Patch release with two bug fixes found on real Shimadzu data.
Fixes
valuespassed toHandler.read_*now follow the files in numeric file-name order (#51). Reaction times or concentrations given asvalues=[...]were zipped onto file paths sorted as text, sosample_10.38min.txtcame beforesample_2.88min.txtand the time course was silently scrambled. Every reader now sorts with one natural-sort key (numbers inside names compared as numbers).
Behaviour change: anyone who passedvaluesin text order of their file names gets a different, now correct, pairing. Zero-padded or single-digit names are unaffected.Handler.add_standardworks with several molecules on one calibration handler (#52, fixes #14). It collected the areas of every assigned peak instead of only the molecule's own, and failed with "Number of peak areas 2N and concentrations N do not match".
CI
- The test step runs under
PYTHONUTF8=1so calipytion's emoji output survives the Windows console.