CSE 583 : Programming Abstractions

Teaching Assistant

Topics and (tentative) plan

Topics in this list will appear as crossed-out, as the semester progresses and we cover the topics in class.

Module 0: Foundations

1. Programming Abstractions for Data Scientists: overview and philosophy, and case studies

2. The Python You Think You Know: dynamic typing, reference semantics, common bugs and design choices, type hints

3. Introduction to OCaml (the "Laboratory" of programming): let bindings, expressions vs statements, REPL-driven development, simple data types

4. OCaml Types and Pattern Matching Basics: records, algebraic data types, type inference

Assignment 1 released

Module 1: Reliable Pipelines

5. Pattern Matching Deep Dive: exhaustiveness, nested patterns, guards; connection to structural pattern matching in Python

6. Lists and Recursion in OCaml: the parametrized list, recursive functions (map, reduce, and filter), tail recursion and stack overflows

7. Python Type Hints and mypy: type hints in Python code; generics, optional, union; mypy usage; limitations of Python's type system

8. Discussion of Assignment 1

Assignment 1 due

9. EXAM I

Assignment 2 released

Module 2: Functional Foundations

10. Buiding Typed Pipelines in Python: dataclass and Pydantic; type-safe pipeline composition

11. First-class Functions: functions as values, higher-order functions, anonymous functions

12. Closures and Lexical Scope: closures in OCaml and Python; partial application; parametrized transformations

13. Lambda Calculus Essentials: syntax and beta-reduction, connection to functional programming

14. From Lambda Calculus to Python: evaluation strategies, currying in practice, rewriting loops with higher-order functions

15. Discussion of Assignment 2

Assignment 2 due

16. EXAM II

17. Immutability: mutable vs immutable structures; parallel computing semantics

18. Functional Python: higher-order functions in practice; functional tools; operations as transformations; performance considerations

19. Practical Functional Data Wrangling: method chaining in pandas; error handling in pipelines

Module 3: Domain-Specific Languages (DSLs)

20. Introduction to DSLs: pandas syntax; internal vs external DSLs

Assignment 3 released

21. Parsing Basics: grammars and abstract syntax trees (ASTs); a simple recursive descent parser in OCaml; Python's ast module

22. Interpreters and Evaluators: from AST to execution; writing a simple interpreter; environments and binding

23. DSL Design: deconstructing pandas syntax; magic methods and operator overloading; domain-aligned abstractions; deferred or symbolic evaluation

24. Discussion of Assignment 3

Assignment 3 due

25. EXAM III

Module 4: Evaluation Strategies

26. Evaluation Strategies: eager vs lazy evaluation; parameter passing; OCaml's Lazy module; lazy evaluation for big data

Assignment 4 released

27. Generators and Iterators in Python: the Iterator pattern; memory-efficient code; vectorization vs loops

28. (Part 1) Lazy Evaluation in Data Systems: Spark's lazy evaluation; pandas and dask + (Part 2) Discussion of Assignment 4

Assignment 4 due

FINAL EXAM