StepCombine

class sofia_redux.instruments.hawc.steps.stepcombine.StepCombine[source]

Bases: StepParent

Combine time series data for R+T and R-T flux samples.

This step averages all chop-subtracted samples for each nod and HWP setting, for the R+T and R-T images. Outliers are identified via iterative sigma-clipping (Chauvenet’s criterion). Errors are propagated from the input variance images or, optionally, reported as the standard deviation across the time samples.

After this step, R-T images are propagated for polarimetry data only (NHWP > 1).

This step should be run after the sofia_redux.instruments.hawc.steps.StepSplit pipeline step. It requires the following extensions: for each HWP angle M and Nod N there should be six images: DATA R-T HWP M NOD N, DATA R+T HWP M NOD N, VAR R-T HWP M, NOD N, VAR R+T HWP M NOD N, VAR R HWP M NOD N, VAR T HWP M NOD N. In addition, there must be a table containing the rows corresponding to a given HWP and Nod, named TABLE HWP M NOD N. Finally, there should be a single bad pixel mask image.

The output extensions are the same as the input extensions, except that VAR R+T and VAR R-T are replaced with ERROR R+T and ERROR R-T extensions.

Methods Summary

comb_table(table, newmask)

Average all rows for a table.

run()

Run the data reduction algorithm.

setup()

Set parameters and metadata for the pipeline step.

Methods Documentation

comb_table(table, newmask)[source]

Average all rows for a table.

Parameters:
tablefits.FITS_rec

The table to average.

newmaskarray-like of bool

Table rows to combine.

Returns:
BinTableHDU

The averaged table.

run()[source]

Run the data reduction algorithm.

Because this step is single-in, single-out (SISO), self.datain must be a DataFits object. The output is also a DataFits object, stored in self.dataout.

This step combines the chop cycles for each HWP angle and Nod separately (and each pixel as well). This happens first for the R-T data, as follows:

  1. For each pixel in R-T, compute the mean value and standard deviation.

  2. Reject any chop cycles more than sigma away from the mean.

  3. Repeat 1-2 until no more chop cycles are removed.

Any masked pixels from R-T deglitching are also masked in R+T. Additional deglitching in R+T follows the same outlier rejection as for R-T, with the sigma cutoff specified by the sum_sigma parameter.

setup()[source]

Set parameters and metadata for the pipeline step.

Output files have PRODTYPE = ‘combine’, and are named with the step abbreviation ‘CMB’.

Parameters defined for this step are:

sigmafloat

Reject outliers more than this many sigma from the mean.

sum_sigmafloat

Reject additional R+T outliers more than this many sigma from the mean.

use_errorbool

Set to True to use the standard deviation across the time samples as the output error, rather than propagating input variances.