covariance module

Functions to calculate 3x2pt Gaussian covariance.

class covariance.Field(field_type, zbin)

Bases: tuple

property field_type

Alias for field number 0

property zbin

Alias for field number 1

class covariance.FieldType(value)

Bases: enum.Enum

The three types of fields: position, shear E and shear B.

POSITION = 1
SHEAR_B = 3
SHEAR_E = 2
class covariance.PowerSpectrum(field_1, field_2)

Bases: tuple

property field_1

Alias for field number 0

property field_2

Alias for field number 1

covariance.bin_combine_cov(n_zbin, lmin_in, lmin_out, lmax, n_bp_min, n_bp_max, input_filemask, save_filemask)

Loop over numbers of bandpowers, and for each one, bin all blocks of unbinned covariance matrix and combine into a single matrix, saved to disk.

Parameters

  • n_zbin (int) – Number of redshift bins, assuming 1 position field and 1 shear field per redshift bin.

  • lmin_in (int) – Minimum l in the unbinned covariance.

  • lmin_out (int) – Minimum l to include in the binned covariance.

  • lmax (int) – Maximum l in the unbinned and binned covariance.

  • n_bp_min (int) – Minimum number of bandpowers to loop over (inclusive).

  • n_bp_max (int) – Maximum number of bandpowers to loop over (inclusive).

  • input_filemask (str) – Path to unbinned covariance blocks output by get_3x2pt_cov, with {spec1_idx} and {spec2_idx} placeholders.

  • save_filemask (str) – Path to save each binned covariance matrix, with {n_bp} placeholder.

covariance.get_3x2pt_cov(n_zbin, pos_pos_filemask, pos_she_filemask, she_she_filemask, lmax_in, lmin_in, lmax_out, lmin_out, pos_nl_path, she_nl_path, noise_lmin, mask_path, nside, save_filemask)

Calculate 3x2pt Gaussian covariance using NaMaster, saving each block separately to disk.

Parameters

  • n_zbin (int) – Number of redshift bins, assuming 1 position field and 1 shear field per redshift bin.

  • pos_pos_filemask (str) – Path to text file containing a position-position power spectrum with {hi_zbin} and {lo_zbin} placeholders.

  • pos_she_filemask (str) – Path to text file containing a position-shear power spectrum with {pos_zbin} and {she_zbin} placeholders.

  • she_she_filemask (str) – Path to text file containing a shear-shear power spectrum with {hi_zbin} and {lo_zbin} placeholders.

  • lmax_in (int) – Maximum l to including in mixing.

  • lmin_in (int) – Minimum l in input power spectra.

  • lmax_out (int) – Maximum l to include in covariance.

  • lmin_out (int) – Minimum l to include in covariance.

  • pos_nl_path (str) – Path to position noise power spectrum.

  • she_nl_path (str) – Path to shear noise power spectrum.

  • noise_lmin (int) – Minimum l in noise power spectra.

  • mask_path (str) – Path to mask FITS file. If None, full sky is assumed, in which case the covariance will be diagonal.

  • nside (int) – HEALPix resolution nside parameter.

  • save_filemask (str) – Path to save each covariance block to disk, with {spec1_idx} and {spec2_idx} placeholders.

covariance.load_cls(signal_paths, noise_paths, lmax_in, lmax_out, signal_lmin, noise_lmin)

Load a list of Cls with appropriate noise, given lists of path to signal and noise Cls. If any entry in either list is None, then zeros are used. Signal and noise are padded with zeros below signal_lmin and noise_lmin and above lmax_in (if less than lmax_out).

Parameters

  • signal_paths (list) – List of paths to signal Cls. If any entry is None, it is taken to be zero.

  • noise_paths (list) – List of paths to noise Cls. If any entry is None, it is taken to be zero.

  • lmax_in (int) – Maximum l to load. If this is less than lmax_out then padded with zeros above this l.

  • lmax_out (int) – Maximum l to return.

  • signal_lmin (int) – First l of signal power spectra. Signal Cls will be padded with zeros below this.

  • noise_lmin (int) – First l of noise power spectra. Noise Cls will be padded with zeros below this.

Returns

List of numpy arrays, each of which is a signal + noise power spectrum in the supplied order.

Return type

list

covariance.spectrum_from_fields(field_a, field_b)

Returns a PowerSpectrum object corresponding to the supplied fields, in a consistent order, which is lower redshift bin first, and if the bins are equal then position first.

(Note that this is not the same convention as used when loading from CosmoSIS output.)

Parameters

  • field_a (Field) – One field in the spectrum.

  • field_b (Field) – The other field in the spectrum.

Returns

Power spectrum corresponding to the two fields.

Return type

PowerSpectrum

covariance.spin_from_field(field)

Returns the spin of a field.

Parameters

field (Field) – Field to get the spin of.

Returns

The spin of the field.

Return type

int

covariance.spins_from_spectrum(spec)

Returns the two spins corresponding to the two fields of the supplied power spectrum.

Parameters

spec (PowerSpectrum) – Power spectrum to get the spins of.

Returns

The spins of the two fields of the power spectrum.

Return type

(int, int)

covariance.workspace_from_spins(spin_a, spin_b, workspace_spin00, workspace_spin02, workspace_spin22)

Returns the appropriate NmtWorkspace object for the two supplied spins.

Parameters

  • spin_a (int) – Spin of one field.

  • spin_2 (int) – Spin of the other field.

  • workspace_spin00 (NmtWorkspace) – Workspace for two spin-0 fields.

  • workspace_spin02 (NmtWorkspace) – Workspace for one spin-0 and one spin-2 field (in either order).

  • workspace_spin22 (NmtWorkspace) – Workspace for two spin-2 fields.

Returns

The workspace object corresponding to the two supplied spins.

Return type

NmtWorkspace