Weatherbit, Visual Crossing, and Open-Meteo.",

        "AI prepared to use the OpenWeatherMap API to retrieve weather data but required a valid API key.",

        "AI noted it could not find a valid OpenWeatherMap API key in memory and requested the user to provide one."

    ]

}        

I give my openweather api key to system (I will delete this key , dont you worry):

User message ('e' to leave):
> openweather API key = 918403ad1a1997bf8650f34d05c9c9a4        

And then magic starts :) see especially when agent 1 makes and executes code in local computer docker container :) asked by Agent0 , and Agent0 delivering only the final answer

Agent 0: Starting a message:

{

    "thoughts": [

        "The user has provided a valid OpenWeatherMap API key.",

        "I will now proceed to retrieve the current weather data for both Helsinki and Stockholm using this API key."

    ],

    "tool_name": "call_subordinate",

    "tool_args": {

        "message": "Now that I have the API key, please retrieve the current weather data for Helsinki and Stockholm using the following API calls:\n1. For Helsinki: https://api.openweathermap.org/data/2.5/weather?lat=60.1699&lon=24.9384&appid=918403ad1a1997bf8650f34d05c9c9a4\n2. For Stockholm: https://api.openweathermap.org/data/2.5/weather?lat=59.3293&lon=18.0686&appid=918403ad1a1997bf8650f34d05c9c9a4.",

        "reset": "false"

    }

}

Agent 0: Using tool 'call_subordinate':

Message:

Now that I have the API key, please retrieve the current weather data for Helsinki and Stockholm using the following API calls:

1. For Helsinki: https://api.openweathermap.org/data/2.5/weather?lat=60.1699&lon=24.9384&appid=918403ad1a1997bf8650f34d05c9c9a4

2. For Stockholm: https://api.openweathermap.org/data/2.5/weather?lat=59.3293&lon=18.0686&appid=918403ad1a1997bf8650f34d05c9c9a4.

Reset: false
        

Agent 1: Starting a message:

{

    "thoughts": [

        "The user has provided a valid API key for the OpenWeatherMap API.",

        "I will prepare the API calls for both Helsinki and Stockholm using the provided API key.",

        "Next, I will execute the API calls to retrieve the current weather data."

    ],

    "tool_name": "code_execution_tool",

    "tool_args": {

        "runtime": "python",

        "code": "import requests\n\n# API key\napi_key = '918403ad1a1997bf8650f34d05c9c9a4'\n\n# URLs for the API calls\nhelsinki_url = f'https://api.openweathermap.org/data/2.5/weather?lat=60.1699&lon=24.9384&appid={api_key}'\nstockholm_url = f'https://api.openweathermap.org/data/2.5/weather?lat=59.3293&lon=18.0686&appid={api_key}'\n\n# Making the API calls\nhelsinki_response = requests.get(helsinki_url)\nstockholm_response = requests.get(stockholm_url)\n\n# Getting the JSON data\nhelsinki_data = helsinki_response.json()\nstockholm_data = stockholm_response.json()\n\n# Printing the results\nprint('Helsinki Weather:', helsinki_data)\nprint('Stockholm Weather:', stockholm_data)"

    }

}

Agent 1: Using tool 'code_execution_tool':

Runtime: python

Code:

import requests

# API key

api_key = '918403ad1a1997bf8650f34d05c9c9a4'

# URLs for the API calls

helsinki_url = f'https://api.openweathermap.org/data/2.5/weather?lat=60.1699&lon=24.9384&appid={api_key}'

stockholm_url = f'https://api.openweathermap.org/data/2.5/weather?lat=59.3293&lon=18.0686&appid={api_key}'

# Making the API calls

helsinki_response = requests.get(helsinki_url)

stockholm_response = requests.get(stockholm_url)

# Getting the JSON data

helsinki_data = helsinki_response.json()

stockholm_data = stockholm_response.json()

# Printing the results

print('Helsinki Weather:', helsinki_data)

print('Stockholm Weather:', stockholm_data)

Agent 1 code execution output:

Traceback (most recent call last):

  File "<string>", line 1, in <module>

ModuleNotFoundError: No module named 'requests'

(venv) root@2c7cada58dbe:~#
        

above the agent 1 counters an error and starts to solve it below:

Agent 1: Starting a message:

{

    "thoughts": [

        "The execution failed because the 'requests' module is not installed in the environment.",

        "I need to install the 'requests' library before I can make the API calls.",

        "I will proceed to install the 'requests' library and then re-execute the API calls."

