docs(examples): the PyPSA file states a constraint once where PyPSA builds one row set, rather than a block per regime

[FBumann]

Prompt: "I now merged 257!" — then, on finding it had not reached main:
"Fix that! Reopen onto main!"

Note

The following content was generated by AI.

This is #257's diff, unchanged, onto main.
git cherry-pick of its merge commit 5233263 applies clean — six files, examples/
and docs/ only, no src/ — and the original authorship is preserved.

#257 was stacked on feat/cases-proved-apart, the branch carrying #168. #168 merged
to main on its own, which left that branch with no route there; merging #257 into it
put the port on an orphaned base. main still has zero cases: in
examples/pypsa.yaml, so no release can carry the port and v0.0.0-alpha.52 cannot be
cut. Nothing is wrong with the change — only with where it landed.

The branch it merged into is now 2 ahead / 5 behind main; it can be deleted once this
lands.

What it is, unchanged from #257

Twelve PyPSA names — sixteen counted once per file — were stated by two to four blocks
apiece, split by a regime the language could not name inside a quantity. Each is now one
block.

	pypsa.yaml	pypsa_linearized_uc.yaml
constraint blocks	116 → 98	25 → 21
names stated by more than one block	17 → 5	4 → 0

The five that remain are the GlobalConstraint sense-splits, where the split is in the
operator rather than in a value — #239, closed not planned, and untouched here.

Verified on this base

• pixi run ci green on main + this commit: lint, 899 tests, mkdocs build --strict, 27 compiled documents. Tree clean after the run.

• The downstream ladder was run against this branch, which is the check docs(examples): the PyPSA file states a constraint once where PyPSA builds one row set, rather than a block per regime #257's own
  symbolic argument could not make — lpspec's differential/pypsa/parity.py builds each
  rung through both engines and compares against live PyPSA 1.3.0. Rung 2 collapses
  exactly as intended: StorageUnit-energy_balance and Store-energy_balance go from
  two recorded structure deviations to none, at an unchanged objective and unchanged
  row, column and dual counts.

  before   rung_02_storage: MATCH · 2 of 26 names differ · duals on 103 rows, 2 negated · 25 equal · 2 region
  after    rung_02_storage: MATCH · 103 rows, 48 columns  · duals on 103 rows, 2 negated · 27 equal · 0 region
  

  Rungs 3 and up are not yet confirmed through that runner: lpspec's rung
  projection — a presentation device that cuts the file down to what each rung builds
  — does not understand cases: yet, and its first cut mis-handles a region whose
  when: is negated. That is a consumer-side gap, filed as work on the lpspec side; it
  says nothing about this diff, and pixi run ci here does not depend on it.

Not done here

• No rebase, no squash, no edit — the point is that this is docs(examples): the PyPSA file states a constraint once where PyPSA builds one row set, rather than a block per regime #257's content and not a
  second opinion about it. Review it there if it needs reviewing again.
• feat/cases-proved-apart is left alone rather than force-moved onto main.

[read-the-docs-community]

Documentation build overview

📚 math-spec | 🛠️ Build #34313621 | 📁 Comparing f6c6166 against latest (40a0503)

  🔍 Preview build  

2 files changed

± examples/pypsa/index.html
± examples/pypsa_linearized_uc/index.html