norm detrend fix

[joshuabmoore]

This PR fixes a bug in which time-series datasets were incorrectly normalised before de-trending. The following changes have been made:

• The order in which pre-processing is performed has been corrected to first de-trend the data (if specified by the user) before normalising. Documentation has been updated to reflect these changes.
• The de-trending and z-score normalising operations can be individually enabled by setting the detrend and normalise flags when instantiating the Calculator() object. Both pre-processing steps are enabled by default (as per previous versions of pyspi).
• Summary statements are printed out to the user when instantiating the calculator object verifying the pre-processing steps that have been applied (detrending and/or normalising if enabled).
• Coloured outputs have been added to enhance clarity via. a new colorama dependency.

While the reversal of pre-processing operations will not introduce breaking changes for existing users, a discrepancy between results obtained for pyspi < 2.0.0 and pyspi >= 2.0.0 is expected.

[anniegbryant]

This is great, Josh! The green-colored print messages explicitly stating whether detrending and/or normalizing were applied are a nice touch, and the docstring for the Data class summarizes the detrending + normalization steps well.

[joshuabmoore]

This is great, Josh! The green-colored print messages explicitly stating whether detrending and/or normalizing were applied are a nice touch, and the docstring for the Data class summarizes the detrending + normalization steps well.

Thanks Annie!