sp_validation.cosmo_val

Cosmology validation package.

CosmologyValidation is the public entry point. cs_plots is re-exported so the cosmo_val.cs_plots attribute path (the cs_util plotting alias used at each call site) keeps resolving for callers that import the package as a module handle.

class CosmologyValidation(versions, catalog_config='./cat_config.yaml', output_dir=None, rho_tau_method='lsq', cov_estimate_method='th', compute_cov_rho=True, n_cov=100, theta_min=0.1, theta_max=250, nbins=20, var_method='jackknife', npatch=20, quantile=0.1587, theta_min_plot=0.08, theta_max_plot=250, ylim_alpha=[-0.005, 0.05], ylim_xi_sys_ratio=[-0.02, 0.5], nside=1024, nside_mask=8192, binning='powspace', power=0.5, n_ell_bins=32, ell_step=10, pol_factor=True, cell_method='map', noise_bias_method='analytic', fiducial_input_inka='coupled', nrandom_cell=10, cell_seed=8192, path_onecovariance=None, cosmo_params=None, blind=None)[source]

Bases: CosebisMixin, PureEBMixin, RealSpaceMixin, PSFSystematicsMixin, CatalogCharacterizationMixin, PseudoClMixin

Framework for cosmic shear validation and systematics analysis.

Handles two-point correlation function measurements, PSF systematics (rho/tau), pseudo-C_ell analysis, and covariance estimation for weak lensing surveys. Supports multiple catalog versions with automatic leakage-corrected variants.

Parameters:

  • versions (list of str) – Catalog version identifiers to analyze. Appending ‘_leak_corr’ to a base version creates a virtual catalog using leakage-corrected ellipticity columns (e1_col_corrected/e2_col_corrected) from the base version configuration.

  • catalog_config (str, default './cat_config.yaml') – Path to catalog configuration YAML defining survey metadata, file paths, and analysis settings for each version.

  • output_dir (str, optional) – Override for output directory. If None, falls back to the COSMO_VAL environment variable, then to the catalog config’s paths.output.

  • rho_tau_method ({'lsq', 'mcmc'}, default 'lsq') – Fitting method for PSF leakage systematics parameters.

  • cov_estimate_method ({'th', 'jk'}, default 'th') – Covariance estimation: ‘th’ for semi-analytic theory, ‘jk’ for jackknife.

  • compute_cov_rho (bool, default True) – Whether to compute covariance for rho statistics during PSF analysis.

  • n_cov (int, default 100) – Number of realizations for covariance estimation when using theory method.

  • theta_min (float, default 0.1) – Minimum angular separation in arcminutes for correlation function binning.

  • theta_max (float, default 250) – Maximum angular separation in arcminutes for correlation function binning.

  • nbins (int, default 20) – Number of angular bins for TreeCorr real-space correlation functions.

  • var_method ({'jackknife', 'sample', 'bootstrap', 'marked_bootstrap'}, default 'jackknife') – TreeCorr variance estimation method.

  • npatch (int, default 20) – Number of spatial patches for jackknife variance estimation.

  • quantile (float, default 0.1587) – Quantile for uncertainty bands in plots (default: 1-sigma ≈ 0.159).

  • theta_min_plot (float, default 0.08) – Minimum angular scale for plotting (may differ from analysis cut).

  • theta_max_plot (float, default 250) – Maximum angular scale for plotting.

  • ylim_alpha (list of float, default [-0.005, 0.05]) – Y-axis limits for alpha systematic parameter plots.

  • ylim_xi_sys_ratio (list of float, default [-0.02, 0.5]) – Y-axis limits for xi systematics ratio plots.

  • nside (int, default 1024) – HEALPix resolution for pseudo-C_ell analysis and area computation.

  • binning ({'powspace', 'linspace', 'logspace'}, default 'powspace') – Ell binning scheme for pseudo-C_ell (powspace = ell^power spacing).

  • power (float, default 0.5) – Exponent for power-law binning when binning=’powspace’.

  • n_ell_bins (int, default 32) – Number of ell bins for pseudo-C_ell analysis (used with binning=’powspace’).

  • ell_step (int, default 10) – Bin width in ell for linear binning (used with binning=’linear’).

  • pol_factor (bool, default True) – Apply polarization correction factor in pseudo-C_ell calculations.

