test(validation): check context equality over every schema, not thirteen

[simontaurus]

Follow-up to #150, which pinned the pyld equivalence on the committed examples only.

What was and was not covered

Corpus	Schemas with a resolvable @context	Was pinned
committed examples	13	yes
committed compliance fixtures	13	no
upstream main	tracks, ~26 today	no

The untested half is the half that matters. The compliance fixtures exist to hold the unusual constructs, so pinning the equivalence on the examples alone tested it against the documents least likely to break it.

Measured before wiring it in: all three corpora agree, 0 divergences, 0 errors.

Changes

• the committed-slice test now sweeps tests/data/oold/compliance/ as well, 13 -> 26 schemas
• a parity test runs the same sweep over OOLD_SCHEMA_DIR, so it tracks main instead of the snapshot the fixture slice was cut from
• the fixture traversal moved to conftest.embedded_schemas so both use one copy

Why this belongs in parity too

context_resolution's docstring argues from what pyld does not return. A snapshot can only hold that against constructs upstream had when it was taken; a new one that breaks the claim appears on main first, and the parity job is the only thing that looks there.

Verification

• pytest tests/test_validation: 498 passed, 7 skipped
• parity with both checkouts: 7 passed
• pre-commit: clean

[github-actions]

Release preview

No version bump from the current commits (stays at v0.20.0). Use conventional commit types (feat, fix, ...) to trigger a release.

Changelog preview (truncated)

_{Preview via python-semantic-release and conventional commits.}

[github-actions]

📊 Benchmark Results

Click to see benchmark comparison

📊 Benchmark Comparison (threshold: 1.3x)
============================================================

➖ Unchanged (within threshold):
  ➖ test_simple_dict_document_store: 0.0016s → 0.0016s (+0.2%)
  ➖ test_sqlite_document_store: 0.0017s → 0.0017s (+0.6%)
  ➖ test_local_sparql_store: 0.0367s → 0.0363s (-1.0%)
  ➖ test_oneof_subschema: 0.0554s → 0.0562s (+1.4%)
  ➖ test_enum_docstrings: 0.0471s → 0.0485s (+2.9%)
  ➖ test_subclass_inheritance: 0.0503s → 0.0535s (+6.4%)
  ➖ test_class_hierarchy: 0.0476s → 0.0481s (+1.1%)
  ➖ test_core[v1]: 0.0354s → 0.0362s (+2.3%)
  ➖ test_core[v2]: 0.0407s → 0.0420s (+3.1%)
  ➖ test_schema_generation[v1]: 0.0015s → 0.0015s (+1.8%)
  ➖ test_schema_generation[v2]: 0.0026s → 0.0026s (+1.0%)
  ➖ test_simple_json: 0.0006s → 0.0006s (+3.4%)
  ➖ test_complex_graph: 0.0014s → 0.0015s (+0.8%)

============================================================
Summary: 0 regressions, 0 improvements, 13 unchanged
============================================================

✅ No significant performance regressions

Threshold: 1.3x (30% slower triggers a regression warning)

Note: Benchmarks are informational only and won't fail the build.

_{💡 Tip: Download the benchmark-results artifact for detailed JSON data}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!