Observation

class jwst.wfss_contam.observations.Observation(direct_image, segmentation_map, grism_wcs, direct_image_wcs, boundaries=None, max_cpu=1, max_pixels_per_chunk=50000.0, oversample_factor=2, band_wavelengths=None)

Bases: object

Define an observation leading to a single grism image.

The Observation class is responsible for calling the various WCS transforms that convert a direct image and a segmentation image into a simulation of the grism image, making assumptions about the spectral properties of the direct image sources. When the disperse_order method is called one or more times, two products are created: the simulated dispersed image (simulated_image attribute) and the simulated source cutouts (simulated_cutouts attribute).

Initialize all data and metadata for a given observation.

Parameters:

direct_imagenp.ndarray

Direct imaging data. May be 2-D (ny, nx) for a single-band direct image, or 3-D (N, ny, nx) for a multi-band cube.

segmentation_mapnp.ndarray

Segmentation map data.

grism_wcsWCS

WCS object from grism image

direct_image_wcsWCS

WCS object from direct image

boundarieslist, optional

Start/Stop coordinates of the FOV within the larger seed image.

max_cpuint, optional

Max number of cpu’s to use when multiprocessing

max_pixels_per_chunkint, optional

Maximum number of pixels per chunk when dispersing sources

oversample_factorint, optional

Factor by which to oversample the wavelength grid

band_wavelengthsarray-like of shape (N,), optional

Central wavelengths (in microns) for each plane of a 3-D direct_image. Required when direct_image is 3-D; ignored when direct_image is 2-D.

Methods Summary

chunk_sources(order, lambdas, sens_waves, ...)	Chunk the sources into groups of max_pixels.
disperse_order(order, wmin, wmax, ...[, ...])	Disperse the sources for a given spectral order, with multiprocessing.

Methods Documentation

chunk_sources(order, lambdas, sens_waves, sens_response, selected_ids=None, max_pixels=100000.0, basis_models=None, trace_pdt=None)

Chunk the sources into groups of max_pixels.

Parameters:

orderint

Spectral order to process

lambdasndarray

Wavelengths at which to compute dispersed pixel values for this order, precomputed once per order.

sens_wavesndarray

Wavelength array from photom reference file

sens_responsendarray

Response (flux calibration) array from photom reference file

selected_idslist, optional

List of source IDs to process. If None, all sources are processed.

max_pixelsint, optional

Maximum number of pixels per chunk.

basis_modelslist of callables, optional

Flux distributions to evaluate at each wavelength. Typically these will be Legendre polynomials from 1st order up to a given max order, e.g. [P1(x), P2(x), …], the coefficients of which are linearly fit later. If None, no models are included in the output.

trace_pdtTracePDT, optional

Precomputed pixel dispersion table for this spectral order. If None, no lookup table is used.

Returns:

disperse_argslist[list]

Outer list has length number of groups, and each inner list contains the arguments to disperse() for that group in the format that multiprocessing starmap expects.

disperse_order(order, wmin, wmax, sens_waves, sens_response, selected_ids=None, basis_models=None, pdt_spacing=None, pdt_wl_oversample=1)

Disperse the sources for a given spectral order, with multiprocessing.

The simulated_cutouts and simulated_image attributes are updated in place.

Parameters:

orderint

Spectral order to process

wminfloat

Minimum wavelength for dispersed spectra

wmaxfloat

Maximum wavelength for dispersed spectra

sens_wavesndarray

Wavelength array from photom reference file

sens_responsendarray

Response (flux calibration) array from photom reference file

selected_idslist, optional

List of source IDs to process. If None, all sources are processed.

basis_modelslist of callables, optional

Flux distributions to evaluate at each wavelength. Typically these will be single polynomial orders, e.g. [lambda x: x, lambda x: x^2], …] the coefficients of which are linearly fit later. If None, no models are included in the output.

pdt_spacingint, optional

Spacing of grid points to sample along the x and y axes when building a cached lookup table for the trace-shape transform. If None (the default), no lookup table is used and the exact transform is evaluated directly for every pixel.

pdt_wl_oversamplefloat, optional

Oversampling factor for the lookup table’s wavelength grid, relative to the native dispersion scale. If None, the lookup table’s wavelength grid instead exactly matches the dispersal wavelength grid, allowing evaluate_grid to skip wavelength-axis interpolation entirely. Only used if pdt_spacing is not None.