Implement better baseline subtraction

[Paradoxdruid]

Currently, baseline subtraction is based simply on minimum data values; implement polynomial fit or other solution.

Initial thought: divide regressed data into overlapping windows, use Anderson-Darling test for normality to identify peak location, remove peak window and do polynomial fit to remaining baseline.

SACMES/sacmes/animation.py

Lines 638 to 657 in e88ad78

	###############################################
	# Absolute Max/Min Peak Height Extraction ###
	###############################################
	# -- If the user selects 'Absolute Max/Min' in the 'Peak Height
	# Extraction Settings'
	# -- within the Settings toolbar this analysis method will be used for PHE
	fit_half = round(len(eval_regress) / 2)
	min1 = min(eval_regress[:fit_half])
	min2 = min(eval_regress[fit_half:])
	max1 = max(eval_regress[:fit_half])
	max2 = max(eval_regress[fit_half:])
	################################################################
	# If the user selected Peak Height Extraction, analyze PHE ###
	################################################################
	Peak_Height = max(max1, max2) - min(min1, min2)
	if cg.SelectedOptions == "Peak Height Extraction":
	data = Peak_Height