Add data for MVP example

[mattwthompson]

This is from #148 (comment)

Closes #196

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.44%. Comparing base (a8ca593) to head (7892f45).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #195   +/-   ##
=======================================
  Coverage   83.44%   83.44%           
=======================================
  Files          58       58           
  Lines        5177     5177           
=======================================
  Hits         4320     4320           
  Misses        857      857           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

🔵 Needs a closer look

It modifies the CI workflow and adds binary data assets, which warrant a human maintainer's final sign-off even though no code defects were found.

0 open findings

What changed in this PR

This PR advances the MVP milestone from issue #148 by replacing the previous compute-submission example with a lightweight data-loading entry point and bundling the example datasets. The old examples/mvp/run.py, which submitted simulation target batches via SimulationWorkflow, is gutted down to a stub that unpacks bundled energy/thermo tarballs and loads them as HuggingFace datasets.Dataset objects. Supporting plumbing is added so the example can be invoked through pixi and (conditionally) from CI.

Changes:

• Add examples/mvp/data/energy.tar.gz and thermo.tar.gz and rewrite run.py to extract and load them.
• Add a run_mvp pixi task that runs the example.
• Add a CI step to run the MVP example, gated to GPU environments.

File	Description
examples/​mvp/​run.py	Replaces the simulation-submission example with a stub that unpacks the bundled tarballs and loads the thermo/energy datasets.
pixi.toml	Adds the run_mvp task to execute the example.
.github/​workflows/​gh-ci.yaml	Adds a Run mvp CI step gated on GPU environments (removes a stray blank line).

Notes from verification (no code changes required):

• The bundled tarballs extract to top-level energy/ and thermo/ directories, so the subsequent Dataset.load_from_disk(root / "data/energy") and .../data/thermo paths are correct.
• tar.extractall(..., filter="data") is valid given requires-python = ">=3.12".
• datasets is pinned to 4.*, where load_from_disk accepts a PathLike, so passing a Path works.
• The new Run mvp step is gated on contains(matrix.environment, 'gpu'); no CI matrix environment (py312amber, py312openeye) matches, so it never runs today. This is explicitly called out as intentional in the inline comment ("not committed to running this in CI").

🧠 Review effort: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.