    ],

    "tool_name": "code_execution_tool",

    "tool_args": {

        "runtime": "terminal",

        "code": "pip install requests"

    }

}

Agent 1: Using tool 'code_execution_tool':

Runtime: terminal

Code: pip install requests

Agent 1 code execution output:

Collecting requests

  Downloading requests-2.32.3-py3-none-any.whl.metadata (4.6 kB)

Collecting charset-normalizer<4,>=2 (from requests)

  Downloading charset_normalizer-3.3.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (33 kB)

Collecting idna<4,>=2.5 (from requests)

  Downloading idna-3.7-py3-none-any.whl.metadata (9.9 kB)

Collecting urllib3<3,>=1.21.1 (from requests)

  Downloading urllib3-2.2.2-py3-none-any.whl.metadata (6.4 kB)

Collecting certifi>=2017.4.17 (from requests)

  Downloading certifi-2024.7.4-py3-none-any.whl.metadata (2.2 kB)

Downloading requests-2.32.3-py3-none-any.whl (64 kB)

   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 64.9/64.9 kB 2.9 MB/s eta 0:00:00

Downloading certifi-2024.7.4-py3-none-any.whl (162 kB)

   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 163.0/163.0 kB 5.4 MB/s eta 0:00:00

Downloading charset_normalizer-3.3.2-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (140 kB)

   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 140.3/140.3 kB 5.3 MB/s eta 0:00:00

Downloading idna-3.7-py3-none-any.whl (66 kB)

   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 66.8/66.8 kB 5.3 MB/s eta 0:00:00

Downloading urllib3-2.2.2-py3-none-any.whl (121 kB)

   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 121.4/121.4 kB 3.5 MB/s eta 0:00:00

Installing collected packages: urllib3, idna, charset-normalizer, certifi, requests

Successfully installed certifi-2024.7.4 charset-normalizer-3.3.2 idna-3.7 requests-2.32.3 urllib3-2.2.2

[notice] A new release of pip is available: 24.1.2 -> 24.2

[notice] To update, run: pip install --upgrade pip

(venv) root@2c7cada58dbe:~#

        

Problem solved Agent 1 proceees on task

Agent 1: Starting a message:

{

    "thoughts": [

        "The 'requests' library has been successfully installed.",

        "Now I will re-execute the API calls to retrieve the current weather data for Helsinki and Stockholm."

    ],

    "tool_name": "code_execution_tool",

    "tool_args": {

        "runtime": "python",

        "code": "import requests\n\n# API key\napi_key = '918403ad1a1997bf8650f34d05c9c9a4'\n\n# URLs for the API calls\nhelsinki_url = f'https://api.openweathermap.org/data/2.5/weather?lat=60.1699&lon=24.9384&appid={api_key}'\nstockholm_url = f'https://api.openweathermap.org/data/2.5/weather?lat=59.3293&lon=18.0686&appid={api_key}'\n\n# Making the API calls\nhelsinki_response = requests.get(helsinki_url)\nstockholm_response = requests.get(stockholm_url)\n\n# Getting the JSON data\nhelsinki_data = helsinki_response.json()\nstockholm_data = stockholm_response.json()\n\n# Printing the results\nprint('Helsinki Weather:', helsinki_data)\nprint('Stockholm Weather:', stockholm_data)"

    }

}


Agent 1: Using tool 'code_execution_tool':

Runtime: python

Code:

import requests

# API key

api_key = '918403ad1a1997bf8650f34d05c9c9a4'

# URLs for the API calls

helsinki_url = f'https://api.openweathermap.org/data/2.5/weather?lat=60.1699&lon=24.9384&appid={api_key}'

stockholm_url = f'https://api.openweathermap.org/data/2.5/weather?lat=59.3293&lon=18.0686&appid={api_key}'

# Making the API calls

helsinki_response = requests.get(helsinki_url)

stockholm_response = requests.get(stockholm_url)

