Python Notebook - Cointegration pairs trading research

I'm looking to build a Python/Jupyter notebook cointegration based pairs-trading compares the efficacy of different unit root tests in identifying stationarity. Here, emphasis is on comparing overused methods like ADF and KPSS that fail to account for structure breaks, against test(s) that do account for these breaks (Zivot Andrews and any other relevant tests). So I would like to use differnet unit root tests in the Engle Granger framework and compare performance/trading activity. As this is for a project, i would ideally like it to have a series of outputs to be analysed including any pnl, t/pvalues, time of structure breaks, etc ( feel free to reccomend any other relevant stats here). End product would ideally be. a Jupyter notebook that takes csv of stocks as inputs (yfinance directly to dataframes). Please let me know if this is something that you can do (I've attached an example of what I have so you get a rough idea idea). Given your background, i welcome any idea you think would be beneficial/relevant for this (ie, probably choosing from multiple pairs for the one that diverges most from the equlibrium or some other measure)