refactor(api): drop parseLit() on typed-literal SPARQL bindings

[marvin-bot-coasys]

Follow-up to #621 — fixes the failing parseLit.test.ts and simplifies every hard-coded SPARQL call site whose target has flipped from envelope URI to typed XSD literal.

Background

With coasys/ad4m#874 (typed RDF literals refactor), scalar @Property fields with no resolveLanguage are stored as deterministic typed XSD literals rather than literal:string:… envelopes. Oxigraph returns their lexical form directly in SPARQL bindings — parseLit() on those values is a no-op that only happens to still work because Literal.fromUrl("plain string").get() throws and the helper falls through to the raw value.

Also: parseLit in ad4m core (commit 3a96f73a) dropped the result.data extraction branch because signed-envelope literals no longer exist for new property writes.

Changes

packages/api/src/utils/parseLit.test.ts

• Fixes the failing case (extracts .data from JSON literal objects). The new behaviour is to JSON-stringify the whole object (matches ad4m core parseLit). Renamed the test accordingly and documented the new contract.

Removed parseLit() wrappers on typed-literal bindings:

• channel/index.ts — body, title, taskName, transcriptStart in allItems() / unprocessedItems(); isConv in recentConversations()
• conversation-subgroup/index.ts — topicNameRaw, body, title, taskName, transcriptStart, relevanceRaw in topics() / itemsData() / topicsWithRelevance()
• conversation/index.ts — topicNameRaw, nameRaw, summaryRaw, transcriptStart in topics() / subgroupsData()
• topic/index.ts — relevance, channelName in linkedConversations() / linkedSubgroups()
• semantic-relationship/index.ts — channelName in allConversationEmbeddings() / allSubgroupEmbeddings() / allItemEmbeddings() / allItemEmbeddingsByType()

Removed now-unused parseLit imports from all five files.

Left untouched (correct as-is)

• Post.image, Community.name/icon — resolveLanguage: FILE_STORAGE_LANGUAGE, still expression URIs resolved via getExpression()
• Channel.pinnedConversations() SPARQL — already uses <ad4m://fn/parse_literal> server-side, correct for typed literals
• binding.embedding in semantic-relationship — expression URI on the embedding language, resolved via getExpression()

Not touched here

Broader Flux code (Kanban/Table/Graph views, packages/utils, WebRTC, ProfileView) still uses Literal.fromUrl(...).get() on values that come from generic/user data where the storage mode isn't fixed by our model definitions. Those need per-case treatment and are out of scope for this CI-unblock work; happy to follow up in a separate PR.

Verification

Local reproduction of CI (ad4m core built from refactor/drop-resolve-literal-flag, which branches off this PR's ad4m target):

• pnpm --filter @coasys/flux-api test: 7 files, 112 tests green (was 1 failing, 111 passing)
• No new TypeScript errors introduced (the one pre-existing PerspectiveProxy mock-shape error in channel-rename.test.ts is unrelated)
• The stderr "SPARQL error" lines in test output are from tests that deliberately mock querySparql to throw and assert graceful handling — those tests pass.

Diff

6 files changed, +61 / −40

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: b211ad81-12e7-4c29-abb0-73bf9eed2be4

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}