Installation

Automatic Installation

For participating to the data generation effort only. If you want to interact with the datasets or submit datasets yourself, you must use one of the other installation methods.

curl https://raw.githubusercontent.com/OpenDrugDiscovery/openfractal-client/main/scripts/auto_install.sh | bash

Docker

Public docker images are available at https://github.com/OpenDrugDiscovery/openfractal-client/pkgs/container/openfractal-client.

docker run --rm -ti ghcr.io/opendrugdiscovery/openfractal-client:main

Using mamba

The QCFractal libraries ecosystem have not yet been released. It is actively being developed in the next branch at https://github.com/MolSSI/QCFractal/tree/next.

For now here is a minimal Conda env.yml file you can use to perform the installation (this will be simplified in the future):

channels:
  - hadim/label/qcportal_next
  - conda-forge/label/libint_dev # for psi4
  - conda-forge

dependencies:
  - python >=3.9
  - pip

  # QCPortal deps
  - hadim/label/qcportal_next::qcportal
  - hadim/label/qcportal_next::qcfractalcompute

  # Compute managers
  - parsl
  - psi4 =1.8
  - openmm
  - openff-forcefields
  - openmmforcefields

  # Optional
  - datamol
  - openff-toolkit
  - zarr

  # Optional utilities
  - loguru
  - typer
  - python-dotenv

Put the above YAML file into env.yml. Then:

micromamba create -n openfractal -f env.yml

In the future, it will be as simple as micromamba create -n openfractal openfractal-client.