Stitch1D special values

[Owen Arnold]

Message from Tim:

It looks like most of my problems with stitching boils down to how NaN’s and Inf’s are handled. Is there a way to disregard them when calculating the scale factors but propagate them in the final data for consistency?  This way they would not  block the scaling or visualization of the rest of the curve.

Yes. This could either be done by throwing those values away, or using ReplaceSpecialValues.

[Owen Arnold]

This doesn't work but should

import numpy
goodx = numpy.array([0.0, 1.0, 2.0,  3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0])
goody = numpy.array([0.8, 0.7, 0.6,  0.5, 0.4, 0.3, 0.2, 0.1, 0.0])
bady = numpy.array([0.7, 0.6,  0.5, 0.4, float('nan'), 0.2, 0.1, 0.0, -0.1])


ws1 = CreateWorkspace(DataX=goodx, DataY=goody)
ws2 = CreateWorkspace(DataX=goodx, DataY=bady)

stitched, sf = Stitch1D(ws1, ws2, Params=1.0, StartOverlap=3.0, EndOverlap=6.0)

plotSpectrum([ws1, ws2, stitched], 0)

This does work

import numpy
goodx = numpy.array([0.0, 1.0, 2.0,  3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0])
goody = numpy.array([0.8, 0.7, 0.6,  0.5, 0.4, 0.3, 0.2, 0.1, 0.0])
bady = numpy.array([0.7, 0.6,  0.5, 0.4, 0.3, 0.2, 0.1, 0.0, -0.1])


ws1 = CreateWorkspace(DataX=goodx, DataY=goody)
ws2 = CreateWorkspace(DataX=goodx, DataY=bady)

stitched, sf = Stitch1D(ws1, ws2, Params=1.0, StartOverlap=3.0, EndOverlap=6.0)

plotSpectrum([ws1, ws2, stitched], 0)

[Owen Arnold]

refs #9779. Add failing tests.

Changeset: 4fd0fbe94eb9a2b6c08ed74fb82fd65e61af5156

[Owen Arnold]

refs #9779. Make fix using ReplaceSpecialValues.

Changeset: e3061c3c055134559616383b6e3c9374be4cc119

[Owen Arnold]

Tester:

The script below should print a real-number scale factor. This would not work previously.

import numpy
goodx = numpy.array([0.0, 1.0, 2.0,  3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0])
goody = numpy.array([0.8, 0.7, 0.6,  0.5, 0.4, 0.3, 0.2, 0.1, 0.0])
bady = numpy.array([0.7, 0.6,  0.5, 0.4, float('NaN'), 0.2, 0.1, 0.0, -0.1])


ws1 = CreateWorkspace(DataX=goodx, DataY=goody)
ws2 = CreateWorkspace(DataX=goodx, DataY=bady)

stitched, sf = Stitch1D(ws1, ws2, Params=1.0, StartOverlap=3.0, EndOverlap=6.0)

plotSpectrum([ws1, ws2, stitched], 0)

print sf

Unit tests have been added to cover this functionality.

[Michael Reuter]

Merge remote-tracking branch 'origin/feature/9779_stitch_special_values'

Full changeset: ​2288c1865ecb5a083d3da9ce131ebd0b104dc20b

[Stuart Campbell]

This ticket has been transferred to github issue ​10621