Verify reproducibility fixture

[sergioald]

Summary

Adds a real regression check for the bundled short reproducibility fixture.

Changed

• examples/reproducible_case1_short/README.md

  • clarifies which reproducibility checks are portable across machines;
  • explains that floating-point reference values are only compared when the recorded scientific-Python environment matches.

• examples/reproducible_case1_short/expected_tree.txt

  • records the final partial variable-history file produced by the documented short run.

• examples/reproducible_case1_short/regenerate.py

  • adds a helper script to regenerate the fixture output tree and record the local Python/scientific stack.

• tests/test_reproducible_example.py

  • runs the documented short case in a temporary workspace;
  • checks the produced output tree against expected_tree.txt;
  • checks stable integer counters and stop criteria against expected_summary.json;
  • checks portable physical/output invariants for generated CSV snapshots;
  • compares pinned floating-point values only when the recorded environment matches the current one.

Why

The repository already ships a short reproducibility fixture under examples/reproducible_case1_short/, but that fixture was not being checked by the test suite.

This PR turns it into an executable regression test while avoiding brittle cross-machine failures from BLAS/NumPy/SciPy floating-point differences.

It also fixes the expected output tree so the documented command includes the final partial variable-history block.

Validation

python -m py_compile examples/reproducible_case1_short/regenerate.py tests/test_reproducible_example.py
python -m pytest tests/test_reproducible_example.py
python -m pytest
python -m run_ldsfl --base-dir . --cases 1 --max-steps 100 --no-plots

Notes

This is a fixture/test/documentation PR.

It does not change the morphodynamic update algorithm, flow solvers, curvature calculation, stability diagnostics, CLI interface, GUI interface, or output naming.

The floating-point comparison is intentionally guarded by environment.json so the test remains useful on CI and on local machines with different scientific-Python stacks.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 00cf711c25

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".