codemonkeys 0.0.5

CodeMonkeys is a highly configurable tool for creating/running strategic AI automations (and more). Beyond that, its a thoughtful and extensible codemonkeys for creating automation abilities. CodeMonkeys' design was heavily influenced by how much it was used in its own development, naturally converging to a sublime synergy of the tool and the codemonkeys files.

CodeMonkeys 🐵

Are you

... a highly-technical individual with an interest in taking more control over your automated AI coding?

... a lightly-technical individual who wants to automate AI and can probably figure out how to edit a config?

... a creature with the capacity to learn how to press buttons and a strong motivation to do the AI?

Overview 🌐

CodeMonkeys, while maintaining a strong emphasis on coding tasks, is designed to be a purposefully more focused alternative to broad spectrum tools like AutoGPT. The goal is to reduce GPT API call wastage and provide more control, harnessing the power of automation for tasks that are reliable, predictable, and precisely tailored to user needs, only involving AI at crucial areas of strength. The project is built around a suite of adaptable Python scripts and user-friendly config utilities. This approach ensures that while you have a solid foundation for customization, you also maintain the efficiency and potency of the core framework. CodeMonkeys aims to strike a balance between adaptability and specificity, providing a less feature-rich but more extensible and capable tool for your unique automation needs.

The Default Automation

The heart of CodeMonkeys is main.py, a generic but complete template for improving whatever is in your WORK_PATH. This file is the default entry-point and is designed to work well out of the box, and to be powerfully custom using only monkey-manifest.yaml configs. main.py is a top-down look at the default Monkey Program. During installation, a copy of this file is stored in programs/my-first-program. The programs directory is a location for your own variations of the main.py file, should you feel the urge to start writing your own programs. If you're thinking "already there, pal", I see you, but I'd recommend you start by playing with the monkey-manifest and default program. If you're thinking "that's too much, man", no worries, the default program and monkey-manifest are very powerful, and the proprietary programs in the programs directory will be expanded.

If the default program main.py is the heart of CodeMonkeys, then

In addition to the monkey-manifest and main.py default program, CodeMonkeys commits to being moddable with a simple and powerful design for using/implementing scripts. The scripts directory is...

Getting Started 🚀

You have three options (try #1 first):

1. Clone CodeMonkeys to use as an automation tool. (or contribute 🥺)
2. Fork CodeMonkeys to use as a framework for making your own automation tool.
3. You do you, boo-boo. The line between tool/framework is blurry enough for the weird thing you're thinking.

This is a source package. It isn't installed per se, but is added to Python's sys.path in the setup script to allow easy module imports while maintaining a paradigm that serves the framework side of things.

Important: Whatever your intentions, co-existing local repos must have distinct directory names. I think this project is best as a source package overall, but I'm relying on sys.path for global imports so namespacing is required or Python cannot accurately select modules. To fix imports run monk fix-namespace.

If you're using 💻🐒s as a framework this is a good way to rename the project. The source code also makes use of the directory name in this spirit in other ways I won't go into right now.

Run the setup script when you're ready:

python3 setup.py

I have relative confidence this works well on macOS/Linux, and I've tried to handle Windows well without over-committing at this early stage. If you have issues with Windows at any time, please open an issue and give me all the info you can.

Usage 🎮

CodeMonkeys operates via the monk command which runs Python scripts located in the scripts directory. For example, to kick off a monkey, simply use monk kickoff [monkey-name] -n50, where n is the maximum number of iterations to run. There is no default value for -n as a measure to encourage thoughtful use and provide a safety measure for those less familiar with the documentation.

Customization and Updates 🛠

CodeMonkeys is designed to grow and adapt with you. The framework provides several built-in features that allow you to tailor the project to your needs.

• Monkey-Manifest: Define and manage your custom 'monkeys' or tasks in a centralized config file.

• Local .env Files: Manage your environment variables locally without affecting the global settings.

• Import/Export: Easily share your customizations or move between different systems.

  • Export: Use monk export to create an archive that includes your custom main.py, monkey configs, and existing backups of your main file. The export is timestamped and can be moved to other systems or shared with others.

  • Import: Use monk import [path-to-exported-file] to import a previously exported archive. This will replace your current main.py and monkey configs with those in the archive. But don't worry - before anything is replaced, your current main.py is backed up with a timestamp, and your backups are merged with those from the import.

I strive to make CodeMonkeys easy to update, even with your customizations. However, extensive modifications may complicate the update process. Please note that while I'm working hard to make updates as seamless as possible, it's always a good idea to back up your work before updating.

For easy export of your monkey-manifest.yaml, automations, barrels, and custom directories, simply run

monk export

The export process will also create a .env-backup file that contains your current .env file. .env-backup is in the default .gitignore, and is not included in the export. Please also verify that you are not committing your exports if you have sensitive information outside your .env file, such as in the monkey-manifest.yaml (please don't do that actually).

Framework Overview 📁

• monk or monk.py: CodeMonkeys CLI base command.
• pack: A "pseudo-package" structured for easy inspection/modification.
• commands: Python, bash, or bat scripts to be run via the monk command.
• monkeys: Generated configs for different 'monkeys' and backups of previous configs.
• monkeys/monkey-manifest.yaml: The centralized config file for all monkeys.
• automations: Python scripts that use monkey configs to run automations.
• automations/default.py: Pre-packaged automation, which enables a range of tasks via monkey configs
• barrels: Python scripts meant to orchestrate more than one automation.
• barrels/default.py: Pre-packaged barrel, which enables running a series of parallel or sequential automations.
• modules: Python modules for both framework support and customized functionality.
• definitions.py: A root-level file providing absolute path variables for both framework and custom use.

Attribution

When forking or creating derivative works from CodeMonkeys, I kindly ask you to:

1. Clearly acknowledge the CodeMonkeys project and provide a hyperlink to the CodeMonkeys repository in the first section of your README.md file.
2. Prominently display an acknowledgement of the usage of CodeMonkeys in any application, framework, or derivative work which is built on top of the CodeMonkeys framework and is distributed through an application store, website, social media, or any other distribution method.
3. Although I expect no acknowledgement when CodeMonkeys is used as a tool to accomplish things, any mention of its usefulness is greatly appreciated.

Things To Include

psuedo-package core/usr monkey-config class env class monkey-config-defaults add-monkey auto-generation of monkeyconfig and env and .env.default replace_prompt_str cop_paths