Reproducibility & Provenance: Develop `aiida-openimpala` Plugin

[jameslehoux]

Labels: ecosystem, hpc-orchestration, phase:4-hpc
Priority: Medium (High for Institutional Adoption)

Description

To support FAIR (Findable, Accessible, Interoperable, Reusable) data principles and seamless HPC execution, OpenImpala should integrate with workflow orchestration tools, specifically AiiDA (the standard for computational materials science).

We need to develop a standalone aiida-openimpala plugin. This will allow researchers to define their OpenImpala simulations in Python, while AiiDA automatically handles the SSH transport to the HPC cluster, Slurm job submission, remote execution, and automatic provenance tracking (saving the exact inputs, versions, and outputs into a directed acyclic graph database).

Acceptance Criteria

• Create a new repository: aiida-openimpala.
• Define the AiiDA CalcJob plugin for the OpenImpala CLI, mapping AiiDA dictionary nodes to OpenImpala's JSON config structure.
• Define an AiiDA Parser to automatically extract the final transport properties (Tortuosity, Diffusivity) from OpenImpala's output CSVs into an AiiDA Dict node.
• Provide an example notebook demonstrating a full automated submission to a remote Slurm cluster via AiiDA.