preserve 8-bit, 0-255 values as native storage for spectral data

[jywarren]

Currently, we use the expression parseInt(line[1] / 2.55)/100 to save image data, which scales it to 0-100%.

In JSON:

spectral-workbench.js/src/SpectralWorkbench.Spectrum.js

Lines 106 to 110 in ef52eac

	_spectrum.average.push({ y: parseInt(line.average / 2.55)/100, x: x })
	if (line.r != null) _spectrum.red.push( { y: parseInt(line.r / 2.55)/100, x: x })
	if (line.g != null) _spectrum.green.push({ y: parseInt(line.g / 2.55)/100, x: x })
	if (line.b != null) _spectrum.blue.push( { y: parseInt(line.b / 2.55)/100, x: x })

and from arrays:

spectral-workbench.js/src/SpectralWorkbench.Spectrum.js

Lines 63 to 66 in ef52eac

	_spectrum.average.push({ y: parseInt(line[1] / 2.55)/100, x: x })
	_spectrum.red.push( { y: parseInt(line[1] / 2.55)/100, x: x })
	_spectrum.green.push( { y: parseInt(line[1] / 2.55)/100, x: x })
	_spectrum.blue.push( { y: parseInt(line[1] / 2.55)/100, x: x })

@Stoft1 has suggested strongly that we make native data storage match original data, which would be 8-bit, 0-255. That'd mean removing the scaling from those lines, and from the matching encodeJSON code at:

spectral-workbench.js/src/SpectralWorkbench.Spectrum.js

Lines 131 to 134 in ef52eac

	average: +(_spectrum.average[i].y * 255).toPrecision(_spectrum.sigFigIntensity),
	r: +(_spectrum.red[i].y * 255).toPrecision(_spectrum.sigFigIntensity),
	g: +(_spectrum.green[i].y * 255).toPrecision(_spectrum.sigFigIntensity),
	b: +(_spectrum.blue[i].y * 255).toPrecision(_spectrum.sigFigIntensity)

Tests would then have to be refactored to look for values in 0-255. We could stick to integer values in sigFigIntensity at:

spectral-workbench.js/src/SpectralWorkbench.Spectrum.js

Line 4 in ef52eac

sigFigIntensity: 4, // since image channel data is 0-255, or 000.4-100.0

On SpectralWorkbench.org, lots of data is stored scaled (not 8-bit). So we'd want to have an easy converter in case people want to compare across.

Update:

• we can create an attribute called spectrum.native8bit which is set to true for all data that does this, so that legacy data will not have this default.
• we can make a spectrum.scaleData() method which checks for spectrum.native8bit and scales data if it is not true.