New metadata classes to perform calibration and odds analysis over a TRestDataSet #392
Conversation
…over a TRestDataSet
for more information, see https://pre-commit.ci
| /// This class is meant to perform the calibration of different runs | ||
| class TRestCalibration : public TRestMetadata { | ||
| private: | ||
| // Name of the output file |
There was a problem hiding this comment.
I would leave one white space between data member definitions so that is easier to read the documentation line corresponding to the metadata member.
Also I think doxygen requires to use a triple /// comment.
| /// TRestCalibration performs the calibration of a TRestDataSet, the | ||
| /// calibration is performed over different peaks provided in the config | ||
| /// file, in which the first peak should corresponds to the maximum in the | ||
| /// spectrum histogram. |
There was a problem hiding this comment.
I understand then that the peaks are sorted in descending energy order? First peak highest energy, latest peak the lowest energy?
There was a problem hiding this comment.
Not necessary, the requirement is that the first peak should correspond to the maximum in the spectra. The rest of the peaks could be randomly ordered.
|
These new classes are clearly in a new type of metadata class that uses the dataset to generate a new column on the input dataset. We should think that there might appear more classes of this type. Perhaps we should create a new directory to put those closes inside, together with I would also think about introducing a name root to these classes, such as |
Co-authored-by: Javier Galan <javier.galan@unizar.es>
Co-authored-by: Javier Galan <javier.galan@unizar.es>
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
New metadata clasess have been added under framework:
TRestCalibrationis meant to perform the calibration of aTRestDataSet, the calibration is performed over different peaks provided in the config file. If an output file name is provided a new dataset will be generated with the corresponding metadata and calibration constants, while adding a new observable,calib_energy. In addition, the spectrum and the fit results will be stored in the output file.TRestOddsperforms the log odds of the different observables given in the config file and for a particular dataSet. To perform the log odds first the probability density funcion (PDF) is obtained for a set of observables in the desired range and used to compute the log odds. New observables are created in the output dataSetodds_obsNameand the addition of all of them inodds_total. If an input odds file is provided, the different PDFs are retrieved from the input file.