Develop - #341
Draft
namedgraph wants to merge 7 commits into
Draft
Conversation
* Serve raw ontology graphs without RDFS inference Resolve the application ontology's owl:imports closure natively via ontapi (OntModelFactory.createModel over a ScopedGraphRepository view) instead of manually flattening the closure and materializing an RDFS-inferred model. No inference is applied anymore — every consumer (constructor/constraint inheritance, client-side (rdfs:subClassOf)* queries) traverses hierarchies explicitly. rdfs:Class terms are promoted to owl:Class in a separate union member so no document graph is polluted; closure union graphs are cached in a bounded map on Application, keyed by ontology URI. The shared repository now only ever holds raw per-document graphs, which ProxyRequestFilter serves directly for closure documents — asserted triples only, identical to a direct document GET. The DESCRIBE fallback over the in-memory closure (terms minted in external namespaces) is inference-free as well. This fixes inferred rdf:type rdfs:Resource leaking into proxied namespace documents, where the extra type produced multi-token @typeof and silently degraded View blocks to a generic property list. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * Fix ScopedGraphRepository.contains() dropping resolvable imports ontapi consults GraphRepository.contains() before get() when resolving an ontology's imports closure. PrefixGraphRepository.contains() reports cache state (loaded graphs), not resolvability, so every import that resolves through a bundled location mapping (dh, sp, spin, foaf, sioc, sd), SPARQL-first loading or HTTP was answered with false on first resolution — and ontapi silently substituted an empty ontology graph for it (its ignoreUnresolvedImports fallback). The closure kept its shape but lost the content of every such import: SPIN constraints vanished, so validation enforced nothing (422 tests wrote through, eventually applying invalid dataspace settings and cascading into NPEs), and vocabulary term lookups came up empty. This is what failed the HTTP test suite in CI. contains() now attempts resolution through the backing repository (which loads and caches the graph) after the cache checks, reporting absent only for genuinely unresolvable ids. Adds a production-shaped regression test: ns# ontology importing the SPARQL-seeded ldh# vocabulary with its transitive imports resolved through the real bundled location mappings. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * Alias only repository-held closure ids under their document URIs ontapi keys imports under their declared ontology IRIs, which need not be repository entries: a content-addressed upload is cached under its uploads/ URI while declaring a foreign ontology IRI. The doc-URI aliasing loop called repository.get() on such declared IRIs, which fell through to an HTTP dereference of the foreign IRI (e.g. https://example.org/test) during ontology load — failing the load and the ontology-import-upload-no-deadlock HTTP test. Guard the loop with isCached() so only graphs the shared repository actually holds get aliased. Adds a mismatched-IRI import case to the closure regression test. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * Fix GET-proxied-ontology-ns.sh assertions for closure DESCRIBE semantics The admin ontologies/namespace/ document stores the ontology but the closure keys it under the <ns#> ontology URI, so a proxied GET of the document URI is answered by the closure DESCRIBE fallback — the document's own #-fragment term descriptions — not the raw graph branch. Assert on a class minted in the document's hash namespace (mirroring the original #related_View regression) instead of the made-up-namespace classes, which only appear under their own namespace document URI. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * Remove Linked Data proxy DESCRIBE-over-closure fallback The proxy is document-keyed transport; term lookups over the ontology closure belong on /ns (SPARQL), not on a DESCRIBE synthesized from the proxy target URI. Drop the fallback and its ParameterizedSparqlString/ QueryExecution imports; the isCached closure-cache branch (raw per-doc graphs, no inference) still serves closure documents. Realign tests: delete GET-proxied-ontology-ns.sh (it only exercised the removed DESCRIBE path and dereferenced ontology terms via the admin doc URI, which was never a supported path). Add GET-ns-no-query.sh (raw /ns ontology graph, no rdfs:Resource leak) and GET-proxied-mapped-vocab.sh (dct:title, foaf:Person, skos:Concept served from the static prefix mapping; skos also covers fragment-strip + xml:base resolution). