Photo-z

These stages deal with photo-z PDF training and estimation

class txpipe.photoz_stack.TXPhotozStack(args, comm=None, aliases=None)[source]

Naive stacker using QP.

Can only cope with hist or interp PDF types. Ideally this should be replaced by a RAIL stage.

Inputs:

  • photoz_pdfs: QPPDFFile

  • tomography_catalog: TomographyCatalog

  • weights_catalog: HDFFile

Outputs:

  • photoz_stack: QPNOfZFile

Parallel: Yes - MPI

Configuration
  • chunk_rows: (int) Default=5000.
  • tomo_name: (str) Default=source.
  • weight_col: (str) Default=shear/00/weight.
  • zmax: (float) Default=0.0.
  • nz: (int) Default=0.
class txpipe.photoz_stack.TXTruePhotozStack(args, comm=None, aliases=None)[source]

Make an ideal true source n(z) using true redshifts

Fake an n(z) by histogramming the true redshift values for each object.

Inputs:

  • tomography_catalog: TomographyCatalog

  • catalog: HDFFile

  • weights_catalog: HDFFile

Outputs:

  • photoz_stack: QPNOfZFile

Parallel: Yes - MPI

Configuration
  • chunk_rows: (int) Default=5000.
  • zmax: (float) Required.
  • nz: (int) Required.
  • weight_col: (str) Default=weight.
  • redshift_group: (str) Required.
  • redshift_col: (str) Default=redshift_true.