rfwtools.extractor.down_sample.lttb_extractor

rfwtools.extractor.down_sample.lttb_extractor(example, signals, n_out, query=None)[source]

Extract features via lttb on individual signals from a set. Loads/unloads data.

LTTB is not a fixed time step method, but produces good graphical results. It uses a Largest Triangle Three Bucket approach which picks points based on which would maximize the size of triangles created but points in adjacent buckets.

Parameters:
  • example (Example) – The example on which to operate

  • signals (List[str]) – A list of the example’s columns to be down sampled (e.g., “1_GMES”).

  • n_out (int) – The number of points to be returned

  • query (Optional[str]) – Argument passed to the ex.event_df to filter data prior to feature extraction, e.g. “Time <= 0”.

Return type:

DataFrame

Returns:

A DataFrame with a single row containing the the down sampled and concatenated signals.