# Getting the JSON data

helsinki_data = helsinki_response.json()

stockholm_data = stockholm_response.json()

# Printing the results

print('Helsinki Weather:', helsinki_data)

print('Stockholm Weather:', stockholm_data)

Agent 1 code execution output:

Helsinki Weather: {'coord': {'lon': 24.9384, 'lat': 60.1699}, 'weather': [{'id': 803, 'main': 'Clouds', 'description': 'broken clouds', 'icon': '04d'}], 'base': 'stations', 'main': {'temp': 289.96, 'feels_like': 289.97, 'temp_min': 288.76, 'temp_max': 291.18, 'pressure': 1005, 'humidity': 87, 'sea_level': 1005, 'grnd_level': 1002}, 'visibility': 10000, 'wind': {'speed': 5.36, 'deg': 317, 'gust': 6.26}, 'clouds': {'all': 75}, 'dt': 1722581370, 'sys': {'type': 2, 'id': 2011913, 'country': 'FI', 'sunrise': 1722564175, 'sunset': 1722624597}, 'timezone': 10800, 'id': 658225, 'name': 'Helsinki', 'cod': 200}

Stockholm Weather: {'coord': {'lon': 18.0686, 'lat': 59.3293}, 'weather': [{'id': 803, 'main': 'Clouds', 'description': 'broken clouds', 'icon': '04d'}], 'base': 'stations', 'main': {'temp': 289.46, 'feels_like': 288.97, 'temp_min': 288.9, 'temp_max': 290.45, 'pressure': 1010, 'humidity': 70, 'sea_level': 1010, 'grnd_level': 1005}, 'visibility': 10000, 'wind': {'speed': 2.06, 'deg': 270}, 'clouds': {'all': 57}, 'dt': 1722581370, 'sys': {'type': 1, 'id': 1788, 'country': 'SE', 'sunrise': 1722566149, 'sunset': 1722625921}, 'timezone': 7200, 'id': 2673730, 'name': 'Stockholm', 'cod': 200}

(venv) root@2c7cada58dbe:~#
        

Agent 1: Starting a message:

{

    "thoughts": [

        "I have successfully retrieved the current weather data for both Helsinki and Stockholm.",

        "The data includes temperature, weather conditions, humidity, and wind speed.",

        "I will now format this information for the user."

    ],

    "tool_name": "response",

    "tool_args": {

        "text": "Here is the current weather data:\n\n**Helsinki:**\n- Temperature: 16.81°C\n- Weather: Broken clouds\n- Humidity: 87%\n- Wind Speed: 5.36 m/s\n\n**Stockholm:**\n- Temperature: 16.31°C\n- Weather: Broken clouds\n- Humidity: 70%\n- Wind Speed: 2.06 m/s"

    }

}

Agent 0: Response from tool 'call_subordinate':

Here is the current weather data:

Helsinki:

- Temperature: 16.81°C

- Weather: Broken clouds

- Humidity: 87%

- Wind Speed: 5.36 m/s

Stockholm:

- Temperature: 16.31°C

- Weather: Broken clouds

- Humidity: 70%

- Wind Speed: 2.06 m/s

Agent 0: Starting a message:

{

    "thoughts": [

        "I have successfully retrieved the current weather data for both Helsinki and Stockholm.",

        "Now I will prepare a response to the user comparing the weather in both locations."

    ],

    "tool_name": "response",

    "tool_args": {

        "text": "Here is the current weather data:\n\n**Helsinki:**\n- Temperature: 16.81°C\n- Weather: Broken clouds\n- Humidity: 87%\n- Wind Speed: 5.36 m/s\n\n**Stockholm:**\n- Temperature: 16.31°C\n- Weather: Broken clouds\n- Humidity: 70%\n- Wind Speed: 2.06 m/s"

    }

}        

Final responce to my question :

Agent 0: reponse:

Here is the current weather data:

Helsinki:

- Temperature: 16.81°C

- Weather: Broken clouds

- Humidity: 87%

- Wind Speed: 5.36 m/s

Stockholm:

- Temperature: 16.31°C

- Weather: Broken clouds

- Humidity: 70%

- Wind Speed: 2.06 m/s