Add whole-document schema validation specs

[HassanAkbar]

Round-trip specs prove the gem keeps what it read. They cannot prove the output is a legal STS document — a model that always drops the same element round-trips fine and still fails every schema.

Validates 4 fixtures against the NISO Extended XSD and ISOSTS.xsd, checking both the shipped file and the re-serialized output. The shipped file is the control; without it a passing round-trip proves nothing.

Adds spec/support/schemas.rb to compile and cache the schemas, and moves spec/iso_sts/schema_validation_spec.rb onto it — that spec was recompiling a 0.5s schema in each of its 11 examples and calling Dir.chdir. Its assertions are unchanged.

Parsing is strict. libxml repairs damage by default, so a truncated file would otherwise compile clean and pass everything.

nokogiri moves into the Gemfile, since the specs call Nokogiri::XML::Schema directly. It stays out of the gemspec — lib/ does not use it.

Interchange and NISO STS 1.2 are not covered. Neither schema is vendored, and a stub test for a missing schema is fake coverage.

bundle exec rspec: 5210 examples, 0 failures, 2 pending. bundle exec rubocop: clean. Runtime grows about 23s, nearly all of it re-parsing two 1MB fixtures.

The 2 pending are tbx-nisosts-0.2.xml, invalid as shipped — a <mml:semantics> with no <mml:annotation>. Raw and round-tripped output give the same error, so the fixture is at fault, not the model.

Ticks the schema validation box in TODO.sts-refactor/10-test-coverage.md.

Metanorma PR checklist

• Breaking changes (list related PRs)
• Documentation update required (create task for this)
• External dependency introduced (documentation update need)
• Gem with native library introduced

[Copilot]

Pull request overview

Adds whole-document XML schema compliance checks to the spec suite to ensure serialized output is actually valid against the vendored STS schemas (not just round-trip stable), while also improving schema-validation spec performance by compiling schemas once per process.

Changes:

• Introduces whole-document schema validation specs for key NISO STS Extended and ISOSTS fixtures (including validating both shipped fixtures and re-serialized output).
• Adds a shared StsSchemas helper that compiles and memoizes Nokogiri schemas and enforces strict XML parsing to avoid libxml recovery masking invalid input.
• Refactors the existing IsoSts fragment schema-validation spec to use the shared schema helper and avoids per-example schema recompilation / Dir.chdir.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
TODO.sts-refactor/10-test-coverage.md	Marks schema validation tests as completed and documents the intentional scope limits.
spec/support/schemas.rb	Adds memoized schema compilation and strict parsing helpers for consistent schema validation across specs.
spec/sts_schema_compliance_spec.rb	Adds whole-document schema validation against vendored NISO Extended and ISOSTS schemas (fixture + re-serialized output).
spec/iso_sts/schema_validation_spec.rb	Switches existing fragment-level ISOSTS validation to the shared schema helper (no assertion changes).
Gemfile	Adds nokogiri for direct Nokogiri::XML::Schema use in the spec suite.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.