Write a ZnTrack Node

Setup environment

1. Install miniconda https://docs.conda.io/en/latest/miniconda.html
2. Install poetry https://python-poetry.org/docs/
3. Create an environment conda create -n asemd python
4. Activate environment conda activate asemd
5. Install package poetry install

Test package

run pytest . to check if all the tests pass.

The package should work as follows:

import ase_md

atoms = ase_md.simulator.generate_atoms(size=2)

    atoms_list = ase_md.simulator.run_simulation(
        atoms=atoms, temperature=300, timestep=1.0, dump_interval=5, steps=20
    )

    rdf = ase_md.simulator.compute_rdf(
        atoms_list=atoms_list, rmax=1.0, nbins=50, elements="Cu"
    )

which builds a DAG like:

Rewrite the package

• Add an Example using DVC (without ZnTrack)
• Add an Example using ZnTrack
• Add an Example using MLFlow
• Add an Example using Hydra

You can also consider other packages you want to try out. Some can be found here https://github.com/pditommaso/awesome-pipeline.

Run experiments:

Test your workflow with different parameters for all Nodes and compare them.