Weights

These stages deal with weighting the lens sample

class txpipe.lssweights.TXLSSWeights(args, comm=None, aliases=None)[source]

Base class for LSS systematic weights

Not to be run directly.

This is an abstract base class, which other subclasses inherit from to use the same basic structure, which is:

  • load and process sytematic (survey property) maps

  • compute 1d density correlations+covariance

  • compute weights with a regression method

Inputs:

  • binned_lens_catalog_unweighted: TomographyCatalog

  • lens_tomography_catalog_unweighted: TomographyCatalog

  • mask: MapsFile

Outputs:

  • lss_weight_summary: FileCollection

  • lss_weight_maps: MapsFile

  • binned_lens_catalog: HDFFile

  • lens_tomography_catalog: HDFFile

Parallel: No - Serial

Configuration
  • supreme_path_root: (str) Default=.
  • nbin: (int) Default=20.
  • outlier_fraction: (float) Default=0.01.
  • allow_weighted_input: (bool) Default=False.
  • nside_coverage: (int) Default=32.
class txpipe.lssweights.TXLSSWeightsLinBinned(args, comm=None, aliases=None)[source]

Compute LSS systematic weights using simultanious linear regression on the binned 1D correlations

Model: Linear Covariance: Shot noise (for now), no correlation between 1d correlations Fit: Simultaniously fits all sysmaps. By calculating a total weight map and calculating Ndens vs sysmap directly

Inputs:

  • binned_lens_catalog_unweighted: TomographyCatalog

  • lens_tomography_catalog_unweighted: TomographyCatalog

  • mask: MapsFile

  • lens_photoz_stack: HDFFile

  • fiducial_cosmology: FiducialCosmology

Outputs:

  • lss_weight_summary: FileCollection

  • lss_weight_maps: MapsFile

  • binned_lens_catalog: HDFFile

  • lens_tomography_catalog: HDFFile

Parallel: No - Serial

Configuration
  • supreme_path_root: (str) Default=.
  • nbin: (int) Default=20.
  • outlier_fraction: (float) Default=0.05.
  • pvalue_threshold: (float) Default=0.05.
  • equal_area_bins: (bool) Default=True.
  • simple_cov: (bool) Default=False.
  • diag_blocks_only: (bool) Default=True.
  • b0: (list) Default=[1.0].
  • allow_weighted_input: (bool) Default=False.
  • nside_coverage: (int) Default=32.
class txpipe.lssweights.TXLSSWeightsLinPix(args, comm=None, aliases=None)[source]

Compute LSS systematic weights using simultanious linear regression at the pixel level using scikitlearn simple linear regression

Model: Linear 1D Covarinace: Shot noise (for now), no correlation between 1d correlations Pixel Covarinace: Shot noise, no correlation between pixels Fit: Simultaniously fits all sysmaps using sklearn

Inputs:

  • binned_lens_catalog_unweighted: TomographyCatalog

  • lens_tomography_catalog_unweighted: TomographyCatalog

  • mask: MapsFile

  • lens_photoz_stack: HDFFile

  • fiducial_cosmology: FiducialCosmology

Outputs:

  • lss_weight_summary: FileCollection

  • lss_weight_maps: MapsFile

  • binned_lens_catalog: HDFFile

  • lens_tomography_catalog: HDFFile

Parallel: No - Serial

Configuration
  • supreme_path_root: (str) Default=.
  • nbin: (int) Default=20.
  • outlier_fraction: (float) Default=0.05.
  • pvalue_threshold: (float) Default=0.05.
  • equal_area_bins: (bool) Default=True.
  • simple_cov: (bool) Default=False.
  • diag_blocks_only: (bool) Default=True.
  • b0: (list) Default=[1.0].
  • regression_class: (str) Default=LinearRegression.
  • allow_weighted_input: (bool) Default=False.
  • nside_coverage: (int) Default=32.
class txpipe.lssweights.TXLSSWeightsUnit(args, comm=None, aliases=None)[source]

Assign weight=1 to all lens objects

Inputs:

  • binned_lens_catalog_unweighted: TomographyCatalog

  • lens_tomography_catalog_unweighted: TomographyCatalog

  • mask: MapsFile

Outputs:

  • lss_weight_summary: FileCollection

  • lss_weight_maps: MapsFile

  • binned_lens_catalog: HDFFile

  • lens_tomography_catalog: HDFFile

Parallel: No - Serial

Configuration
  • nside_coverage: (int) Default=32.