refactor(language): a partition is a class of its own rather than a direction

[FBumann]

Prompt: "Should we treat partitions differently? A different class?" — "Do the stacked PR"

Note

The following content was generated by AI.

Stacked on #494.

What this changes

program.Partition holds what a translation, a window and a grouped position read off a relation: along, the key column stepped along; group, the value columns within= named; and joined, the other key columns. Translate.partition, Window.partition and DimensionPositionNode.partition are Partition | None. Direction keeps sum and at alone, and its docstring loses the paragraph that redefined its fields for a partition. RelationNode carries either kind as use, and each consumer asserts the kind its operator takes.

Why

A partition is not a direction. sum and at consume columns and produce columns, so the frame changes and "which way" is the fact. A partition leaves the frame as it is: it steps along one key column and reads the group columns at the frame's own key. So consumed and produced were false names for its two fields, and Direction had to say "for a partition, consumed is…" to be read at all. Walk had the same paragraph.

No consumer took both kinds. Resolution already built them through two functions, and every reader was one or the other. The one shared piece was the typesetter's _relation_read, which reads the relation's name, key and values, so it takes the RelationDeclaration now.

The break

Partition joins program.__all__. The three nodes above no longer hold a Direction, so a consumer reading .consumed or .produced off a partition reads .along and .group instead. lpspec pins math-spec by tag and needs the same follow-up #494 already owes it.

Rebased onto main at bda1fc5

Merged, not replayed: the hard rule here is never to force-push. main came
in at the bottom of the stack (#494) and was carried up through every branch
between, so this one's base is current.

Clean at this level: no conflict, and the suite passes unchanged.

Gates

pixi.sh is refused by this environment's egress proxy, so the gates ran from a uv environment on Python 3.12. That is a departure from the "everything runs in a pixi environment" default.

gate	result
pytest -n auto	1330 passed, 6 skipped
ruff check, ruff format	clean
pyrefly check	0 errors
typos, reuse lint	clean
mkdocs build --strict	clean, with the docs.python.org inventory dropped for the run — the proxy refuses it with a 403
compile-tex	not run, for want of a TeX distribution

Every generator was re-run: neither the schema, nor the golden .out files, nor the generated pages moved. One test moved with the field it reads: test_a_partition_keeps_its_group_when_the_relation_gains_a_value_column asserts .group where it asserted .produced.

🤖 Generated with Claude Code

https://claude.ai/code/session_01UnbTmmYp7auaiCwSgkKKsB

---

Generated by Claude Code

---

Generated by Claude Code

[read-the-docs-community]

Documentation build overview

📚 math-spec | 🛠️ Build #34659743 | 📁 Comparing e2552ad against latest (0d93f01)

  🔍 Preview build  

17 files changed · ± 17 modified

± Modified

• CHANGELOG/index.html
• examples/commitment/index.html
• examples/operators/index.html
• examples/pypsa/index.html
• examples/pypsa_linearized_uc/index.html
• howto/declare-a-column/index.html
• reference/notation/index.html
• reference/language/expressions/index.html
• reference/language/operators/index.html
• reference/math_spec/advice/index.html
• and 7 more...