Experiment Automation Agents (EAA)

Installation

Option 1: install via pip

First, create a conda environment with Python 3.11:

conda create -n eaa python=3.11
conda activate eaa

Then clone the respository to your hard drive. CD into the repository's root, and install it with

pip install -e .

The -e flag allows any changes made in the source code to immediately take effect without reinstallation when you import the package in Python.

Option 2: install via uv

uv is a Python environment and package manager that is fast and dependency-deterministic, offering better reproducibility guarantee. Unlike a conda environment, a uv virtual environment is installed at the root directory of a project instead of in a centralized location, making it more portable.

First, install uv using pip:

pip install uv

Then clone the repostiroy, CD into the respository's root, and create a new environment there:

uv venv --python 3.11

This creates a new virtual environment in ./.venv.

Activate the environment using

source .venv/bin/activate

Then install the dependencies using

uv pip install -r requirements.txt

The requirements.txt in this repository is generated by uv on the developers' side and contains all dependencies and their exact version numbers, thereby maximizing the deterministicness of the installation.

Finally, install the package itself:

uv pip install -e .