Wire GUI equivalence stability controls

[sergioald]

Summary

Wires the equivalence-style sinuosity stability controls through the GUI and configuration layer.

Changed

• ldsfl/gui_utils.py

  • adds GUI run-control fields for the equivalence stability diagnostic;
  • includes equivalence settings in config round-tripping and preview output;
  • ignores unknown future run-control fields when loading saved configs;
  • validates equivalence diagnostic settings before a run.

• gui_ldsfl.py

  • exposes equivalence diagnostic controls in the advanced GUI run settings;
  • forwards those settings from the GUI to the solver;
  • uses the configured equivalence settings when refreshing the live sinuosity panel;
  • safely defaults missing GUI variables in lightweight/headless tests and older config states.

• ldsfl/main.py

  • adds a guard against solver runs with no enabled stop criteria;
  • forwards the selected equivalence method into the combined sinuosity stability diagnostic.

• run_ldsfl.py

  • exposes the equivalence diagnostic method through the CLI.

• tests/test_gui_equivalence_controls.py

  • checks config round-tripping for equivalence settings;
  • checks preview output;
  • checks invalid equivalence settings are rejected;
  • checks the no-enabled-stop-criteria guard;
  • checks that the combined diagnostic forwards the selected equivalence method.

Why

The GUI already exposed the basic sinuosity stability stop switch, but the equivalence-style diagnostic parameters were still effectively hard-coded in the GUI/live-panel path.

After PR #36, the equivalence diagnostic has a selectable statistical method, with increment as the safer default for unit-root-like histories and hac retained for explicit comparison. The GUI should therefore preserve, validate, display, and forward the same controls as the solver/CLI path.

This PR keeps the GUI, saved config files, preview output, CLI options, and solver diagnostics consistent.

Validation

python -m py_compile ldsfl/gui_utils.py ldsfl/main.py run_ldsfl.py gui_ldsfl.py tests/test_gui_equivalence_controls.py
python -m pytest tests/test_gui_equivalence_controls.py tests/test_gui_sinuosity_metrics.py
python -m pytest
python -m run_ldsfl --base-dir . --cases 1 --max-steps 100 --no-plots

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: ec24bad6bb

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".

[chatgpt-codex-connector]

Preserve blank transient step when saving GUI configs

When a user leaves the new “Equiv. transient step” field blank to request all-history analysis, _build_config() correctly turns it into None, but this gui_state entry rewrites the blank value to the default string "40000". On load, the GUI applies config.run and then overwrites the field from gui_state, so a saved all-history configuration silently comes back as post-40000-only and future runs use the wrong diagnostic window. Preserve the raw empty string here or omit the GUI override when the field is blank.

Useful? React with 👍 / 👎.