[Josef-Haupt]

• Segments: map the audio-speed slider like every other tab; 0 and negative slider values no longer reach the library as a raw speed
• Review: ignore confidences outside [0, 1] (the old check could never trigger); parse the filename in one place
• Evaluation: the "Overall" row follows programmatic changes of the averaging radio (reset to defaults, presets)
• Evaluation: an unparsable recording duration or an empty metric selection raises an error instead of silently falling back
• Evaluation: a duration that differs from the one the processor was built with is rejected on calculate and export
• Train: validate the cache location per mode (existing file for "load", existing directory plus a non-empty name for "save") instead of a raw TypeError
• Train: switching the cache mode clears the picked location (behaviour change)
• Train: derive crop-overlap visibility and interactivity from cache mode and crop mode together, so preset loads can't leave it inconsistent
• Analysis / Train: use gu.slider_to_value instead of hand-copied formulas; stop clamping fmin/fmax after they were already validated
• Plot download: take the extension from the file name, not the first dot in the path (C:\Users\j.doe\plot.png became plot.png.png); save JPG/JPEG, WebP, BMP, GIF and TIFF as chosen
• Network shortcuts: one unreadable .lnk no longer hides all the others
• Single file: the Kaleidoscope export dialog offers *.csv, matching the file it writes
• Localization: two new keys in all language files, eval-tab-error-no-metric-selected and validation-no-valid-cache-file-name
• Tests: handler-level tests for every fix (new tests/gui/test_{analysis,evaluation_tab,review_tab,segments_tab,single_file_downloads,utils}.py)

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

An organization owner can choose the repository organization as the billing source for reviews requested by members with a paid Copilot license. Learn about Copilot code review billing.