Maps

These stages deal with making different kinds of maps for analysis and plotting.

class txpipe.maps.TXBaseMaps(args, comm=None, aliases=None)[source]

A base class for mapping stages

This is an abstract base class, which other subclasses inherit from to use the same basic structure, which is: - select pixelization - prepare some mapper objects - iterate through selected columns - update each mapper with each chunk - finalize the mappers - save the maps

Inputs: None

Outputs: None

Parallel: Yes - MPI

Configuration
class txpipe.maps.TXSourceMaps(args, comm=None, aliases=None)[source]

Generate source maps directly from binned, calibrated shear catalogs.

This implementation uses DASK, which offers a numpy-like syntax and hides the complicated parallelization details.

Inputs:

  • binned_shear_catalog: HDFFile

Outputs:

  • source_maps: MapsFile

Parallel: Yes - Dask

Configuration
  • block_size: (int) Default=0.
  • chunk_rows: (int) Default=100000.
  • pixelization: (str) Default=healpix.
  • nside: (int) Default=0.
  • sparse: (bool) Default=True.
  • ra_cent: (float) Default=nan.
  • dec_cent: (float) Default=nan.
  • npix_x: (int) Default=-1.
  • npix_y: (int) Default=-1.
  • pixel_size: (float) Default=nan.
class txpipe.maps.TXLensMaps(args, comm=None, aliases=None)[source]

Make tomographic lens number count maps

Uses photometry and lens tomography catalogs.

Density maps are made later once masks are generated.

Inputs:

  • photometry_catalog: HDFFile

  • lens_tomography_catalog: TomographyCatalog

Outputs:

  • lens_maps: MapsFile

Parallel: Yes - Dask

Configuration
  • block_size: (int) Default=0.
  • chunk_rows: (int) Default=100000.
  • pixelization: (str) Default=healpix.
  • nside: (int) Default=0.
  • sparse: (bool) Default=True.
  • ra_cent: (float) Default=nan.
  • dec_cent: (float) Default=nan.
  • npix_x: (int) Default=-1.
  • npix_y: (int) Default=-1.
  • pixel_size: (float) Default=nan.
class txpipe.maps.TXExternalLensMaps(args, comm=None, aliases=None)[source]

Make tomographic lens number count maps from external data

Same as TXLensMaps except it reads from an external lens catalog.

Inputs:

  • lens_catalog: HDFFile

  • lens_tomography_catalog: TomographyCatalog

Outputs:

  • lens_maps: MapsFile

Parallel: Yes - Dask

Configuration
  • block_size: (int) Default=0.
  • chunk_rows: (int) Default=100000.
  • pixelization: (str) Default=healpix.
  • nside: (int) Default=0.
  • sparse: (bool) Default=True.
  • ra_cent: (float) Default=nan.
  • dec_cent: (float) Default=nan.
  • npix_x: (int) Default=-1.
  • npix_y: (int) Default=-1.
  • pixel_size: (float) Default=nan.
class txpipe.maps.TXDensityMaps(args, comm=None, aliases=None)[source]

Convert galaxy count maps to overdensity delta maps

delta = ngal / (weight * <ngal>/<weight>) - 1

This has to be separate from the lens mappers above because it requires the mask, which is created elsewhere (right now in masks.py)

Inputs:

  • lens_maps: MapsFile

  • mask: MapsFile

Outputs:

  • density_maps: MapsFile

Parallel: No - Serial

Configuration
  • mask_threshold: (float) Default=0.0.
class txpipe.noise_maps.TXSourceNoiseMaps(args, comm=None, aliases=None)[source]

Generate realizations of shear noise maps with random rotations

This takes the shear catalogs and tomography and randomly spins the shear values in it, removing the shear signal and leaving only shape noise

Inputs:

  • shear_catalog: ShearCatalog

  • shear_tomography_catalog: TomographyCatalog

  • mask: MapsFile

Outputs:

  • source_noise_maps: LensingNoiseMaps

Parallel: Yes - MPI

Configuration
  • chunk_rows: (int) Default=100000.
  • lensing_realizations: (int) Default=30.
  • true_shear: (bool) Default=False.
class txpipe.noise_maps.TXLensNoiseMaps(args, comm=None, aliases=None)[source]

Generate lens density noise realizations using random splits

This randomly assigns each galaxy to one of two bins and uses the different between the halves to get a noise estimate.

Inputs:

  • lens_tomography_catalog: TomographyCatalog

  • photometry_catalog: HDFFile

  • mask: MapsFile

Outputs:

  • lens_noise_maps: ClusteringNoiseMaps

Parallel: Yes - MPI

Configuration
  • chunk_rows: (int) Default=100000.
  • clustering_realizations: (int) Default=1.
  • mask_in_weights: (bool) Default=False.
class txpipe.noise_maps.TXExternalLensNoiseMaps(args, comm=None, aliases=None)[source]

Generate lens density noise realizations using random splits of an external catalog

This randomly assigns each galaxy to one of two bins and uses the different between the halves to get a noise estimate.

Inputs:

  • lens_tomography_catalog: TomographyCatalog

  • lens_catalog: HDFFile

  • mask: MapsFile

Outputs:

  • lens_noise_maps: ClusteringNoiseMaps

Parallel: Yes - MPI

Configuration
  • chunk_rows: (int) Default=100000.
  • clustering_realizations: (int) Default=1.
  • mask_in_weights: (bool) Default=False.
class txpipe.noise_maps.TXNoiseMapsJax(args, comm=None, aliases=None)[source]

Generate noise realisations of lens and source maps using JAX

This is a JAX/GPU version of the noise map stages.

