Three layered gawk tools. Each is one self-contained binary, a few hundred bash lines wrapping embedded heredocs of awk. Each does one thing. Each can be read in a sitting. Each builds on the previous.
Object sugar for gawk. .it.field → HEAP[it]["field"].
Preprocessor + 9-line runtime + helpers.
Column types on top of dot: Num (running mean+stdev), Sym (counts/mode/entropy),
Data (CSV table). One-call --get-data for 30 curated CSVs.
Overview · Stats demo · Manual
Start with the dot tour — install in 30 seconds, write a 5-line program, see it run. Then layer up.