You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Originally it seemed like a decent idea to hardcode the cuts used for the CDL to generate the reference spectra. Certainly for the 2014/15 case, because those cannot change anymore. They are just a fixed reference.
Given how different fits can change these values easily however, it seems ridiculous and is a lot of work to update those values always an changes to the fits.
New cuts are generated by running the cdl_spectrum_creation tool on a fully reconstructed H5 file of the CDL datasets, as for example:
The code automatically calculates the the correct minimum and maximum value for the cut.
However, instead of hardcoding the fit results, this should be done in a more flexible way!
which calculates those values also based on a given fit parameter text file.
We could maybe add something to our config.toml file - an option to hand a path to a good fit parameter file and then read that at run time. The "currently correct file" could be stored in some place in the repository maybe.
The text was updated successfully, but these errors were encountered:
Originally it seemed like a decent idea to hardcode the cuts used for the CDL to generate the reference spectra. Certainly for the 2014/15 case, because those cannot change anymore. They are just a fixed reference.
Given how different fits can change these values easily however, it seems ridiculous and is a lot of work to update those values always an changes to the fits.
New cuts are generated by running the
cdl_spectrum_creation
tool on a fully reconstructed H5 file of the CDL datasets, as for example:which will generate a
fitparameters_accurate_<timestamp>.txt
file, such as:https://github.com/Vindaar/TimepixAnalysis/blob/master/resources/archive/fitparams_accurate_1600446157.txt
From here one has to pick the main peaks of the charge fits (those with highest amplitude essentially) and take the mean
gmu
and sigmags
. Those can be inserted here:https://github.com/Vindaar/TimepixAnalysis/blob/master/Analysis/ingrid/private/cdl_cuts.nim#L122-L149
for each of the target + filter combinations.
Note that the energy ranges are from lowest to highest (index 0 is C with 0.6 kV, 7 is Ni with 15 kV).
The code automatically calculates the the correct minimum and maximum value for the cut.
However, instead of hardcoding the fit results, this should be done in a more flexible way!
Think about it! There is a mini script here:
https://github.com/Vindaar/TimepixAnalysis/blob/master/Tools/calcCdlCutsFromFitParams.nim
which calculates those values also based on a given fit parameter text file.
We could maybe add something to our
config.toml
file - an option to hand a path to a good fit parameter file and then read that at run time. The "currently correct file" could be stored in some place in the repository maybe.The text was updated successfully, but these errors were encountered: