Make the configuration file optional. This is particularly useful when only generating fluxes for instance. In the process the description of the input file has been clarified.

Get rid of all masked arrays. It seems to work at least when there is no upper limit. It should save memory (perhaps) and speed up computation (likely).

The test did not work for some reason. If I understand the code correctly, it should be enough to test the min and max values of the histogram.

Improve information regarding the analysis speed. Now we update the data for every object analysed. Also fixed a bug that flooded the terminal when computing models.

Use numpy version of all rather than the stock version. Cuts down th analysis from 1.47s to 0.10s on a test case.

Use numpy version of min/max rather than the stock version. Cuts down th analysis from 5.86s to 1.47s on a test case.

Set the astropy log level to ERROR rather than WARNING to avoid clogging the terminal with unimportant messages.

Given the current formula, an AGN fraction of 1 would yield an infinite luminosity. We are clearly out of the accepted bounds for the model. So rather than ruining the fits, we rather raise an Exception and terminate.

simplification of the pdf analysis using fixed number of bins and a minimum error to 5% of the values

Estimate the parameters by binning the parameters into bins, with the height of the histogram based on the likelihood of the points in each bin. Adapted from an initial patch by Denis.

Check that the id and redshift columns are present in the input file. Freely inspired from an initial patch written by Denis.