[ENHANCEMENT] improve validate_series checks and messages

[dbrakenhoff]

Describe the proposed enhancement

In a course a lot of users ran into their time series not being ready for pastas, I (and @martinvonk) think adding these checks would further help users along.

• Add an error and message specifically for multi-column DataFrames:

# specific msg for multi-column DataFrame.
msg = "DataFrame with multiple columns. Please select one."
raise ValueError(msg)

• Add a check for NaT types in series.index

# check if index contains NaNs
if s.index.hasnans:
    logger.info("The time series index contains NaNs. Try to remove these with `series.loc[series.index.dropna()]`.")

[martinvonk]

Goes wrong with this kind of csv:

date,level
2016-01-21 13:00,9.55
2016-01-21 14:00,9.55
2016-01-21 15:00,9.55
,
,

series = pandas.read_csv(csv, index_col=0, parse_dates=True)
pastas.validate_stress(series) raises:
ValueError: The time-indices of series level are not monotonically increasing. Try to use series.sort_index() to fix it.