-
Notifications
You must be signed in to change notification settings - Fork 1
Open
Labels
enhancementNew feature or requestNew feature or request
Description
Context
ADR-201 Phase 5 introduced GraphFacade (topology acceleration) and batch hydration (grounding/confidence caching). The API routes now use these optimized paths, but four client applications need verification to ensure they're calling the right endpoints with the right parameters.
This is the biggest architectural change to the query layer so far. All clients should benefit from the acceleration without any client-side changes (the optimization is server-side), but we need to verify:
- Clients request
include_grounding: truewhere appropriate - Clients use the batch-enabled endpoints (not deprecated ones)
- Response shapes haven't changed in ways that break client parsing
Clients to Audit
Web workstation (web/src/api/client.ts)
- Inventory all query-related API calls
- Verify search, connect, related endpoints hit the accelerated routes
- Check response parsing for grounding/confidence fields
MCP server
- Inventory tool implementations that make graph queries
- Verify grounding parameters are passed through
- Check for any hardcoded endpoint paths
CLI (cli/src/)
- Inventory
kg search,kg search connect,kg conceptetc. - Verify they request grounding and use batch-enabled endpoints
- Check response display code handles new fields
FUSE driver (fuse/)
- Inventory graph queries made by the filesystem layer
- Verify query patterns are consistent with API changes
Deliverable
- Document findings per client (what's correct, what needs updating)
- Apply fixes where needed
- Update
.claude/todo-adr-201-graph-accel.mdwith results
References
- Phase 5g in
.claude/todo-adr-201-graph-accel.md - PR feat: unified GraphFacade with graph_accel integration (ADR-201 Phase 5b) #274 (GraphFacade), PR feat: generation-aware hydration caches (ADR-201 Phase 5f) #276 (batch hydration)
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
enhancementNew feature or requestNew feature or request