Output from 1D searching not very useful

[MatthewWhiting]

We advertise that Duchamp works fine for 1D spectra, but the output is not really good enough:

• The final catalogue lists the channels in the X-column, rather than the Z (which is what you'd expect).
• The RA/DEC calculations appear confused.
• The spectra output appears to use the 2D case, plotting a map of the spectrum (ie. as a line of pixels) when it should plot the spectrum.

In short, the 1D processing just seems to be a special case of the 2D processing, albeit with the 1D search. Work needs to be done specialising the handling of the search results for the 1D case.

[MatthewWhiting]

Here's what should happen in the 1D case:

• When searching, make the xvalue a zvalue, and set x & y to zero
  • Similar transforms are done for search3DarraySpectral
• Perhaps we could even make the dimension array [1,1,zdim], so that we maintain a mock 3D cube, just with degenerate axes.
  • Do a check in the initialise step, and if we have only one good dimension, assume it is the z one and set the axes accordingly.
• spatial detection plot should instead be a spectral plot with the detection ranges marked
• duchamp-Spectra.ps should have the same output as for 3D case, but without the moment map
• moment map plotting is not done

[MatthewWhiting]

The first part of this is done in [977] and [978], making it a mock 3D cube. This results in the searching returning z-values instead of x-values, which is what we want.

Now have to adapt the graphical output.

There is still an error with a double free occuring at the end of the execution -- this needs to be followed up.

[MatthewWhiting]

This is now greatly improved. The graphical output now consists of a spectrum with each final detection highlighted, as well as a bar along the top indicating which channels were detected (including those rejected for failing the minimum criteria).

I've also put the detection and growth thresholds so you can easily see what is detected and what was just missed. (These have also been added to the regular spectral output, at least for the peak flux spectra.)

Closing this ticket, but might make another one to improve the formatting of the plots, as well as the class design for the relevant code.

[inetryconydot]

Pluto's weather is more active than we thought With each batch of images that rolls in from New Horizons, scientists learn more about Pluto's characteristics.