Skip to main content

A simple, lightweight loop for your agent.

Project description

A simple, lightweight loop for your agent. Start/stop, step-through, and more.

Lint and Test PyPI version

Quickstart

from agentloop import start, stop

def step_one(next_output, loop_data):
    print("step_one")
    return next_output

def step_two(next_output, loop_data):
    print("step_two")
    return next_output

# Run the loop
loop_data = start(steps=[step_one, step_two])

# Pause the loop
pause(loop_data)

# Unpause the loop
unpause(loop_data)

# Stop the loop
stop(loop_data)

Installation

pip install agentloop

Usage

This package provides a set of functions to perform a stepped or continuous loop of operations in a separate thread. This can be used for various purposes like running a continuous process that can be controlled from the outside, a debugging tool for a cycle of operations, etc.

Steps

Each step must take in input from the last step and return output for the next step. The first step will receive None as input, and this will need to be handled. You can either start with an initialization step that returns the initial input, or you can check for None in the first step and return the initial input if it is None.

Example steps:

def step_one(next_output, loop_data): # next output None first run, then received from step_two
    print("step_one")
    return next_output # next output sent to step_two

def step_two(next_output, loop_data): # next output received from step_one
    print("step_two")
    return next_output # next output sent to step_one

steps = [step_one, step_two]

Function start

start(steps, stepped=False, step_interval=0.0)

Description

Starts the main loop in a separate thread. This loop will run the steps given, in a continuous or stepped manner.

Parameters

  • steps : a list of functions that should be executed in the loop. Each function should accept a single argument and return a single value which will be passed as an argument to the next function. The first function will receive None as an argument.

  • paused (optional): a boolean value that determines whether the loop should run in paused step mode or not. Defaults to False.

  • step_interval (optional): a float value that determines the time interval between steps in seconds. Defaults to 0.0.

Returns

A dictionary containing

  • stop_event: an instance of threading.Event that's used to control stopping of the loop.
  • step_event: an instance of threading.Event that's used to control stepping.
  • started_event: an instance of threading.Event that's set when the loop starts running.
  • thread: an instance of threading.Thread that's running the main loop.

Function stop

stop(loop_data)

Description

Handles stopping of the loop.

Parameters

  • loop_data: a dictionary containing the stop_event and thread which is returned by the start function.

Returns

None


Sure, here are the updated sections for the pause and unpause functions in your README file.


Function pause

pause(loop_data)

Description

Pauses the loop. When paused, the loop will not execute the next step until it's either stepped using the step function or unpaused using the unpause function.

Parameters

  • loop_data: a dictionary containing the pause_event which is returned by the start function.

Returns

None


Function unpause

unpause(loop_data)

Description

Resumes the loop after it has been paused with the pause function. If the loop is not paused, calling this function has no effect.

Parameters

  • loop_data: a dictionary containing the pause_event which is returned by the start function.

Returns

None

Function step

step(loop_data)

Description

Performs a single step in the loop.

Parameters

  • loop_data: a dictionary containing the step_event which is returned by the start function.

Returns

None


Function loop

loop(steps, paused=False, loop_data=None, step_interval=0.0)

Description

Runs the step array in a loop until stopped.

Parameters

  • steps: a list of functions that should be executed in the loop. Each function should accept a single argument and return a single value which will be passed as an argument to the next function. The first function will receive None as an argument.

  • paused (optional): a boolean value that determines whether the loop should run in paused / stepped mode or not. Defaults to False.

  • loop_data (optional): a dictionary containing stop_event and step_event instances. If not provided, new events will be created.

  • step_interval (optional): a float value that determines the time interval between steps in seconds. Defaults to 0.0.

Returns

None

Function use_keyboard

use_keyboard(loop_data, input_key=keyboard.Key.space)

Description

Creates a keyboard listener and attaches it to the provided loop data object. This listener listens for a specified key press, and when detected, steps the loop (sets the 'step_event').

Parameters

  • loop_data: A loop data object, which is typically created by the start function.

  • input_key (optional): The keyboard key that the listener will react to. Defaults to keyboard.Key.space.

Returns

Returns the updated loop dictionary with the newly created keyboard listener added to it. Note: Pass the updated dictionary to the stop function to also stop the keyboard listener

Defaults to Spacebar

Example

loop_data = {
    "stop_event": threading.Event(),
    "step_event": threading.Event(),
    "started_event": threading.Event(),
    "thread": None,
}

updated_loop_dict = use_keyboard(loop_data, input_key)

Testing

pytest test.py

Publishing

bash publish.sh --version=<version> --username=<pypi_username> --password=<pypi_password>

Contributions Welcome

If you like this library and want to contribute in any way, please feel free to submit a PR and I will review it. Please note that the goal here is simplicity and accesibility, using common language and few dependencies.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

agentloop-0.2.2.tar.gz (5.7 kB view hashes)

Uploaded Source

Built Distribution

agentloop-0.2.2-py3-none-any.whl (6.2 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page