Removal of HOUGH algorithm and cosmic killer as they are obsolete and never used
Correction of rotation and transposition of images. Now ROOT and midas files are modified so that they are straight before operation (meaning top is top and right is right with point 0,0 on the top left corner of the image)
Correction of x and y storage of data in the TTree (rows and columns were wrongly interpreted as x-y, but they are y-x)
Transformation from square sensor to sensor with potentially different dimensions
Addition of QUEST camera as a possible camera to be analysed.
CARE:
Old pedestals and the .so library of the cython part will need to be remade with this reco.
Requires cygnolib above version 17 (daq_cam2array was modified for the quest in the cygno libs)
No vinetting map exists for QUEST yet
TEST performed:
Using MANGO, LNF and LNGS data with iron source located on the right (MANGO) and on top (LNF, LNGS) it was possible to correctly determine the orientation of the pictures. The shape of the noise bands was also exploited.
The pedestal was created flipped by 180 deg or mirrored right-to-left and the best performances (less noise clusters) was confirmed to be found with current implementation
Back compatibility of the ROOT data format was tested
The reconstruction was tested before and after the changes checking that for squared sensors (LNGS, LNF) the clusters found were compatible in the two versions
After x-y inversion in the saving of the data, the output was crosschecked with the raw images and the debug mode output was also aligned
After x-y inversion it was confirmed that the cluster variables which depend on the eigenvector rotation (width, length, tgausssigma etc..) were not affected by the change
Multicore analysis was tested
NOT TESTED yet on cloud (image is still missing version 17 of cygnolibs
Large update on reconstruction code. Main things:
CARE:
TEST performed: