Ayush-Gole8/tabml ? reverse-engineered prompt

Build me a simple command line app that can take a tabular dataset from either a CSV file or OpenML, figure out whether it’s a classification or regression problem, and then run an end to end analysis on it.

I want it to clean and profile the data, split it in time order if I give a timestamp column, train a few strong baseline models and boosting style models, compare a few tuning approaches, and then explain the best model in a way a normal person can understand. It should also flag predictions that seem unreliable, check for concept drift over time, compare a static model with one that gets retrained periodically, and run a statistical test to see whether the results are meaningfully different.

At the end, please generate one self contained HTML report in the results folder that shows the dataset summary, model comparison, explanations, drift findings, and the final winner. I only want to edit one config file for a new dataset, and if you need current library details, look up the latest docs online.