pypeit.core.spectrum module

Temporary light-weight spectrum object.

Ideally this would be replaced by specutils.Spectrum

class pypeit.core.spectrum.Spectrum(wave, flux, ivar=None, gpm=None, meta=None)

Bases: object

A light-weight container class for a spectrum.

The flux array can be either 1D or 2D. If 2D, its shape is always assumed to be \((N_{\rm pix},N_{\rm spec})\), where \(N_{\rm pix}\) is the number of pixels per spectrum and \(N_{\rm spec}\) is the number of spectra. The wavelength array must always be provided as 1D; i.e., for a 2D flux array, the wavelength vector must be correct for each flux vector along the 1st axis of the 2D array.

Parameters:

• wave (array-like) – Vacuum wavelengths in angstrom. Must be 1D, and its length must match the first axis of flux.

• flux (array-like) – Flux data at each wavelength. Can be 2D or 1D; see class description.

• ivar (array-like, optional) – Inverse variance in the flux. Shape must match flux. If None, assumes uncertainties are unknown.

• gpm (array-like, optional) – Boolean good-pixel mask. Shape must match flux. If None, assumes all pixels are valid.

• meta (dict, optional) – Collection of relevant metadata.

copy()

Return a deepcopy of the object.

inverse()

Replace the spectrum with its multiplicative inverse.

This modifies the spectrum in place. If uncertainties are available, they are propagated. Any divisions by 0 result in an inverse variance of 0 and the good pixel mask is set to False.

multiply(a)

Multiply the spectrum by a scalar, vector, or another spectrum.

This modifies the spectral data in place. If uncertainties are available, they are propagated. Any divisions by 0 result in an inverse variance of 0 and the good pixel mask is set to False.

Parameters:

a (scalar, array-like, pypeit.core.spectrum.Spectrum) – Multiplicative factor. If an array, its shape must match flux. If a spectrum, the wavelength arrays of the two spectrum must be identical.

property ndim

The dimensionality of the flux array

resample(new_wave, pixel_fraction_threshold=0.8, conserve=False)

Resample the spectrum to a new wavelength array.

If available, errors and masking are both propagated through the calculation. This function is basically a wrapper for Resample.

Parameters:

• new_wave (array-like) – New wavelength vector for the spectrum

• pixel_fraction_threshold (float, optional) – The resampling calculates the fraction of each output pixel that has unmasked contributions from the original spectrum. Fractions below this threshold will be masked in the output spectrum.

• conserve (bool, optional) – Conserve the flux in the resampled spectrum. If the units of the spectrum are flux integrated over the pixel, this should typically be True; if the units are flux density (e.g., \({\rm ergs/s/cm}^2{\rm /angstrom}\)), this should typically be False.

Returns:

A new spectrum object with the resample data.

Return type:

Spectrum

property shape

The shape of the flux array

property size

The size of the flux array

to_magnitude(zeropoint=0.0)

Convert the spectrum to magnitudes.

For fluxes, \(f\), this returns

\[m = -2.5 \log_{\rm 10} (f) + Z,\]

where \(Z\) is the provided zeropoint.

This modifies the spectrum in place. If uncertainties are available, they are propagated. Any pixels with non-positive fluxes are masked.

Parameters:

zeropoint (float, optional) – The magnitude conversion zeropoint (see above)