feat(profile): DCAT-AP v3 SHACL validator via pyshacl + multi-resource framework

[jqnatividad]

Summary

Closes the final §6 follow-up from profile3-handoff.md. DCAT-AP v3 ships SHACL constraints upstream (not JSON Schema), so qsv's built-in JSON-Schema validator can't validate emitted DCAT-AP v3 blocks. This PR generalizes the validation.external framework from #3911 to support multi-file validators and wires up pyshacl as the canonical DCAT-AP v3 reference validator.

Framework generalization

• New ExternalValidatorResource struct + resources: Vec<...> on ExternalValidator. Each entry declares a logical embedded name resolved at spawn time against EMBEDDED_RESOURCES (include_str!-bundled), plus a token name used as {<name>} in args.
• resolve_args / substitute_tokens rewritten for multi-token substitution. {file} keeps its "append if absent" fallback; named tokens ({shapes}, etc.) substitute in place; unknown {tokens} pass through verbatim. OsString path-fidelity from feat(profile): Croissant mlcroissant validator + generic external validator framework #3911 preserved.
• Reserved name: "file" and unknown embedded values rejected with Severity::Required warnings so a misconfigured profile fails fast.

Vendored SHACL bundle

• resources/dcat-ap-v3/shacl/dcat-ap-SHACL.ttl (176KB, 2321 lines) — verbatim from SEMICeu DCAT-AP v3.0.0 at https://github.com/SEMICeu/DCAT-AP/blob/master/releases/3.0.0/shacl/dcat-ap-SHACL.ttl
• resources/dcat-ap-v3/shacl/README.md — source + re-vendor procedure.

Profile wiring

• resources/profiles/dcat-ap-v3.yaml's validation.external declares pyshacl with args: [-s, {shapes}, -sf, turtle, -df, json-ld, -f, human, {file}], references the embedded shapes via resources: [{name: shapes, embedded: dcat-ap-v3-shacl-shapes, suffix: .ttl}], and carries an install_hint pointing at pip install pyshacl (https://github.com/RDFLib/pySHACL).

End-to-end verified. On a host without pyshacl installed:

$ qsv profile <input> --profile dcat-ap-v3 --validate-dcat -o out.json
$ jq '.dcat_warnings[] | select(.field=="external_validate")' out.json
{
  "field": "external_validate",
  "severity": "info",
  "message": "\`pyshacl\` not installed; skipped dcat-ap-v3 validation. Install: pip install pyshacl (https://github.com/RDFLib/pySHACL)"
}

When pyshacl IS installed, qsv writes the rendered JSON-LD AND the embedded shapes to tempfiles, spawns pyshacl with both substituted, and surfaces findings as ProjectionWarnings with the stable external_validate field + pyshacl: <line> message format from #3911.

Test plan

• cargo test --bin qsv -F profile,feature_capable cmd::profile:: — 161 unit tests pass (7 new in external_validate::tests + extended embedded smoke for DCAT-AP v3)
• cargo test --test tests -F profile,feature_capable -- test_profile:: — 53 integration tests pass (incl. trust-gate coverage from feat(profile): Croissant mlcroissant validator + generic external validator framework #3911)
• All 4 binaries build clean: qsv, qsvmcp, qsvlite, qsvdp
• cargo clippy --bin qsv -F profile,feature_capable — no new warnings
• cargo +nightly fmt applied
• python3 scripts/docs-drift-check.py — no drift
• resources/profiles/README.md updated with resources config + DCAT-AP example + instructions for adding new EMBEDDED_RESOURCES entries
• Vendored 176KB SHACL bundle includes README.md with source + re-vendor procedure (mirrors the GSA bundle pattern)

Notes

With this PR, all 5 §6 follow-ups from PR #3908's handoff are complete:

1. ✅ CI gate for embedded profile dry_compile (feat(profile): YAML projection follow-ups — CI gate, per-dist merge, catalog inheritance #3910)
2. ✅ Per-distribution identity-based discovery merge (feat(profile): YAML projection follow-ups — CI gate, per-dist merge, catalog inheritance #3910)
3. ✅ Template-driven catalog inheritance (feat(profile): YAML projection follow-ups — CI gate, per-dist merge, catalog inheritance #3910)
4. ✅ Croissant mlcroissant validator + external-validator framework (feat(profile): Croissant mlcroissant validator + generic external validator framework #3911)
5. ✅ SHACL backend for DCAT-AP v3 (this PR)

🤖 Generated with Claude Code

[codacy-production]

Up to standards ✅

🟢 Issues 0 issues

Results:
0 new issues

View in Codacy

_{NEW Get contextual insights on your PRs based on Codacy's metrics, along with PR and Jira context, without leaving GitHub. Enable AI reviewer
TIP This summary will be updated as you push new changes.}

[Copilot]

Pull request overview

Generalizes the validation.external framework introduced in #3911 to support multiple input files (beyond the implicit rendered JSON-LD), and wires DCAT-AP v3 to pyshacl using a vendored SEMICeu DCAT-AP 3.0.0 SHACL bundle bundled into the qsv binary via include_str!.

Changes:

• Add ExternalValidatorResource + resources: Vec<...> on ExternalValidator; resolve_args/substitute_tokens rewritten for multi-token substitution with reserved-name + unknown-embedded validation.
• Vendor resources/dcat-ap-v3/shacl/dcat-ap-SHACL.ttl (2321 lines) and register a single EMBEDDED_RESOURCES entry; documented re-vendor procedure.
• Wire dcat-ap-v3.yaml's validation.external to pyshacl with {shapes}/{file} tokens; update README and unit/integration tests.

Reviewed changes

Copilot reviewed 5 out of 6 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
src/cmd/profile/external_validate.rs	Adds EMBEDDED_RESOURCES table, resource validation, multi-token substitute_tokens, and tests.
src/cmd/profile/profile_spec.rs	Adds ExternalValidatorResource struct + resources field; extends embedded DCAT-AP v3 smoke test.
resources/profiles/dcat-ap-v3.yaml	Wires pyshacl with embedded SHACL shapes via {shapes} token.
resources/profiles/README.md	Documents resources field and DCAT-AP/pyshacl example.
resources/dcat-ap-v3/shacl/dcat-ap-SHACL.ttl	Vendored SEMICeu DCAT-AP v3.0.0 SHACL shapes (2321 lines).
resources/dcat-ap-v3/shacl/README.md	Documents source URL and re-vendoring procedure.