radon_wrapper.readme

###############
####Summary####
###############

The program, radon_wrapper.pro, is my own personal script to load
MaNGA data, prepare the data, apply the radon transformation, analyze
the radon transformation, and save the output.  The steps it goes through are:

1) Load MAPS data (either stellar or gas velocities)

2) recenter the maps so the optical center is directly over the central pixel

3) (Optional) apply an additional offset if a alternate center is known

4) Mask any bad data (combine existing masks as well as a SNR cut)

5) apply the radon transform, including propagating the covariance
matrix if requested

6) Trace the ridge in the radon transformation defined by where the
value of theta where the transformation is minimized as a function of
theta

7) Save the output to a fits file, and create a diagnostic plot

###########################
###Description of Output###
###########################

The wrapper writes a fits file with 13 different extensions each
containing different useful information. The following is a short
explanation of each extension (extension number and extension name,
followed by the description):

Extension 0 -- Empty

Extension 1 -- 'velocity'

	  The velocity field (after preprocessing)

Extension 2 -- 'velocity_error'

	  The input velocity field errors (after preprocessing)

Extension 3 -- 'velocity_mask'

	  The velocity field mask (after preprocessing)

Extension 4 -- 'radon'

	  The absolute radon transform of the veloctity field

Extension 5 -- 'radon_rescale'

	  A rescaled version of the radon transform where the range of
	  values at fixed rho (each row in the matrix) range from 0 to
	  1. This is mostly for viewing purposes.

Extension 6 -- 'radon_mask'

	  A mask indicating where the radon transform may be
	  unreliable.  It flags infinite values, zeros, regions where
	  the radon "aperture" overlapped masked data in the velocity
	  field, and regions where the radon "aperture" extended over
	  the edge of the map.

Extension 7 -- 'radon_error'

	  Error on radon transform

Extension 8 -- 'radon_theta'

	  Array of theta values for this transform (x-axis)

Extension 9 -- 'radon_rho'

	  Array of rho values for this transform (y-axis)

Extension 10 -- 'radon_length'

	  Number of pixels used to calculate the transform at a given
	  (rho,theta) location. This can vary due if the radon
	  aperture is not finite, near the edge or masked regions of a
	  velocity field, and small variations will occur to the the
	  finite pixel size and nearest-neighbor interpolation used
	  when calculating the radon transformation.

Extension 11 -- 'radon_maskfrac'

	  Map indicating hte fraction of data in the velocity field
	  that was within the radon aperture but masked (these values
	  can be biased)

Extension 12 -- 'radon_edge'                                   

	  Map indicating whether the radon aperture at a given point
	  overlapped the edge of the velocity field

Extension 13 -- 'trace'   

	  The results of the radon transform trace along the ridge
	  where the transform is minimized as a function of rho.  The
	  contents are:

   	  RHO -- distance from center           

   	  THETA -- angle where RT is minimized (basically position angle but
   	  offset by 90 degrees)

   	  ETHETA -- uncertainty on theta         

   	  STATUS -- status returned by MPFIT (see their documentation
   	  for further details). Typically we want to avoid anything <
   	  1.

   	  SLICE_LENGTH -- fraction of data at fixed rho which was
   	  usable.  The trace can perform poorly when slice_length gets
   	  small, although this depends on exactly which data are
   	  masked (see "bias_flag" for an attempt at a more robust way
   	  of flagging theta measurements where missing data becomes a
   	  problem).

   	  BIAS_FLAG -- indicates that the estimate of theta may be
   	  biased by missing data. For each radius, rho[i], I check to
   	  see if the value of theta at the previous radius, rho[i-1],
   	  falls inside a masked region at the current rho[i]. If it
   	  does, then it is flagged

	  When I plot these traces, I typically use the selector:

	  sel=where(status gt 0 and 1-bias_flag)