Draft Task: Handling empty sensor data from Actfast

[Asanto32]

Description

Currently, we raise a Value Error if the sensor data (measurement or time) is empty from Actfast. However, it is only acceleration that is a mandatory field for further analysis (enmo, angle z, sleep, physical activity metrics, etc.) This task is to built on the current architecture to handle these potentially empty Measurement fields

Tasks

• [ ] Task 1

Freeform Notes

@ReinderVosDeWael Perhaps we raise the value error only for acceleration? And if a general measurement we just give a user warning??

def validate_accelereation(cls, v)
if v.measurements.size == 0:
raise ValueError...

but we try to give a warning for general empty measurements??

def validate_measurements_not_empty...
if v.size == 0:
warn("this field is empty")

[nx10]

You probably want to check len(data) == 0 first and then something like datetime[-1] - datetime[0] < minimum time interval to be able to process, right?

[ReinderVosDeWael]

For further discussion: an error seems to be the correct behavior, but do we want to allow for an option to ignore errors in sensors we don't strictly need?