For linear regression (including Ridge regression) one can literally get the coefficients from P123 predictions output (unable to show a screenshot), multiply the feature values times the coefficient for that feature, sum this product over all of the features (from a download or from the API) and sort this sum (largest first). Features in the columns and Tickers in the rows. The sum of the products (for each ticker) as a new column to be sorted.
You would have to know what RankPos cutoff you want to use to trade this. But once you have this RankPos value, you could enter the trades into a port manually—or use InList to rebalance in a port.
It can be done—at a small price with a small resource overhead--with just a spreadsheet by a member.