aces.holography.clean_holography

Clean holography data

Module Contents

Classes

BeamModel

Class - to be improved 1. hold data in hdf5 file, not pickle.

Functions

tft(t)
report_flags(flags, tag)
robust_poly_fit(spec[, poly_ord, sigma_lim, max_iter, ...])	Fit a polynomial to data with outliers
fit_spec(spec[, poly_ord, sigma_lim, max_iter])	Fit a polynomial to data with outliers
fit_all_spectra(hm[, poly_ord])	Determines the best fitting models and generates that model for a given mean cube
add_metadata_to_header(→ None)	Common function to add metadata from a MapSet into the FITS header. Some
write_to_fits(data_clean, mapset, interp_factor, ...)	Write data cube to FITS:
write_tt_to_fits(beamtt, mapset, fref, bw, dxy, output_dir)	Write Taylor Terms to FITS:
interp2D_cmplx(z, pixpos[, k])	param z: (ndarray) Complex cube of shape Nt,Na,Nb,Np,Nf,nx,ny
normalise(v, bpos, refant)	Given an array of raw holography visibilities v, normalise to give
volt_to_power(v)	Given a Jones matrix v, return the corresponding brightness
feed_to_stokes(f)	Given a matrix (voltage or power) in the feed frame, return the corresponding
raw_to_stokes(raw_file_name, refant)	Reads holography data in hdf5 MapSet format. Performs operations on the
get_quality_flags(mapset[, amp_max])	Take MapSet object holding Holography power patterns in the IQUV
form_mean(mapset)	Take MapSet object holding Holography power patterns in the IQUV
save_mean(mean_cube, mapset, out_name)	Takes the array-wide mean cube and writes it to an hdf5 file in the BeamSet format
centre_beams(mean_cube, mapset[, kxy, nxy])	Takes the mean holography cube (averaged over antennas) and
tt(fstart, beam_cube, nterm)	For correction of continuum images we turn the frequency axis
get_beampos_errs(mapset, bmodel)	TODO
get_shift(mapset, ref_im, target_im[, frac])	Compute the shift of target_im relative to ref_im.
centre_mass(im, xaxis)
get_beampos_errs_from_fits(fits_header, fits_data)
shifts_to_mapset(mapset, shifts, resids, amps)	Package the beam shift data into a Mapset object
save_shifts(beamset, holo_dir)	Save a BeamSet object holding beam errors to an hdf5 file:
design_mat(x, op)
pfit(x, y, op)
get_posfit_flags(resid, amps, res_limits, amp_limits)	TODO
set_resid_lim(resids[, order, iqr_fac])
get_shift_flags(shifts, shift_fit_flags)	TODO
main(sbid[, holo_dir, remake_stokes, max_order, ...])	Main script
cli()

Attributes

log

aces.holography.clean_holography.log

aces.holography.clean_holography.tft(t)

aces.holography.clean_holography.report_flags(flags, tag)

aces.holography.clean_holography.robust_poly_fit(spec, poly_ord=2, sigma_lim=0.8, max_iter=20, low_fact=10.0)

Fit a polynomial to data with outliers

The test for the presence of outliers is a comparison between standard deviation (std) and the inter-quartile range (IQR). The IQR is a robust measure of the distribution width, and for a gaussian distribution std/IQR = approx 0.6. The value of std/IQR at which to cease the search for outliers is sigma_lim.

Parameters:

• spec (np.ndarray) – Spectrum to fit

• poly_ord (int, optional) – Order of polynomial, defaults to 5

• sigma_lim (float, optional) – Sigma limit, defaults to 0.8

• max_iter (int, optional) – Maxium interations, defaults to 20

Returns:

TODO

Return type:

TODO

aces.holography.clean_holography.fit_spec(spec, poly_ord=5, sigma_lim=3.0, max_iter=50)

Fit a polynomial to data with outliers

Parameters:

• spec (np.ndarray) – Spectrum to fit

• poly_ord (int, optional) – Order of polynomial, defaults to 5

• sigma_lim (float, optional) – Sigma limit, defaults to 3.0

• max_iter (int, optional) – Maxium interations, defaults to 50

Returns:

TODO

Return type:

TODO

aces.holography.clean_holography.fit_all_spectra(hm, poly_ord=3)

Determines the best fitting models and generates that model for a given mean cube Args:

hm ([array]): data cube (masked array) poly_ord (int):

Returns:

[array]: Returns the best fitting model over all pixels and the associated frequencies.

aces.holography.clean_holography.add_metadata_to_header(header: astropy.io.fits.header.Header, metadata: dict[str, Any]) → None

Common function to add metadata from a MapSet into the FITS header. Some items from the MapSet metadata structure are recorded as a list of values. Here these are handled by taking the first element of the list.

