Attempting to calibrate to an empty dataset not caught by input checks of log_posterior_probability

[twallema]

What I Did

Sliced a dataset I wanted to calibrate a model to (input argument data of log_posterior_probability) in a wrong way making it an empty series:

Series([], Name: Weekly_Hosp, dtype: float64)

This was not caught by the input checks resulting in the following error in pySODM's internals:

  File "/Users/twallema/Documents/academic/github/pySODM/src/pySODM/optimization/objective_functions.py", line 428, in compute_log_likelihood
    interp = out_copy.interp({time_index: df.index.get_level_values(time_index).unique().values}, method="linear")

when trying to calibrate that model.

Proposed fix

Add an appropriate input check for empty datasets.

[twallema]

Caught by #98