Python as a Hackable Language for Interactive Data Science
summary
Demonstrates concrete hacks: an optional chaining AST rewrite that evaluates a?.b()?.c style chains, a non‑blocking assignment transform that runs expensive assignments in background and only blocks on loads, and a reactive execution system that reruns dependent cells and skips unrelated side effects. Implements these with Piccolo and the ipip flow kernel, shows memorization that makes repeat calls instantaneous, and shows the composed AST transforms actually behave as demoed.
AI-written summary · about
chapters (10)
- 0:00Python for data science
- 1:50Python in interactive data science
- 3:36AST transformations in Python
- 5:24Optional chaining in Python
- 9:01Non-blocking assignments for speed
- 12:18Notebook state with reactivity
- 17:21Runtime introspection essentials
- 19:32Memoization and caching optimization
- 22:05Transformations with Piccolo
- 24:36Conclusion and resources
