Phase Gradient Analysis

Functions for analysis of calibration solution phase gradient fits.

ska_sci_ops_data_analysis.phase_gradient.fit_phase_angle(freqs: ndarray, phases: ndarray, guess: float)

Fit provided phases over freqs, returning m and sig_m.

This does the optimisation on complex numbers, not just the phase of those complex numbers, for better results.

This function needs a guess for the gradient within ~0.05 of the truth.

Parameters:

  • freqs (np.ndarray) – array of frequencies (MHz)

  • phases (np.ndarray) – array of measured phases (radians)

  • guess (float) – guess for gradient (rad/MHz)

Returns:

fitted gradient, error in gradient

Return type:

float, float

ska_sci_ops_data_analysis.phase_gradient.fit_soln_set_phase_angles(soln_dir: str, force_zero_phase_offset: bool = False)

Fit the phase angles for all antenna-polarisations in a soln set.

Parameters:

  • soln_dir (str) – path to directory containing calibration solutions

  • force_zero_phase_offset (bool) – If True, fit phases while forcing the y-intercept to be zero. Otherwise, use ska_low_mccs_calibration fitting function. Defaults to False.

Returns:

Arrays of fitted m, sig_m, c, sig_c. m is the gradient, c is the y-intercept. m units are rad/MHz, c units are rad

Return type:

np.ndaray, np.ndarray, np.ndarray, np.ndarray

ska_sci_ops_data_analysis.phase_gradient.get_guess(freqs: ndarray, phases: ndarray, guess0: float)

Derive a guess for the gradient of the phase angle of complex gains.

Done by fitting the phase angle on the first half of the given data.

Parameters:

  • freqs (np.ndarray) – array of frequencies (MHz)

  • phases (np.ndarray) – array of measured phases (rad)

  • guess0 (float) – initial gradient guess (rad/MHz)

Returns:

new guess (rad/MHz)

Return type:

float

ska_sci_ops_data_analysis.phase_gradient.phase_angle_model(f: float, m: float)

Model the phase angle of complex gain solutions.

phi(f) = m*f

Parameters:

  • f (float) – Frequency at which to evaluate the model (arb. units)

  • m (float) – Gradient of the phase angle (rad/freq unit)

Returns:

phi(f) = m*f (rad)

Return type:

float

ska_sci_ops_data_analysis.phase_gradient.plot_mean_fmax_phase_errs(means, ax_hist, ax_pie)

Plot the distributions of solution mean fmax_phase_errs.

Parameters:

  • means (array-like) – array-like of floats

  • ax_hist (matplotlib.pyplot.axis) – axis to plot histogram on

  • ax_pie (matplotlib.pyplot.axis) – axis to plot pie chart on

ska_sci_ops_data_analysis.phase_gradient.soln_set_phase_fit_quality(soln_dir: str, fmax_phase_err_thresh: float = 5.74398001037828, flagged_ant_idxs: list = [], force_zero_phase_offset: bool = False)

Determine the quality of the phase angle fits for a solution set.

Parameters:

  • soln_dir (str) – Path to directory containing calibration solutinos

  • fmax_phase_err_thresh (float) – Threshold for errorbar at maximum frequency in degrees, defaults to FMAX_PHASE_ERR_THRESH

  • flagged_ant_idxs (list) – List of indexes of flagged antennas to exclude, defaults to empty

  • force_zero_phase_offset (bool) – If True, fit phases while forcing the offset to be zero. Otherwise, use ska_low_mccs_calibration fitting function. Defaults to False.

Raises:

ValueError – If soln_dir doesn’t contain station name

Returns:

quality from (“Healthy”, “Faulty”), mean fmax_phase_err, array of all measured fmax_phase_err, array of phase offsets, array of errors in phase_offsets. All arrays in degrees with shape (NPOL, NANT)==(2, 256).

Return type:

str, float, np.ndarray, np.ndarray, np.ndarray

ska_sci_ops_data_analysis.phase_gradient.summary_plot(soln_dir_base: str)

Plot goodness metric for a particular time’s solutions.

Includes all antennas present.

Parameters:

soln_dir_base (str) – Path to directory containing solutions for a particular time step

Returns:

figure and axis of summary plot

Return type:

plt.Figure, plt.Axis