Custom Ebin binning, Component analysis following 4FGL

[mireianievas]

• Allow custom Ebin binning in the SED.
• Add the possibility to use a component analysis similar (although not exactly the same because the core radius here is fixed to 2, it does not depend on the source) to the one in the 4FGL (up to 15 components with different energy ranges, ev types, zmax, rads, binning per decade).
• Use energy dispersion by default (except for the diffuse components) ... maybe we want to discuss this?
• Catch Runtime errors and retry running (up to 5 times) the LAT Science Tool binaries ... just trying, in an ugly way, to make the code more robust against job crashes in clusters.
• Raise the Ecut value from 30GeV to 100GeV.
• Change the wget addresses as the FTP in LAT stopped working [maybe temporarily ?]

[davidsanchez]

Thanks a lot do you know the Fermi team recommendations for the energy dispersion ?

[mireianievas]

Thanks a lot do you know the Fermi team recommendations for the energy dispersion ?

They seem to recommend applying them globally (particularly if you are going below a few hundred MeV, https://fermi.gsfc.nasa.gov/ssc/data/analysis/documentation/Pass8_edisp_usage.html) except for the diffuse components. Fermipy also seems to enable it by default (https://fermipy.readthedocs.io/en/latest/config.html).

PS. I discovered a few minor bugs related to the xml generation and component tagging in the code and an old piece of code for Front/Back events which looks to me like it is old and totally unneeded. I am doing some tests and I will probably make today another small pull request with the proposed changes.