fix(validation): treat an absent @context as every property unmapped

[simontaurus]

Fixes #141.

Before

FAIL roundtrip.generated    N.schema.json: property lost through RDF despite being mapped: v
6 ok, 1 failed, 0 warning(s), 1 skipped

v is not mapped; the schema declares no @context. OOLD-SCH-2d05 says an unmapped term must not be a conformance failure.

After

WARN OOLD-SCH-21d7 context.coverage   N.schema.json: schema declares no @context, so none of its 1 property
                                      will reach RDF: v. Declare @vocab to map them into a default
                                      namespace, or add the terms to x-oold-context.
7 ok, 0 failed, 1 warning(s), 1 skipped

Change

_unmapped_properties split the two cases its docstring already distinguished. A context that cannot be resolved still returns the empty set, because the checks that report resolution failures own that. A context that resolves to nothing is the opposite: no term exists, so every declared property is unmapped.

context.coverage also reports the empty-context case, which it skipped past. Without that, a schema with no @context produced no signal at all once the wrong failure was removed - quieter than a schema that maps nine properties out of ten.

Verification

• pytest: 560 passed, 9 skipped, plus a no_context.schema.json fixture and a regression test
• oold-schema/examples 380 ok / 0 failed and examples/compliance 72 ok / 0 failed, both unchanged
• parity against validate.mjs: 6 passed

Parity failed on main before this branch, for an unrelated reason: the remote meta cache still held a copy of refs/heads/main from before OO-LD/oold-schema#152, so coverage.vocab demanded a fixture for a keyword that no longer exists. oold meta fetch --force clears it. That is #138, noted there.

[github-actions]

Release preview

Merging this PR would release v0.18.2 (current: v0.18.1).

Changelog preview (truncated)

## v0.18.2 (2026-08-29)

### Bug Fixes

- **validation**: Treat an absent @context as every property unmapped
  ([`25cdbf5`](https://github.com/OO-LD/oold-python/commit/25cdbf5ddf074359ed4c1110d83b782761cb6aa1))

_{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: 0.0012s → 0.0012s (-0.2%)
  ➖ test_sqlite_document_store: 0.0013s → 0.0013s (-1.4%)
  ➖ test_local_sparql_store: 0.0299s → 0.0298s (-0.4%)
  ➖ test_oneof_subschema: 0.0478s → 0.0474s (-0.9%)
  ➖ test_enum_docstrings: 0.0429s → 0.0404s (-5.7%)
  ➖ test_subclass_inheritance: 0.0480s → 0.0428s (-10.7%)
  ➖ test_class_hierarchy: 0.0461s → 0.0422s (-8.5%)
  ➖ test_core[v1]: 0.0336s → 0.0305s (-9.1%)
  ➖ test_core[v2]: 0.0410s → 0.0352s (-14.2%)
  ➖ test_schema_generation[v1]: 0.0012s → 0.0012s (-0.6%)
  ➖ test_schema_generation[v2]: 0.0020s → 0.0020s (-0.4%)
  ➖ test_simple_json: 0.0004s → 0.0004s (-0.5%)
  ➖ test_complex_graph: 0.0011s → 0.0011s (-2.6%)

============================================================
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!