PSF

source

PSF(
   fits_ext, padto, *, dtype = np.complex128, norm = True
)

PSF object class.

Helper class to handle PSF data.

Args

  • fits_ext (astropy.io.fits) : opened fits file containing PSF data and required
  • padto (list of int) : shape of desired Fourier transform array.
  • dtype (np.dtype, optional) : desired complex dtype of Fourier array. header data (XPOS,YPOS,LAMBDA).

Attributes

  • fft_data (np.ndarray) : array storing the (minimal) rfft2 data of the given PSF.
  • xpos (float) : x-position of PSF point source in arcsec.
  • ypos (float) : y-position of PSF point source in arcsec.
  • Lambda (float) : wavelength used to capture the PSF.
  • norm (bool) : whether or not to normalize the PSF.

TileGenerator

source

TileGenerator(
   source, psfs_file, gauss_width_pix, *, dtype = np.complex128, which_psf = None,
   pixsize = 0.00375, norm_psf = True
)

Object for generating tiles to be sliced into final image.

Args

  • source (Source object) : Object containing all of the source data, as defined in Source.py.
  • psfs_file (str) : path to .fits file containing all PSFs and metadata.
  • gauss_width_pix (int) : support size in pixels to build the Gaussian star kernel.
  • dtype (np.dtype, optional) : complex data type to work with in DFT space.
  • which_psf (int, optional) : If specified, use fits HDU[which_psf+1] PSF only.

Attributes

  • source (Source object) : collection of sources in Source-type object.
  • psfs (list of PSF objects) : list of PSF objects (see PSF class).
  • pixsize (float) : pixel size in arcsec used to build tile.
  • gauss_width_pix (int) : width of Gaussian square support in pixels
  • static (bool) : if True, use a static PSF, otherwise use field variable PSF.

Methods:

.get_effective_psf_fft

source

.get_effective_psf_fft(
   s_pos
)

Takes star information and computes effective PSF.

From star position, the convex combination PSF is found (equivalent to bilinear interpolation since stars are defined on square grid). The resulting effective PSF is added to the internal _psf_array to be used in the get_tile pipeline.

Args

  • star_pos (np.ndarray) : position [arcsec].

.get_tile

source

.get_tile(
   index
)

Get the tile corresponding to source[index]

From the tile_generator object tgen, calling tgen.get_tile(index) will generate the tile corresponding to the tgen.source_pos[index] star by interpolating the 4 neighbouring PSFs and convolving this effective PSF with a sub-pixel shifted Dirac-delta function, and if requested, a Gaussian kernel defined by tgen.cov_mat .

The output of this is a tile which has been trimmed down to the input PSF dimensions, as well as the coordinates of the bottom-left-corner of the tile so that it may be sliced into the final image properly.

Args

  • index (int) : index of star in source table to generate tile for.

Returns

  • out (real-valued np.ndarray) : tile to be sliced into final image
  • bottom_left_corner (np.ndarray) : coordinates of bottom left corner of bottom left pixel in arcsec

.get_star_kernel_fft

source

.get_star_kernel_fft(
   flux, mu
)

Compute star Gaussian based in DFT space.

Directly computes the FFT of the Gaussian kernel with appriate amplitude, width, and offset to suit the tile being generated.

Uses optimised np.einsum so requires running optimize_star_kernel() first.

Args

  • flux (float) : flux of star.
  • mu (np.ndarray) : position of star.

Returns

  • gaussian_fft (np.ndarray) : star Gaussian kernel in FFT space.

.optimize_star_kernel

source

.optimize_star_kernel()

Runs star kernel once to optimise np.einsum

ImageGenerator

source

ImageGenerator(
   array_width_pix, source, psfs_file, pixsize = 0.00375, gauss_width_pix = 34,
   which_psf = None, norm_psf = True
)

Generate image from sliced tiles, one tile per source object.

This is the core object to work with when generating a MAVISIM image from a Source object.

Args

  • array_width_pix (int) : width of full image in pixels before rebinning.
  • pixsize (float) : pixel size in arcsec before rebinning.
  • source (Source object) : source list as Source-type object
  • psfs_file (str) : filename for fits file containing all PSFs and PSF metadata.
  • gauss_width_pix (int) : width of Gaussian square support in pixels.
  • which_psf (int, optional) : If specified, use fits HDU[which_psf+1] PSF only.

Attributes

  • pixsize (float) : pixel size in arcsec of image.
  • fov (float) : FoV of full image.
  • full_image (real-valued np.ndarray) : final image at original pixel size (i.e., before rebinning).
  • tile_gen (TileGenerator object) : tile generator object used to create tiles to slice into final image.

Methods:

.main

source

.main()

Loop over all stars and add the tile to the full image.

.get_rebinned_cropped

source

.get_rebinned_cropped(
   rebin_factor, cropped_width_as
)

Rebin self.full_image after cropping to desired rebin factor.

Args

  • rebin_factor (int) : rebinning factor from high-res image to rebinned image.
  • cropped_width_as (float) : desired width of final image in arcsec. Note that no checking is done on the validity of this, so use with care.

Returns

  • rebinned_im (real-valued np.ndarray) : complete image, rebinned and cropped.