Handle non-equidistant data gracefully

[ecomodeller]

>>> from datetime import datetime
>>> import pandas as pd
>>> from anomalydetection.detectors import DiffRangeDetector
>>>
>>> dates = [
...     datetime(2000,1,1),
...     datetime(2000,1,2),
...     datetime(2000,1,7)]
>>>
>>> normal_data = pd.Series([0.0,0.5,0.0])
>>> normal_data.index = dates
>>>
>>> df_detector = DiffRangeDetector()
>>>
>>> df_detector.fit(normal_data)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "c:\users\jan\code\anomalydetection\anomalydetection\detectors.py", line 155, in fit
    data_diff = self.diff_time_series(data)
  File "c:\users\jan\code\anomalydetection\anomalydetection\detectors.py", line 147, in diff_time_series
    time_diff = data.index.shift() - data.index
  File "C:\Users\JAN\Miniconda3\lib\site-packages\pandas\core\indexes\datetimelike.py", line 532, in shift
    result = arr._time_shift(periods, freq=freq)
  File "C:\Users\JAN\Miniconda3\lib\site-packages\pandas\core\arrays\datetimelike.py", line 1148, in _time_shift
    raise NullFrequencyError("Cannot shift with no freq")
pandas.errors.NullFrequencyError: Cannot shift with no freq