Parameters:

• header (fits.header.Header) – The HDU header that is being added to

• metadata (dict) – The key-value pair of items to add to header

aces.holography.clean_holography.write_to_fits(data_clean: numpy.ndarray, mapset: aces.beamset.mapset.MapSet, interp_factor: float, output_dir: str)

Write data cube to FITS: :param data_clean: Cleaned array (with Stokes axis). :type data_clean: np.ndarray :param mapset: Input holography dataset :type mapset: MapSet :param interp_factor: Intperpolation grid used in deg :type interp_factor: float :param output_dir: Directory to save file to :type output_dir: str

aces.holography.clean_holography.write_tt_to_fits(beamtt: numpy.ndarray, mapset: aces.beamset.mapset.MapSet, fref: float, bw: float, dxy: float, output_dir: float)

Write Taylor Terms to FITS: :param beamtt: Taylor term beams :type beamtt: np.ndarray :param mapset: Input holography dataset :type mapset: MapSet :param fref: Reference frequency for Taylor terms :type fref: float :param bw; The observing bandwidth that is recorded to the header :type bw: float :param dxy: Interpolation grid in deg :type dxy: float :param output_dir: Directory to save file to :type output_dir: str

aces.holography.clean_holography.interp2D_cmplx(z, pixpos, k=2)

Parameters:

• z – (ndarray) Complex cube of shape Nt,Na,Nb,Np,Nf,nx,ny

• pixpos – (ndarray) Array of shape (nbeams,2) giving beam positions in pixels

• k – (int) Order of spline interpolation used.

:return interp (ndarray) Interpolated values at beam positions - shape Nt,Na,Nb,Np,Nf

aces.holography.clean_holography.normalise(v, bpos, refant)

Given an array of raw holography visibilities v, normalise to give the response of each antenna relative to its response to a point source at beam centre. Each map value at beam centre is determined by interpolation: vxx0, vyy0. The normalised quantities are computed as Nxx = vxx/vxx0, Nxy = vxy/vxx0, Nyx = vyx/vyy0, Nyy = vyy/vyy0 :param v:(numpy.ndarray) visibilities :param bpos: (np.ndarray) beam position in pixels (nb,2) :param refant: (int) Reference antenna number (0 based) :return N: (numpy.ndarray) normalised visibilities

NOTE: This scheme works provided the reference antenna has a lower index than all target antennas For cases where the correlation is Target * Reference we need: scale_xy = scale_yy scale_yx = scale_xx

aces.holography.clean_holography.volt_to_power(v)

Given a Jones matrix v, return the corresponding brightness matrix b as b = v * v(h) where (h) is the hermitian transpose. In both v and b, the polarisation axis is the 4th of 7 axes: sh = (nt,ne,nb,4,nf,nx,ny) :param v: (numpy.ndarray) voltage vector of shape sh :return b: (numpy.ndarray) brightness vector of shape sh

aces.holography.clean_holography.feed_to_stokes(f)

Given a matrix (voltage or power) in the feed frame, return the corresponding Stokes matrix. Do this for ASKAP, assuming the measurements were made with the antenna rotated +45 degrees. In both f and s, the polarisation axis is the 4th of 7 axes: sh = (nt,ne,nb,4,nf,nx,ny) :param f: (numpy.ndarray) vector in feed frame: XX,XY,YX,YY :return s: (numpy.ndarray) vector in Stokes frame: I,Q,U,V

aces.holography.clean_holography.raw_to_stokes(raw_file_name, refant)

Reads holography data in hdf5 MapSet format. Performs operations on the data hyper-cube: * transforms to axes aligned with celestial coords * determines map values at beam positions in Vxx, Vyy * derives normalising factors for Vxx, Vxy, Vyx, Vyy visibilities * normalises the complex data * determines brightness vector XX, XY, YX, YY * converts to Stokes I Q U V * saves both Stokes vector and scale values in hdf5 MapSet format

Parameters:

• raw_file_name – (str) Name of hdf5 holography data file

• refant – (int) Reference antenna

:return stokes_obj (MapSet) Holography maps, normalised and in Stokes brightness I Q U V :return centre_obj (MapSet) Normalising scales - visibilities at beam positions (interpolated).

aces.holography.clean_holography.get_quality_flags(mapset, amp_max=1.4)

Take MapSet object holding Holography power patterns in the IQUV Stokes parameters; look for and flag image planes whose maximum stokes I value significantly exceeds the expected value of 2.0.

Typically, the large value flags arise from planes that were effected by RFI and produce large values at random locations away from the beam peak: :param mapset: MapSet object holding cube of beam images :type mapset: aces.beamset.mapset.MapSet :param amp_max: Maximum allowed amplitude :type amp_max: float, optional :return: result - a dict holding flags and correspondingly masked data :rtype: dict

aces.holography.clean_holography.form_mean(mapset)

Take MapSet object holding Holography power patterns in the IQUV Stokes parameters; Compute the mean over the antenna axis for all beams, polarisations and spectral points using the MapSet flags to include only image planes judged to be of high quality: :param mapset: Object holding cube of beam images :type mapset: aces.beamset.mapset.MapSet :return: mean_cube: Mean cube of shape [nB, nP, nF, nx, ny] :rtype: numpy.ndarray

aces.holography.clean_holography.save_mean(mean_cube, mapset, out_name)

