Parallel Processing: Integrate Parallel Helpers Into Functions That Loop Or Groupby-Apply

[mdancho84]

Running list of functions that should be evaluated for parallel processing / progress bars for engine = 'pandas'.

Core

• [ ] summarize_by_time - Not sure if we need to parallelize agg() function
• [ ] apply_by_time
• [x] future_frame
• [x] pad_by_time - Implementing Cartesian Product (Vectorized Padding Technique)

Anomalies

• [x] anomalize

Feature Engineering

• [ ] augment_lags
• [ ] augment_leads
• [x] augment_rolling
• [x] augment_rolling_apply
• [x] augment_expanding
• [x] augment_expanding_apply
• [ ] augment_hilbert
• [ ] augment_wavelet
• [ ] augment_fourier

TS Summaries

• [x] ts_features
• [x] ts_summary

[mdancho84]

Keeping this quote for posterity. THIS is why vectorization is important.

[mdancho84]

Parallel Processing update - I've switched to using pathos because it handles lambda functions whereas concurrent futures ProcessPoolExecutor does not. https://github.com/uqfoundation/pathos