Installation

AMPLFI can be installed with pip from PyPI

$ pip install amplfi

or directly from source

$ git clone git@github.com:ML4GW/amplfi.git 
$ cd amplfi
$ pip install .  

It is highly recommended that you install AMPLFI in a virtual environment using tools like venv or uv. AMPLFI is managed using uv, which is the tool we recommend.

uv

$ git clone git@git.ligo.org:ml4gw/amplfi.git
$ cd amplfi
$ uv sync

venv

$ git clone git@git.ligo.org:ml4gw/amplfi.git
$ cd amplfi
$ python -m venv ./venv 
$ source ./venv/bin/activate

Supported python versions: 3.10-3.12.

Data Generation

Currently, running the data generation workflow that queries strain data requires utilizing the AMPLFI container. First, set the $AMPLFI_CONTAINER_ROOT where you would like the image stored.

export AMPLFI_CONTAINER_ROOT=~/amplfi/images

Then you can either pull the container from the remote github repository

apptainer

$ apptainer pull ${AMPLFI_CONTAINER_ROOT}/amplfi.sif docker://ghcr.io/ml4gw/amplfi/amplfi:main

docker

$ docker pull ghcr.io/ml4gw/amplfi/amplfi:main

Supported python versions: 3.10-3.12.

Or build the container locally

$ git clone git@git.ligo.org:ml4gw/amplfi.git
$ cd amplfi
$ apptainer build $AMPLFI_CONTAINER_ROOT/amplfi.sif apptainer.def

This container can be also be used for launching the training and tuning pipelines, but is not strictly required. See the container documentation for more information.