  • nrandom_cell (int, default 10) – Number of random realizations for C_ell error estimation.

  • cell_seed (int, default 8192) – Seed for the random-rotation noise realizations in the pseudo-C_ell noise debiasing, making those realizations reproducible run-to-run.

  • cosmo_params (dict, optional) – Cosmological parameters to pass to get_cosmo(). If None, uses Planck 2018.

versions

Validated catalog versions after processing _leak_corr variants.

Type:

list of str

cc

Loaded catalog configuration with resolved absolute paths.

Type:

dict

catalog_config_path

Resolved path to the catalog configuration file.

Type:

Path

treecorr_config

Configuration dictionary passed to TreeCorr correlation objects.

Type:

dict

cosmo

Cosmology object for theory predictions.

Type:

pyccl.Cosmology

Notes

  • Path resolution: Relative paths in catalog config are resolved using each version’s ‘subdir’ field as the base directory.

  • Virtual _leak_corr versions: These create deep copies of the base version config, swapping e1_col/e2_col with e1_col_corrected/e2_col_corrected.

  • TreeCorr cross_patch_weight: Automatically set to ‘match’ for jackknife, ‘simple’ otherwise, following TreeCorr best practices.

static _split_seed_variant(version)[source]

Return the base version and seed label if version encodes a seed.

static _materialize_seed_path(base_cfg, seed_label, version, base_version, catalog_config)[source]

Render the seed-specific shear path using Python string formatting.

_output_path(*parts)[source]

Absolute path under the catalog config’s output directory.

Joins *parts onto self.cc["paths"]["output"] and absolutises the result, mirroring the os.path.abspath(f"{output}/...") pattern used throughout the mixins. A single parts string may contain / separators.

get_redshift(version)[source]

Load redshift distribution for a catalog version.

Parameters:

version (str) – Catalog version identifier

Returns:

  • z (ndarray) – Redshift values

  • nz (ndarray) – n(z) probability density

Notes

If self.blind is set, the redshift path is modified to use the specified blind (A, B, or C) by replacing the blind suffix in the configured path.

_write_catalog_config()[source]
color_reset()[source]
_cprint(color, msg, end='\n')[source]

Print msg in color, then restore the default foreground.

print_blue(msg, end='\n')[source]
print_start(msg, end='\n')[source]
print_done(msg)[source]
print_magenta(msg)[source]
print_green(msg)[source]
print_cyan(msg)[source]
init_results(objectwise=False)[source]
property results
property results_objectwise
basename(version, treecorr_config=None, npatch=None)[source]
_binning(min_sep=None, max_sep=None, nbins=None, **extra)[source]

treecorr_config with min_sep/max_sep/nbins overridden.

None falls back to the instance’s treecorr_config value for that key; any further keys in extra override on top.

_read_shear_cols(ver, *keys)[source]

Read shear-catalog columns by their config-key names.

Each key in *keys (e.g. "e1_col", "w_col") is resolved to a column name via self.cc[ver]["shear"][key] and indexed out of self.results[ver].dat_shear. Must be called inside a self.results[ver].temporarily_read_data() context, since it touches dat_shear directly.

Returns one array per key (a bare array, not a 1-tuple, when a single key is requested).

_calibrated_g(ver)[source]

Calibrated shear components (g1, g2) for a catalog version.

Applies additive-bias subtraction and the multiplicative response: g = (e − c) / R. For DES the response is the catalog-averaged per-component R11/R22 (column names in the config); for every other version it is the scalar R from the config. Used identically by calculate_2pcf() and calculate_aperture_mass_dispersion().

Must be called inside a self.results[ver].temporarily_read_data() context, since it reads dat_shear columns.

property colors
summarize_bmodes(fiducial_scale_cut=(12, 83), versions=None)[source]

Print and return B-mode PTE summary across all statistics.

Collects PTEs from pure E/B, COSEBIs, and pseudo-Cl at the specified fiducial scale cut. Statistics that haven’t been computed show ‘–‘.

Parameters:

  • fiducial_scale_cut (tuple, optional) – (min_theta, max_theta) for extracting PTEs (default: (12, 83)).

  • versions (list, optional) – Versions to summarize. Uses self.versions if None.

Returns:

{version: {statistic: pte_value, ...}, ...}

Return type:

dict

Submodules