Skip to content
FAIRSCAPE

Quick start

This walks one crate through every step: create it, check it, view it, grade it and publish it. The pages at the bottom are what you end up with.

Install

Terminal window
pip install fairscape-models fairscape-conversion
pip install git+https://github.com/fairscape/fairscape_artifacts
pip install git+https://github.com/fairscape/AIreadiness-grader
pip install git+https://github.com/fairscape/fairscape_publish

The five steps

  1. Create. Convert a Datasheet for Datasets into a crate with fairscape-conversion. Other formats (Snakemake, Cromwell, Galaxy, MLflow, REDCap, …) work the same way.

    Terminal window
    mkdir my-crate
    python -m fairscape_conversion.core.cli convert d4d import datasheet.yaml my-crate/ro-crate-metadata.json

  2. Validate. Check it against the data model in fairscape-models:

    import json
    from fairscape_models import ROCrateV1_2
    

    ROCrateV1_2.model_validate(json.load(open("my-crate/ro-crate-metadata.json")))

  3. View. Build the datasheet, evidence graph and preview with fairscape-artifacts:

    Terminal window
    fairscape-artifacts all my-crate
    # writes ro-crate-datasheet.html, ro-crate-evidence-graph.html and ro-crate-preview.html into my-crate/

  4. Assess. Grade it against the AI-Ready rubric with the AI-Readiness grader:

    Terminal window
    fairscape-evidence my-crate -o review
    # writes review/ai-ready-review.html

  5. Publish. Check that it’s ready, then create a draft deposit with fairscape-publish:

    Terminal window
    fairscape-publish check my-crate
    fairscape-publish zenodo my-crate --token $ZENODO_TOKEN --sandbox

    To host crates on your own server instead, use fairscape-lite.

What you end up with

These are the pages steps 3 and 4 produce, for a crate that came out of the Snakemake variant-calling tutorial. Each is a single HTML file.