Masked data in time series

[jfigui]

This issue concerns the process_point_measurement function in process_timeseries.py.

The function stores the data at the point of interest but ignores the mask. Masked data will therefore take an arbitrary value but information about the data being masked will be lost. On the other hand, if we get the actual masked value, which is "--" we have to be careful to convert it to an arbitrary float when writing the .csv file of time series in function write_ts_polar_data of PLOT_AND_WRITE_POINT and reconvert bug to a masked value in function read_timeseries to be able to perform the right time series plot.

@wolfidan, what do you thing we should do?

[wolfidan]

Hi @jfigui , couldn't we just use NaN ?

I know that pandas digests NaN both in reading and writing, but I know that we don't use pandas for these particular functions. I will check on monday how we could do it, but I would prefer NaN than any arbitrary float.