Installation
There are 3 options for installing locuaz. Conda + pip (Option 1) is recommended, since its both lightweight and quick. A pure conda install is also available (Option2). It is also possible to compile from source (Option 3). In all cases a Post-installation step is required.
Prerequisites
We recommend using Mambaforge instead of pure conda, since it’s much faster
Stable release
conda + pip (Option 1)
Create a conda environment from the usr_deps.yaml
, which looks like this:
name: locuaz
channels:
- conda-forge
dependencies:
- conda-forge::python>=3.10
- conda-forge::ambertools>=22.0.0
- conda-forge::tensorflow
- conda-forge::openbabel
To create the environment:
mamba env create -f usr_deps.yaml
Then, activate the environment and install the protocol through pip:
mamba activate locuaz
pip install locuaz
This option takes an extra step with respect to using only conda, but the process will be faster and the resulting environment won’t be as heavy.
apptainer container (Option 2)
You can pull any version from the GitHub repository like this:
apptainer pull oras://ghcr.io/pgbarletta/locuaz.sif:0.6.2
These containers will be signed with a public key, so you can verify them:
apptainer verify --url https://keys.openpgp.org locuaz.sif
apptainer containers are available for a plug-and-play approach. Just change the way you call locuaz. If you’d call the locuaz entry-point like this:
locuaz config.yaml
this is how you’d call it when inside a container:
apptainer exec --nv locuaz.sif locuaz config.yaml
Check Job submission for more info on using locuaz with containers.
From sources (Option 3, developers only)
Clone the repo:
git clone https://github.com/pgbarletta/locuaz
Create the environment and install all the necessary dependencies at once:
mamba env create -f dev_deps.yaml
And inside the newly cloned dir, install locuaz in development mode:
pip install -e .
That’s it. You can also change the environment’s name by editing the name field of the dev_deps.yml file, before creating it.
Post-installation
If you want to use the dlp
mutator You’ll also have to get DLPacker’s weights and place them on a dedicated
dlpacker
(actual name doesn’t matter) directory, more info on Mutators.
You also may want extra scorers, check the Scorers