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Make the Linked Data proxy dumb: drop ontology-closure serving Remove the getOntology()/isCached branch that answered ?uri= requests from the app ontology owl:imports closure cache, plus the now-unused ontology injection, getOntology(), and PrefixGraphRepository/Application/EndUserApplication/ OntModel imports. The proxy is now dumb transport: bundled-vocab file cache (isMapped) + SSRF-checked external fetch. Ontology terms are served by /ns. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Fix GET-proxied-mapped-vocab SIGPIPE on large vocab responses The assertions piped the whole vocabulary graph through `echo | grep -q`. `grep -q` exits on first match and closes the pipe while `echo` is still writing; with `set -o pipefail` the SIGPIPE'd `echo` (write error: broken pipe, exit 141) fails the pipeline whenever the response exceeds the ~64 KiB pipe buffer — so a *successful* match killed the test. Only this test trips it, being the only one that returns entire vocabulary documents (40-113 KiB). Read from a here-string instead (temp file, no pipe to break), and match the full language-tagged label literal the bundled documents actually carry. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
) The four view controls (order-by dropdown, facet headers, parallax properties, rdf:type facet values) resolved predicate/class labels by proxying each term's vocabulary document (GET ?uri=<ac:document-uri(term)>), then applying ac:label to the response. With the proxy no longer serving ontology terms, route these through the app's /ns SPARQL endpoint instead. - view-results chain: add a property-metadata load step (mirroring the existing object-metadata step) that DESCRIBEs the result predicates over /ns before ldh:render-view, threading $property-metadata through to the order-by and facet-header rendering. - order-by + facet headers: render labels synchronously from $property-metadata via ac:label (local-name fallback for terms absent from the closure); drop the per-predicate proxy promises and the ldh:order-by-response / ldh:facet-filter-response handlers. - parallax + rdf:type facet values: swap each per-term proxy GET for a POST /ns DESCRIBE $Type VALUES { <term> }; the response is still application/rdf+xml so the existing handlers are unchanged. Verified against the unesco-thesaurus demo: on /concepts/concept1002/ the redundant vocab-document proxy fetches drop (skos/core 9->5, rdf 2->1, prov 2->1), the SKOS view predicates now resolve via /ns, and the 429/502 the redundant load triggered are gone. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
POST-proxied-external-query.sh proxied a live query to dbpedia.org/sparql, whose public endpoint intermittently returns 502 in CI. Rename it to POST-proxied-cross-origin-query.sh and point it at the admin app's SPARQL endpoint (a different origin than the end-user app), so it still exercises the ProxyRequestFilter remote-fetch + SPARQL-results re-serialization path but runs deterministically against a local target. Use the owner cert since the admin endpoint is ACL-protected, and assert the response is actual SPARQL results rather than only a 200 status. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Replace /add and /generate endpoints with client-orchestrated graph writes Move the "Add data" and "Generate containers" orchestration to the client over the uniform Graph Store Protocol interface, removing the server-side fetch/SSRF surface (pen-test LNK-002) and collapsing two special-purpose endpoints into standard per-document writes. - Add data: the submit handler splits by the presence of a spin:query input. The add/clone variant fetches dct:source through the same-origin ?uri= proxy as RDF/XML (CORS + Jena format conversion) and POST-appends it to the sd:name target (ldh:add-data-source-response / ldh:add-data-form-error). The import-ontology variant is unchanged and still RDF/POSTs to /transform. - Generate containers: builds one container document per checked class and PUTs them via parallel ixsl:all fan-out (ldh:generate-containers-fanout -> ldh:generate-containers-join), seeded by ixsl:resolve so the requests run in an active promise context. The view block is now correctly wrapped as ldh:Object -> rdf:value -> ldh:View (the endpoint's bare ldh:View bypassed ldh:InvalidContentBlockType validation). - Delete Add.java, Generate.java and their Dispatcher locators. /transform is retained until a client-side SPARQL engine lands. - http-tests: drop the obsolete /add and /generate tests (the system/ suite is unregistered); add add/GET-proxied-source-POST-append.sh and add/PUT-generate-container.sh. CHANGELOG updated. