Calibrating model requires data with date as first index

[woldmuyn]

Technical questions

• Python version: 3.10.6
• Operating System: Windows 64 bit

Description

While trying to calibrate my model, I got following error: Exception: Shapes of model prediction (275, 2) and (2, 275) do not correspond; np.arrays 'ymodel' and 'ydata' must be of the same size

Solution was to reorder my data so that date was the first index.

suggestion

Give users a warning that date must be the first index or do an automated reorder

[twallema]

I fixed this in the objective_functions.py of pySODM by automatically rearranging the dataframe with dataframe.reorder_levels()

df = df.reorder_levels([self.time_index,]+self.additional_axes_data[idx])

where, for the influenza demo,

self.time_index = 'date'
self.additional_axes_data[idx] = ['Nc',]