suggestion for new task -- re-do spectra given list of detections

[Matthew Whiting]

A suggested enhancement: It might be nice to be able to re-draw the spectra of a list of detections (or even a subset of a given detection list) with different specfications (eg. draw the summed spectra as opposed to the peak spectra). Why is this necessary? If a search takes a long time (eg. for a large cube) and you just want to see the integrated spectra rather than the peak spectra, it would be good to have a program to read the detected pixels from the log file and produce the spectrum from them. Note that it may not be possible to do the reconstructed spectrum, if it was not saved -- I don't think it would be worth incorporating any reconstruction into this program.

[Matthew Whiting]

Mostly done, in revision [218]. New program works, but might want to change the command-line options to make more user friendly.

[Matthew Whiting]

With the new way of reporting the central pixel (centroid, average, peak flux), this could also be changed using this program. The central location that is printed out above each spectrum would then be changed in the new spectra.

[MatthewWhiting]

Email from Mary Putman

Hi Matt,

Tina and I are hoping to put some Duchamp results in a couple of papers.  
One problem we are having is when we dig into the data it picks up junk and real 
clouds and the resulting moment map shows both.   Is there any way to trick the 
final ps plotter to only show the clouds we define as real?
Tina thought maybe we could use the mask fits file, but we haven't had much luck 
with that file yet.

Thanks much,

Mary

This suggests another way to re-do the plots: give a selection of sources that are done as individual images. This could be done either as an input parameter, eg:

objectList   1,3,5-7,11

or via a command-line argument: -o 1,3,5-7,11

[MatthewWhiting]

OK, time to re-think this. Rather than defining a new program (which we have been calling Selavy), how about just making this an option for Duchamp? The way it would work would be to have a user parameter, say usePrevious, which, when set to true, would read the detections from the logfile, rather than doing the searching.

We could also incorporate the objectList option shown above, so that if it is given, only those objects are printed. They could be done as individual plots as well as the single ps file with all of them.

[MatthewWhiting]

This was introduced in v1.1.6, with the usePrevious parameter enabling its use.

Closing this ticket.