Weights

Classes inheriting form Weight create weights arrays for spectra. They work very similar to Masks, but instead of returning a boolean mask they return a float array, containing a weight for every pixel in a spectrum.

spexxy comes with a few pre-defined weight classes:

Weight

class spexxy.weight.Weight(*args, **kwargs)

Weight is the base class for all objects that can create a weight array for spectra.

__call__(spectrum: spexxy.data.spectrum.Spectrum, filename: str) → numpy.ndarray

Creates a new weight for a spectrum.

Parameters:
  • spectrum – Spectrum to create weight for.
  • filename – Name of file containing spectrum to create weight for.
Returns:

Array containing weight for given spectrum.
__init__(*args, **kwargs)

Initialize a new weight.

WeightFromSigma

class spexxy.weight.WeightFromSigma(squared: bool = False, *args, **kwargs)

Reads the SIGMA extension from the given file and creates weights from it as 1/SIGMA.

This class, when called, loads the SIGMA extension from the given file and returns the weights as 1/SIGMA.

__call__(spectrum: spexxy.data.spectrum.Spectrum, filename: str) → numpy.ndarray

Creates a new weight for a spectrum from its SIGMA extension.

Parameters:
  • spectrum – Spectrum to create weight for.
  • filename – Name of file containing spectrum to create weight for.
Returns:

Array containing weight for given spectrum.
__init__(squared: bool = False, *args, **kwargs)

Initializes a new weight.

Parameters:squared – Return 1/SIGMA**2 instead of 1/SIGMA.

WeightRanges

class spexxy.weight.WeightRanges(ranges: List[Tuple[float, float, float]] = None, initial: float = 1.0, *args, **kwargs)

Creates a weights array from given ranges.

This class, when called, creates a weights array from the given wavelength ranges.

__call__(spectrum: spexxy.data.spectrum.Spectrum, filename: str) → numpy.ndarray

Creates a new weight for a spectrum from the ranges given in the configuration.

Parameters:
  • spectrum – Spectrum to create weight for.
  • filename – Name of file containing spectrum to create weight for.
Returns:

Array containing weight for given spectrum.
__init__(ranges: List[Tuple[float, float, float]] = None, initial: float = 1.0, *args, **kwargs)

Initializes a new weight.

Parameters:
  • ranges – List of tuples of (wave start, wave end, weight).
  • initial – Initial value for whole array.

WeightFromSNR

class spexxy.weight.WeightFromSNR(keyword: str = 'HIERARCH SPECTRUM SNRATIO', *args, **kwargs)

Reads the S/N ratio from the given file and creates weights from it as 1/SQRT(FLUX/SNR).

This class, when called, loads the SNR from the given file and returns weights from it as 1/SQRT(FLUX/SNR).

__call__(spectrum: spexxy.data.spectrum.Spectrum, filename: str) → numpy.ndarray

Creates a new weight for a spectrum from its S/N.

Parameters:
  • spectrum – Spectrum to create weight for.
  • filename – Name of file containing spectrum to create weight for.
Returns:

Array containing weight for given spectrum.
__init__(keyword: str = 'HIERARCH SPECTRUM SNRATIO', *args, **kwargs)

Initializes a new weight.

Parameters:keyword – FITS header keyword containing S/N.

WeightFromGrid

class spexxy.weight.WeightFromGrid(filename, initial: float = 0.0, max_line_depth: float = 0.5, center_weight: float = 1.0, max_step: int = 1, mask_lines: Union[bool, str, List[T]] = True, max_change=(300, 0.3), *args, **kwargs)

This class loads the weights from a grid depending on the initial values of the fit parameters by linear interpolation. It returns an array containing the weights.

__call__(spectrum: spexxy.data.spectrum.Spectrum, filename: str) → numpy.ndarray

Creates and returns weight array.

Parameters:
  • spectrum – Spectrum to create weight for.
  • filename – Name of spectrum file.
Returns:

Array containing the weight for given spectrum.
__init__(filename, initial: float = 0.0, max_line_depth: float = 0.5, center_weight: float = 1.0, max_step: int = 1, mask_lines: Union[bool, str, List[T]] = True, max_change=(300, 0.3), *args, **kwargs)

Initializes a new weight.

Parameters:
  • filename – Name of grid file.
  • initial – Initial value for the whole weight array.
  • max_line_depth – Central pixel for lines with larger line depth are masked out.
  • center_weight – Factor that increases the weight of the central pixel of each line.
  • max_step – In iteration steps <= max_step new weights are loaded from the grid.
  • mask_lines – List of absorption lines that are always masked out in their centers.

WeightFromGridNearest

class spexxy.weight.WeightFromGridNearest(filename, initial: float = 0.0, max_line_depth: float = 0.5, center_weight: float = 1.0, max_step: int = 1, mask_lines: Union[bool, str, List[T]] = True, max_change=(300, 0.3), *args, **kwargs)
This class loads the weights from a grid depending on the initial values of the fit parameters by choosing the
nearest neighbour in the grid. It returns an array containing the weights.
__call__(spectrum: spexxy.data.spectrum.Spectrum, filename: str) → numpy.ndarray

Creates and returns weight array.

Parameters:
  • spectrum – Spectrum to create weight for.
  • filename – Name of spectrum file.
Returns:

Array containing the weight for given spectrum.
__init__(filename, initial: float = 0.0, max_line_depth: float = 0.5, center_weight: float = 1.0, max_step: int = 1, mask_lines: Union[bool, str, List[T]] = True, max_change=(300, 0.3), *args, **kwargs)

Initializes a new weight.

Parameters:
  • filename – Name of grid file.
  • initial – Initial value for the whole weight array.
  • max_line_depth – Central pixel for lines with larger line depth are masked out.
  • center_weight – Factor that increases the weight of the central pixel of each line.
  • max_step – In iteration steps <= max_step new weights are loaded from the grid.
  • mask_lines – List of absorption lines that are always masked out in their centers.