Handle input intensity measures in the form of RotD50

[bodlukas]

The current version computes the geometric mean (average horizontal) from the provided stationlist.json file. This is currently handled within Stations.get_recordings . In some use cases, the station data might already be pre-processed and contain for example the RotD50 intensity measures.

To allow for such use cases, the code needs to be modified as followed:

• Add a keyword argument im_definition to Stations.get_recordings, which is either 'geometric_mean' or 'rotD50'. If it is 'rotD50' it will only read the corresponding column in the dataframe, while for 'geometric_mean', it performs the same computations as now.

Importantly, the dataframe created from the stationlist.json file should contain the correct column names :

• 'geometric_mean' : the columns should end with 'E' and 'N' (e.g., 'sa(1.0)_E' and 'sa(1.0)_N') -'rotD50' : the column should end with 'rotD50' (e.g., 'sa(1.0)_rotD50')

The provided utility function to read the station data in utils.py assumes a .json file (similar to the one provided for the 2023 Turkey/Syria earthquakes). And it assumes that both horizontal directions are provided (ending with '1' or '2' or 'E' and 'N'). Yet, this largely on the file and format of the station data available to the user. The user just needs to make sure that the resulting dataframe contains the columns mentioned above.

Note: The currently used GMMs in the example notebooks are defined for the geometric mean. If one uses RotD50, then the user need to make sure that one uses a GMM which is defined for RotD50.

[bodlukas]

I already started to modify the Stations.get_recordings method in branch feature/rotd50_inputs.