thusser / spexxy Goto Github PK

When the spectrum.valid property is not None, the information is lost when running the method. One fix would be to add the following before returning the new spectrum.

if self.valid is not None:
    spec.valid = np.copy(self.valid[i1:i2])

https://stackoverflow.com/questions/58904695/pkg-resources-resolutionerror-not-found/61955965

When resampling a spectrum that contains NANs using Spectrum.resample() and changing the number of pixels, the output will contain NANs in wavelength regions that were valid before. A possible fix would be to replace the following statement in Spectra.resample().

output.flux[start_index:start_index+len(ip)] = ip[start_index:start_index+len(ip)]

with

i_min, i_max = output.indices_of_wave_range(part.wave.min(), part.wave.max())
output.flux[i_min:i_max] = ip[i_min:i_max]

Hi, I have installed spexxy and all its dependencies, but I am having a problem with spexxytools as I can't seem to produce any grid.csv output. Testing out the program, the code seems to run (using the commands from the documentation: spexxytools grid create test_spexxy_fits, where test_spexxy_fits contains a FITS file) and produces no error, but I get no output. Is this a known problem? Apologies if I am missing something. I have tried it with several different FITS formats (the spexxy example in tests/testdata/specta/, an example from ULySS, and another from my own simulations) and get the same result for each.

I am hoping to use spexxy as a python alternative to ULySS, as a way to get stellar parameters from spectra, so any further advice on how to achieve this would be greatly appreciated.

Many thanks.

An exception should be return when either of the following conditions is fulfilled.

if wave < 0.5 * (w[1] + w[0]):
    raise spexxyValueTooLowException()
if wave >= 0.5 * (w[-1] + w[-2]):
    raise spexxyValueTooHighException()

Otherwise Spectrum.indices_of_wave_range() throws an error as it receives None.

The line should be changed to if self.crval2 > 0: - Because self.crval1 refers to the sampling of the LSF at a given wavelength, which is centred on zero, this condition is probably never fulfilled. So even if the LSF changes with wavelength, only a constant LSF is used.

Should expand if clause:

        if self._wave_mode == PyCrocoSpectrum.Mode.LAMBDA:
            # wavelength in AA means a multiplication
            self._wavelength = self.wave * (1. + vrad / 299792.458)
            # and non-constant step size
            self._wave_step = 0
        elif self._wave_mode == PyCrocoSpectrum.Mode.LOGLAMBDA:
            # wavelength in log domain is simpler
            self._wavelength = self.wave + math.log(1. + vrad / 299792.458)
            self._wave_start = self._wavelength[0]
        elif self._wave_mode == Spectrum.Mode.LOG10LAMBDA:
            self._wavelength = self.wave + math.log10(1. + vrad / 299792.458)
            self._wave_start = self._wavelength[0]
        else:
            raise NotImplementedError('Unsupported wave mode: {}'.format(self._wave_mode))