You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This commit was created on GitHub.com and signed with GitHub’s verified signature.
Changed
Ruff replaces pylint.ruff.toml selects ALL under ruff>=0.16.10,<0.17, each
rule left off says why, and pixi run lint runs ruff and ty. The library is fully
annotated; plot_sweep and sweep_sequential take any sequence for input_vars and result_vars, and a Parameter-keyed mapping for const_vars.
Input checks raise TypeError, not AssertionError: a non-list override_defaults, and ResultFloat(units=...) with a non-str unit.
bn.curve raises ValueError when its x and y lengths differ. It used to drop the
extra points silently.
The optuna Pareto message, the zip_results1D* output and the file_server messages
go to logging instead of stdout. Shape._deep_copy is now Shape.deep_copy.
Fixed
A loaded or cached result kept its ResultReference plot.save_result and the
result cache drop the referenced objects, so the panes plot raised IndexError and
was removed from the report. It now shows a placeholder that says the object was not
persisted. An over_time history cell shows the same placeholder instead of the latest
run's object, and a run that did not record the reference leaves a gap.
to_holomap_list returned an empty column. It built each HoloMap and dropped it.
The optuna NaT guard never fired. It checked isnull, which pd.NaT lacks; it now
uses pd.isna.