Core Concepts

Auto-handling

fling inspects your DataFrame and handles common data problems before training begins.

Behaviours

ProblemWhat fling does
Text / categorical columnsOneHotEncodes them
Missing numeric valuesFills with column median
Missing categorical valuesFills with most frequent value
Columns with > 50% missingDrops them, logs in report
ID-like columns (all unique)Drops them automatically
Choosing a modelPicks based on dataset size
Train/test split orderingAlways splits before preprocessing
All of this happens inside a sklearn Pipeline - fit on training rows only, applied to test rows using training statistics.