keppca: perform principal component analysis upon a target pixel file

pyke.keppca.keppca(infile, outfile=None, maskfile='ALL', components='1-3', plotpca=False, nmaps=10, overwrite=False, verbose=False, logfile='keppca.log')[source]

keppca – Perform principal component analysis upon a target pixel file

keppca provides a method to mitigate for motion-derived systematic artifacts via Principle Component Analysis (PCA). This method was demonstrated on Kepler light curves by Harrison et al. (2012). It provides an alternative to cotrending data using basis vectors (kepcotrend) and correlating aperture photometry struture with time-series centroid measurements (kepsff). PCA will perhaps become a more widespread tool in the K2 era where the magnitde of target motion across the detector over a Kepler quarter is experienced by a K2 target over just 6-hours during its regular sequence of thruster firings that counteract boresight roll motion Pixel-level PCA employs only those pixels collected around a specific target and separates photometric trends common to all pixels from trends localized to individual targets or pixels in a series of principal component curves.

The user has the option to choose the specific set of pixels to sample in this analysis. Principal components are plotted by the tool and written out to an output FITS file in an output extension called PRINCIPAL_COMPONENTS. The extension contains a 2D table with one row per timestamp recorded in the input file and one column for every principal component. Summing all principal components together will reconstruct a normalized version of the summed pixel within the chosen aperture. The user also has the choice of which principal components to optimally-subtract from the aperture-derived light curve in order to remove motion systematics from the time-series data. The aperture light curve and the corrected light curve are written to the LIGHTCURVE extension of the output file. The first populates the SAP_FLUX data column and the second is written to a column called PCA_FLUX. This output file can be used as input for other PyKE tasks and can be e.g. inspected using kepdraw.


infile : str

The name of a standard format FITS file containing Kepler or K2 target pixels within the first data extension.

outfile : str

Filename for the output light curves and principal components. This product will be written to the same FITS format as archived light curves. Aperture photometry will be stored in the SAP_FLUX column of the first FITS extension called LIGHTCURVE. A version of this light curve with principal components subtracted is stored in column PCA_FLUX and a normalized version is stored in PCA_FLUX_NRM. The individual principal components are stored within a new FITS extension called PRINCIPAL_COMPONENTS.

maskfile : str

This string can be one of three options:

  • ‘ALL’ tells the task to calculate principal components from all pixels within the pixel mask stored in the input file.
  • ‘APER’ tells the task to calculate principal components from only the pixels within the photometric aperture stored in the input file (e.g. only those pixels summed by the Kepler pipeline to produce the light curve archived at MAST (note that no such light curves are currently being created for the K2 mission)
  • A filename describing the desired photometric aperture. Such a file can be constructed using the kepmask or kepffi tools, or can be created manually using the format described in the documentation for those tools. Note that if an aperture provided is not stricly rectangular, keppca will increase the size of the aperture so that it defines the smallest possible rectangle that contains all of the specified pixels.

components : str

A list of the principal components to subtract from the aperture light curve. The strings ‘1 2 3 4 5’, 1,‘2,3,4,5’ and ‘1,2,3-5’ yield the same result.

plotpca : bool

If True, keppca will produce plots containing individual principal components, correlation maps and light curves, both aperture and PCA-corrected versions. The will be stored as hardcopies in PNG format.

nmaps : int

The number of correlation maps and principal components to plot as output. This can be any positive integer up to the number of pixels within the mask, although note that many hundreds of plots will likely become prohibitive and is unlikely to be informative.

overwrite : bool

Overwrite the output file?

verbose : bool

Print informative messages and warnings to the shell and logfile?

logfile : str

Name of the logfile containing error and warning message


$ keppca ktwo202073445-c00_lpd-targ.fits.gz --plotpca