We should generate all plots for a file being reconstructed into its own directory.
Essentially we run the reconstruction (either from raw or to apply calibrations) and based on maybe the creation date (or modification date?) of the input file we create a new directory in which all plots are put.
We have to make sure that if individual reconstruction runs are done one after another with different arguments that all plots are put into the same directory.
So maybe use creation date of the reconstruction file, generate some prefix from a mix of filename and date and store that as an attribute in the root of the file. That way we can both always determine the plots by looking at a H5 file and can programmatically store the plots in the correct directory.
We should generate all plots for a file being reconstructed into its own directory.
Essentially we run the reconstruction (either from raw or to apply calibrations) and based on maybe the creation date (or modification date?) of the input file we create a new directory in which all plots are put.
We have to make sure that if individual reconstruction runs are done one after another with different arguments that all plots are put into the same directory.
So maybe use creation date of the reconstruction file, generate some prefix from a mix of filename and date and store that as an attribute in the root of the file. That way we can both always determine the plots by looking at a H5 file and can programmatically store the plots in the correct directory.