Takes the array-wide mean cube and writes it to an hdf5 file in the BeamSet format :param mean_cube: Data cube of shape (nb, npol, nf, nx, ny) :param mapset: BeamSet object from full Stokes cube; used to derive new metadata for output :param out_name: Name of output file :type out_name: str :return:

aces.holography.clean_holography.centre_beams(mean_cube, mapset, kxy=5, nxy=101)

Takes the mean holography cube (averaged over antennas) and interpolates each beam onto a grid centred on its nominal position and with the given grid spacing

Parameters:

• mean_cube (np.ndarray) – shape Nb,Np,Nf,Nx,Ny

• mapset ([type]) – parent object - for metadata

• kxy (int, optional) – Degrees of bivariate spline used in RecBivariateSpline, defaults to 5

• nxy (int, optional) – Size of output grid (must be odd integer)., defaults to 101

Returns:

TODO

Return type:

TODO

aces.holography.clean_holography.tt(fstart, beam_cube, nterm)

For correction of continuum images we turn the frequency axis into a taylor term axis.

• Note this requires picking a centre frequency f0, for which the

Taylor coefficients are computed

Parameters:

• fstart (float) – Starting frequency in MHz

• beam_cube (np.ndarray) – Beam maps, averaged over antennas

• nterm (int) – Number of Taylor terms

Returns:

TODO

Return type:

TODO

class aces.holography.clean_holography.BeamModel(bs_obj)

Bases: object

Class - to be improved 1. hold data in hdf5 file, not pickle.

get(pixoff, freq)

TODO

Parameters:

• pixoff (TODO) – TODO

• freq (TODO) – TODO

Returns:

TODO

Return type:

TODO

static _radial_gau_gen()

Parameters used here were derived from a ‘good set’ of mid-band beams (SBID=16881), the Gaussian parameter (0.6945) minimises the difference between model gaussian and measured average beam shape at 1438.5 MHz.

aces.holography.clean_holography.get_beampos_errs(mapset, bmodel)

TODO

Parameters:

• mapset (TODO) – TODO

• bmodel (TODO) – TODO

Returns:

TODO

Return type:

TODO

aces.holography.clean_holography.get_shift(mapset, ref_im, target_im, frac=0.25)

Compute the shift of target_im relative to ref_im. Use Fourier shift theorem The sense of the returned shifts: If the returned value shift = (dl, dm), and if the position of reference and target beams are (lr, mr) and (lt, mt) respectively, then dl > 0.0 indicates lt > lr and if dm > 0.0 then tm > rm. Coordinate $l$ increases towards the west; $m$ increases towards the north.

Parameters:

• mapset ([type]) – Primary mapset

• ref_im (np.ndarray) – Reference image ndim=2

• target_im (np.ndarray) – Target image

• frac (float, optional) – Mask transform values < ampMax * frac in phase fit, defaults to 0.25

Returns:

TODO

Return type:

TODO

aces.holography.clean_holography.centre_mass(im, xaxis)

aces.holography.clean_holography.get_beampos_errs_from_fits(fits_header, fits_data)

aces.holography.clean_holography.shifts_to_mapset(mapset, shifts, resids, amps)

Package the beam shift data into a Mapset object :param mapset: Input mapset object, holdng the holography maps :param shifts: Stokes I beam position offsets :param resids: Residuals to offset fit :param amps: :return: Mapset object holdong shift data

aces.holography.clean_holography.save_shifts(beamset, holo_dir)

Save a BeamSet object holding beam errors to an hdf5 file: :param beamset: BeamSet object holding beam pos errors :type beamset: BeamSet :param holo_dir: Directory to receive hdf5 file :type holo_dir: str

aces.holography.clean_holography.design_mat(x, op)

aces.holography.clean_holography.pfit(x, y, op)

aces.holography.clean_holography.get_posfit_flags(resid, amps, res_limits, amp_limits)

TODO

Parameters:

• resid (TODO) – TODO

• amps (TODO) – TODO

Returns:

n_flags New flags of shape to match obj.flags; True for bad data

Return type:

TODO

aces.holography.clean_holography.set_resid_lim(resids, order=2, iqr_fac=2.0)

aces.holography.clean_holography.get_shift_flags(shifts, shift_fit_flags)

TODO

Parameters:

• shifts –

• shift_fit_flags –

Returns:

n_flags New flags of shape to match obj.flags; True for bad data

Return type:

TODO

aces.holography.clean_holography.main(sbid, holo_dir='.', remake_stokes=False, max_order=5, param='BIC', max_iter=50, snr_limit=3)

Main script :param sbid: SBID :type sbid: int :param holo_dir: Directory containing holo data, defaults to ‘.’ :type holo_dir: str, optional :param remake_stokes: Remake Stokes parameters, defaults to False :type remake_stokes: bool, optional :param max_order: Maximum polynomial to fit, defaults to 5 :type max_order: int, optional :param param: Criterion to evaluate fit, defaults to ‘BIC’ :type param: str, optional :param max_iter: Maximum interations for fitting, defaults to 50 :type max_iter: int, optional :param snr_limit: SNR limit for fitting, defaults to 3 :type snr_limit: int, optional

aces.holography.clean_holography.cli()