Compare resample_poly with decimate

[kkappler]

Currently using scipy.signal.decimate in time_series_helpers.py. resample_poly seems faster and has more capabilities.

Before switching over, setting up a tool to compare results of both and running it on some example time series should be done. This could be put into a test.

A preliminary test has been created in aurora/tests/synthetic/test_decimation_methods.py

This currently compares aurora's prototype_decimate function with mth5's built-in run_ts.sps_filters.decimate