refactor(mprgp): extract the three moves, and document the solvers by example

[tschm]

What

MPRGP's loop body held the conjugate-gradient, expansion and proportioning moves inline, so the proportioning switch and the stopping test were buried in the numerics of the moves themselves.

• Each move is now its own function over an Iterate = (x, g, p) state.
• _proportional_step makes the feasibility choice between the CG and expansion moves and reports the Hessian products it consumed.
• _mprgp keeps only the loop, the stopping test and the switch; the per-move counters become a dict keyed by the move name the step returns.

Docs

Doctests added to ActiveSetSolver, MPRGP and the nncg.inner module, covering:

• the operator-not-array calling convention,
• the KKT certificate scoring the answer independently,
• the interchangeable inner solvers,
• MPRGP's step counts summing to iterations.

Verification

Behaviour is unchanged — 138 tests pass at 100% line coverage.

🤖 Generated with Claude Code

[Copilot]

Pull request overview

This PR refactors the MPRGP solver implementation to make the three algorithmic “moves” (CG, expansion, proportioning) explicit functions over a shared (x, g, p) iterate state, while also adding doctest-style documentation examples to key public-facing solver entry points.

Changes:

• Extract MPRGP’s CG / expansion / proportioning moves into dedicated helpers and simplify _mprgp to primarily own the loop, stopping test, and proportioning switch.
• Introduce a move-keyed step counter map in MPRGP and keep Hessian product accounting explicit via the extracted step helpers.
• Add doctest examples to ActiveSetSolver, MPRGP, and the nncg.inner module to document the operator interface, KKT certification, and solver interchangeability.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
src/nncg/solver.py	Adds doctest examples illustrating operator usage, KKT certification, fallback reporting, and inner-solver interchangeability.
src/nncg/mprgp.py	Refactors the MPRGP loop by extracting the three moves and selecting CG vs expansion via _proportional_step; adds doctest examples.
src/nncg/inner.py	Adds doctest examples showing how different inner solvers affect the active-set outer loop and how iteration counts behave.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.