Skip to content

Flux

Jump to bottom
Patrick Rauer edited this page Nov 16, 2019 · 2 revisions

Flux

Phosphorpy provides a system to convert magnitudes to fluxes. Based on the flux zero points from SVO the magnitudes are converted with respect to the used magnitude system (Vega or AB). To extract the necessary values the package armapy is used, which provides the opportunity to download filter bandpasses and the general filter information from SVO.
To get the fluxes of all filters use:

fluxes = ds.flux.data

and the fluxes parameter will contain a pandas DataFrame with the fluxes of the targets in the same order as coordinates or magnitudes. The labeling of the columns will be the same as for the magnitudes.

Plotting

The flux environment also provides a basic plotting system. In this case one can plot the SED.
For example, if the first SED of the data set should be plotted in a log-log environment

ds.flux.plot.sed(1, x_log=True, y_log=True)

output_6_0

As in all plotting functions there is the option to give a path string to store the resulting graph directly on ones machine.

Fitting

The flux environment provides a system to fit the SED's. The fitting system has the advantage that it checks automatically if data are missing and fit's only the non NaN values.
At the moment it has two possibilities to fit the SED. The first one is a fit by a polynomial

ds.flux.fit_polynomial(4, error_weighted=True, 
                       x_log=True, y_log=True, 
                       lower_limit=3500, upper_limit=20000)
a b c d e
1 -46.623054 733.619297 -4328.883092 11351.961849 -11177.727124
2 -41.370800 651.196488 -3843.728544 10082.202070 -9930.335327
3 -41.717332 658.645777 -3900.593758 10268.576531 -10154.307088
4 -31.537566 501.435547 -2990.777005 7929.834929 -7900.377867
5 -37.536985 591.408458 -3494.057726 9173.216647 -9044.204596
... ... ... ... ... ...
13557 -36.199599 570.667527 -3374.208532 8867.666862 -8754.640144
13558 -35.324727 558.530888 -3312.204648 8730.078710 -8643.785576
13559 -30.393473 484.873625 -2901.170994 7714.933610 -7707.423603
13560 -38.916592 611.689858 -3605.424703 9444.016989 -9290.117830
13561 -24.311663 385.814025 -2296.270627 6073.646342 -6038.408832

13561 rows x 5 columns

'4' is the degree of the polynomial, 'error_weighted' means that the inverse error is used as a weight for the fit. 'x_log' and 'y_log' have the same meaning as in the plotting case, and 'upper_limit' and 'lower_limit' are limits for the wavelength coverage. If no astropy unit is given then they are taken as angstrom like in the case above.

The second option to fit the SED is more general. Instead of the usage of already implemented functions, the fitting function is given as a parameter. The style of the function is the same as the function curve_fit of the scipy package requires.

def fit_function(wavelength, *parameters):
    y = np.zeros(len(wavelength))
    for p in parameters:
        y += wavelength*p
    return y

zero_points = [0, 0, 0, 0]
ds.flux.func_fit(fit_function, 
                 error_weighted=True, 
                      x_log=True, y_log=True, 
                      lower_limit=3500, upper_limit=20000,
                      p0=zero_points)
a b c d
1 -0.732213 -0.732213 -0.732213 -1.946794
2 -1.586122 -1.586122 -0.703275 -0.182668
3 -0.525108 -0.884023 -2.737084 0.022580
4 -1.326313 -1.326313 -0.645434 -0.649124
5 -1.936848 -1.936848 0.515907 -0.710470
... ... ... ... ...
13557 -2.069036 -2.069036 -0.288950 0.333608
13558 -0.292970 -0.292970 -0.506712 -3.063576
13559 -0.177273 -0.177273 -1.822169 -1.774176
13560 -1.323851 -1.323851 -0.439786 -0.892992
13561 -2.023593 -2.023593 -0.076104 -0.076104

13561 rows x 4 columns

Note, that func_fit needs a function to fit with an explicit declaration of parameters or a list with zero-points. Otherwise Phosphorpy can not determine the number of parameters and will raise an error.

If the latest results are needed later again, they are can be accessed with

ds.flux.fit
a b c d
1 -0.732213 -0.732213 -0.732213 -1.946794
2 -1.586122 -1.586122 -0.703275 -0.182668
3 -0.525108 -0.884023 -2.737084 0.022580
4 -1.326313 -1.326313 -0.645434 -0.649124
5 -1.936848 -1.936848 0.515907 -0.710470
... ... ... ... ...
13557 -2.069036 -2.069036 -0.288950 0.333608
13558 -0.292970 -0.292970 -0.506712 -3.063576
13559 -0.177273 -0.177273 -1.822169 -1.774176
13560 -1.323851 -1.323851 -0.439786 -0.892992
13561 -2.023593 -2.023593 -0.076104 -0.076104

13561 rows x 4 columns

Previous results are not stored.

Clone this wiki locally