Need to update to stop assuming lens and source are the same and split into two stages.

Inputs:

  • shear_catalog: ShearCatalog

  • lens_tomography_catalog: TomographyCatalog

  • shear_tomography_catalog: TomographyCatalog

  • mask: MapsFile

  • lens_maps: MapsFile

Outputs:

  • source_noise_maps: LensingNoiseMaps

  • lens_noise_maps: ClusteringNoiseMaps

Parallel: Yes - MPI

Configuration
  • chunk_rows: (int) Default=4000000.
  • lensing_realizations: (int) Default=30.
  • clustering_realizations: (int) Default=1.
  • seed: (int) Default=0.
class txpipe.auxiliary_maps.TXAuxiliarySourceMaps(args, comm=None, aliases=None)[source]

Stage TXAuxiliarySourceMaps

Inputs:

  • shear_catalog: ShearCatalog

  • shear_tomography_catalog: HDFFile

  • source_maps: MapsFile

Outputs:

  • aux_source_maps: MapsFile

Parallel: Yes - Dask

Configuration
  • block_size: (int) Default=0.
  • flag_exponent_max: (int) Default=8.
  • psf_prefix: (str) Default=psf_.
  • chunk_rows: (int) Default=100000.
  • pixelization: (str) Default=healpix.
  • nside: (int) Default=0.
  • sparse: (bool) Default=True.
  • ra_cent: (float) Default=nan.
  • dec_cent: (float) Default=nan.
  • npix_x: (int) Default=-1.
  • npix_y: (int) Default=-1.
  • pixel_size: (float) Default=nan.
class txpipe.auxiliary_maps.TXAuxiliaryLensMaps(args, comm=None, aliases=None)[source]

Generate auxiliary maps from the lens catalog

This class generates maps of:

  • depth

  • bright object counts

Inputs:

  • photometry_catalog: HDFFile

Outputs:

  • aux_lens_maps: MapsFile

Parallel: Yes - Dask

Configuration
  • block_size: (int) Default=0.
  • bright_obj_threshold: (float) Default=22.0.
  • depth_band: (str) Default=i.
  • snr_threshold: (float) Default=10.0.
  • snr_delta: (float) Default=1.0.
class txpipe.auxiliary_maps.TXUniformDepthMap(args, comm=None, aliases=None)[source]

Generate a uniform depth map from the mask

This is useful for testing on uniform patches. It doesn’t generate all the other maps that the other stages that make aux_lens_maps do, so may not always be useful.

Inputs:

  • mask: MapsFile

Outputs:

  • aux_lens_maps: MapsFile

Parallel: No - Serial

Configuration
  • depth: (float) Default=25.0.
class txpipe.masks.TXSimpleMask(args, comm=None, aliases=None)[source]

Generate a simple binary mask using cuts on depth and bright object maps.

Inputs:

  • aux_lens_maps: MapsFile

Outputs:

  • mask: MapsFile

Parallel: No - Serial

Configuration
  • depth_cut: (float) Default=23.5.
  • bright_object_max: (float) Default=10.0.
class txpipe.masks.TXSimpleMaskSource(args, comm=None, aliases=None)[source]

Generate a binary mask for source galaxies using positive lensing weights across source bins.

Inputs:

  • source_maps: MapsFile

Outputs:

  • mask: MapsFile

Parallel: No - Serial

Configuration
class txpipe.masks.TXSimpleMaskFrac(args, comm=None, aliases=None)[source]

Make a simple mask using a depth cut and bright object cut Include the fractional coverage of each pixel using a high-res survey property map (e.g. N-exposures for given band(s) )

#NOTE: This assumes all cuts to the mask are being done within TXpipe at the config Nside #if we want to apply cuts on the survey property map nside (potentially higher resolution), we will need to add this feature

Inputs:

  • aux_lens_maps: MapsFile

Outputs:

  • mask: MapsFile

Parallel: No - Serial

Configuration
  • depth_cut: (float) Default=23.5.
  • bright_object_max: (float) Default=10.0.
  • supreme_map_file: (str) Required.
class txpipe.convergence.TXConvergenceMaps(args, comm=None, aliases=None)[source]

Make a convergence map from a source map using Kaiser-Squires

This uses the wlmassmap library, which is included as a submodule in TXPipe.

Inputs:

  • source_maps: MapsFile

Outputs:

  • convergence_maps: MapsFile

Parallel: No - Serial

Configuration
  • lmax: (int) Default=0.
  • smoothing_sigma: (float) Default=10.0.
class txpipe.map_correlations.TXMapCorrelations(args, comm=None, aliases=None)[source]

Plot shear, density, and convergence correlations with survey property maps

The Supreme code generates survey property maps; this stage makes plots of the correlations with those maps with a simple linear fit.

Since the Supreme maps are loaded from a directory, outside the pipeline, we don’t know in advance what plots will be generated, so the formal output is a directory.

Inputs:

  • lens_maps: MapsFile

  • convergence_maps: MapsFile

  • source_maps: MapsFile

  • mask: MapsFile

Outputs:

  • map_systematic_correlations: FileCollection

Parallel: No - Serial

Configuration
  • supreme_path_root: (str) Default=/global/cscratch1/sd/erykoff/dc2_dr6/supreme/supreme_dc2_dr6d_v2.
  • nbin: (int) Default=20.
  • outlier_fraction: (float) Default=0.05.