What is a good way to do "triple barrier" and "trend scanning" with Backtesting and Scikit-Learn?

[BradKML]

Just saw that this is being used to create classification data for making strategy.
would like to see the "triple barrier" to be used to tag different times (early stop loss vs horizontal moves vs high potential to win) https://www.mlfinlab.com/en/latest/labeling/tb_meta_labeling.html
Similarly "trend scanning" would be good for finding ways to enter and exit in the most minimal way https://www.mlfinlab.com/en/latest/labeling/labeling_trend_scanning.html

[BradKML]

Noted that others have made an "extension" of the idea https://github.com/nkonts/barrier-method
See also this notebook https://github.com/maxzager/Financial-series-and-Triple-Barrier-Method