Early development roadmap

[lagru]

A rough work in progress roadmap for docstub during its early development phase:

v0.3 (MVP)

Minimum threshold to create stubs for scikit-image.

• Add basic CI tests
  • Add basic CI configuration #8
• Errors and warnings point to parsed source if possible
  • Point to precise line in parsed source for parsing problems #10
• Cache static analysis
  • Cache results of static analysis #15
• Support module attributes, class attributes, and type aliases
  • Support attributes and type aliases #18
• CI testing with mypy
• Report statistics at the end: error count, file count, runtime, ...
  • Basic stats in 0b048f3
• Minimal documentation

v0.x

• Explore parallelization
• Support checking & transforming a single doctype for testing & debugging
• Add support for grammar & transformation extensions
• Evaluate on scikit-image
  • Generate typing stubs with docstub scikit-image/scikit-image#7546
• Evaluate on networkx
• "Dogfooding" -- publish and distribute stubs for docstub
• Support docstring styles other than NumPyDoc (already supported)
• Add automated release infrastructure
• Add website with documentation
• CI testing with other type checkers (Pytype, Pyright / Pylance, Pyre, ...)
• Profile & optimize
  • Evaluate using Lark's LALR(1) parser
• Cache transformed doctypes
• Cache created stubs

Long-term / v1.0

• Actively used by major projects
• Do runtime analysis to find types?
• Support inline type annotations?