hifast.cube Regridding#

hifast.cube

python -m hifast.cube **/data/*-fc*.hdf5 --outname ./test_cubes.fits --bwidth 60 -p SIN --method gaussian

Workflow#

  1. Input files containing spectra corrected for Doppler (coordinate system) shifts, identified by -fc in their filenames. (For now, all files need have the same spectral sampling times.)

  2. Read the input files’ coordinates to create a WCS header, forming the RA-DEC grid. Grid point spacing is set by --bwidth. By default, the WCS RA and DEC range is defined by the input files, but can also be set using --ra_range and --dec_range.

    • --type3: The third axis: vopt, vrad, or freq (default: vrad)

    • --range3: Third axis range (default: None)

    • -p: Celestial projection method. See arXiv:astro-ph/0207413, section 7.2, for choices (default: AIT)

    • --wcs_ra_center, --wcs_dec_center: manually set “CRVAL1” and “CRVAL2” in WCS header.

  3. “Convolution” - --r_cut: The range within which spectral lines contribute to each grid point, in arcseconds. - --beam_fwhw: Beam size (FWHM); unit: arcminutes (default: 2.9) - --method: Convolution kernel types: gaussian, bessel_gaussian, or sinc_gaussian (default: gaussian, Mangum et al., arXiv:0709.0553)

    • gaussian: Used parameters:

      • --gaussian_fwhw: Unit: arcminutes; default: beam_fwhw/2

      • --r_cut: Defaults to 3*gaussian_sigma, i.e., 3*(gaussian_fwhw/(sqrt(8ln(2))))

    • bessel_gaussian: Used parameters:

      • --bsize: Unit: arcminutes; default: 1.55*beam_fwhw/3

      • --gsize: Unit: arcminutes; default: 2.52*beam_fwhw/3

      • --r_cut: Defaults to 3.8317059702075*bsize/pi

    • sinc_gaussian: Used parameters:

      • --bsize: Unit: arcminutes; default: 1.55*beam_fwhw/3

      • --gsize: Unit: arcminutes; default: 2.52*beam_fwhw/3

      • --r_cut: Default is bsize

    • --frac_finite_min FRAC_FINITE_MIN: For a grid point with n spectra within r_cut, if the count of finite value (neither nan nor infinite) at a channel (frequency sampling) is below FRAC_FINITE_MIN * n, the output for that channel is set as nan (default: 1)

    • --polar {XX,YY,M}: Polarization choice (default is M, merging both polarizations.)

Parameters#

To view more parameter details, use python -m hifast.cube -h | more.