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Fix generate-containers dialog: parent init, disabled Generate, async schema load - Initialise the parent typeahead with the current container (ldh:LoadTypeaheads from ldh:base-uri, matching btn-save-as), which was lost when the form stopped carrying a source param. - Disable the Generate button until the schema is loaded; enable it in ldh:endpoint-classes-response once the class list is populated. - Make Load schema fully async: the service-endpoint resolution used a blocking document() fetch (the "TO-DO: asynchronous request"). Replaced with an ixsl:resolve -> ldh:load-schema-endpoint -> ldh:load-schema-results promise chain, so the request no longer blocks and the progress cursor shows. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Add RDF data: seed target graph from the local dataspace, reject remote targets The "Add RDF data" dialog seeded its target graph typeahead from ldh:base-uri(.), which resolves to the proxied remote resource when viewing one (ac:document-uri(ac:uri())). So adding data while viewing a proxied document defaulted the write target to that remote and proxied the append to it (403). The source is still ldh:base-uri (the remote being imported). - btn-save-as now seeds the graph typeahead from ac:absolute-path(ldh:request-uri()) (the local browser location), which is always a local dataspace document. - Guard the append: a cross-origin target renders an inline error instead of proxying the write. ldh:add-data-form-error now tolerates a missing response so pre-fetch validation can reuse it. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Condense Unreleased CHANGELOG entries to one line each Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Merge /ns ontology labels into client-side object-metadata Object labels for rdf:type (class) objects were missing because $object-metadata-query only queried the instance SPARQL endpoint (sd:endpoint() -> /sparql); class labels live in the ontology served at /ns (in-memory OntModel), not in instance data. So class objects fell through ac:object-label to their URI fragment. Object-metadata is now the union of labels from /sparql (instances) and /ns (ontology/classes), merged client-side into the single $object-metadata document ac:object-label already consumes. No SERVICE federation and no Java: ARQ's SERVICE bypasses LDH's proxy/truststore and a query to /sparql is executed by an isolated Fuseki container that can't reach the app's public /ns, so a two-request merge is used instead, mirroring the existing property-metadata -> /ns pattern. - client.xsl: add $object-metadata-ns-query (the label CONSTRUCT without the GRAPH ?graph wrapper, since /ns serves the ontology in the default graph); wire the /ns fetch + merge into the document-load chain. - merge-rdfxml.xsl: add ldh:merge-metadata(), delegating to the existing ldh:MergeRDF mode (guards an absent side with empty rdf:RDF). - block.xsl: ldh:load-object-metadata now builds both /sparql and /ns requests; add ldh:set-object-metadata-ns + ldh:merge-object-metadata; wire the ontology-view chain. - view.xsl: wire the view-results chain. - form.xsl / modal.xsl: add the /ns tuple + merge step to all five ldh:fire-load-set-parallel chains. - object.xsl: build the /ns request, fire it in the metadata thunk, merge both bodies in the response handler. ac:object-label and its $object-metadata tunnel are unchanged; only the assembly of that document changes. Server-side (SSR) assembly is left untouched: ldh:rdf-document-response replaces the SSR-rendered document-body on load, so the visible labels come from the client merge. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Merge /ns ontology labels into server-side object-metadata (initial paint) The client render already merges instance labels (/sparql) with ontology/class labels (/ns) for object URIs, but the initial server-rendered paint (layout.xsl, the counterpart of client.xsl's document-load chain) fetched object-metadata from sd:endpoint() only. So on first paint an rdf:type/class object showed its URI fragment, snapping to the real label once the client re-rendered the body. Give the server's first paint the same merged metadata: - declare $object-metadata-ns-query (the label CONSTRUCT without the GRAPH ?graph wrapper, since /ns serves the ontology in the default graph), mirroring client.xsl; - include client/merge-rdfxml.xsl so ldh:merge-metadata() + the ldh:MergeRDF mode are available server-side (pure XSLT, no ixsl); - the initial-paint $object-metadata variable now fires both requests (each in its own try/catch) and merges them via ldh:merge-metadata, passing the merged document as the bs2:TabBody object-metadata tunnel. Verified: layout.xsl compiles clean under Saxon-HE (product=SAXON, the server-side processor) — a full execution plan is produced with no static errors, matching the pre-edit baseline. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.