Skip to content

Skill idea: migrate-rng — drive the np.random → Generator campaign #38

Description

@mmcky

Parked skill idea, moved from FUTURE-IDEAS.md when that file was retired in favour of tracker issues (2026-08-13 maintenance round).

The idea

Drive the np.random.*Generator migration campaign (QuantEcon/meta#299): ~13 near-identical PRs so far, many lectures remaining. Mechanical rewriting with one defined judgement checkpoint — distinguishing figure changes caused by RNG-stream changes, which are never auto-applied.

Assessment (2026-08-13)

Time-limited by construction — the campaign ends when the lectures are migrated, so the skill is only worth building if the campaign runs long enough to amortize it. Two notes that sharpen the call:

  • The detection half already exists: qe/skills/check-code checks for legacy np.random.* usage by rule ID. What this idea adds is only the mechanical rewrite plus the figure-change judgement checkpoint.
  • The unsettled convention (rng explicit argument vs rng=None fallback — lecture-python.myst#874 vs lecture-python-intro#741) should be settled in the manual regardless of whether the skill is built, and is a prerequisite for building it: a migration skill cannot apply a convention that does not exist.

Recommended disposition: settle the convention upstream first; revisit the skill only if a large tranche of lectures remains after that.

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementImprovement to existing content or functionalitylow-priorityNice to have, no time pressure

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions