The freshly rendered 607-record graph pages in draft #924 present every point as a METPO trait embedding, while the verified data contains 375 direct matches and 232 parent proxies. The two new records (traitmech:000211 DISARM and traitmech:000212 CBASS) both use the parent node METPO:1000188. Their metadata is correct, but a reader cannot see that distinction in the graph UI.
Reproduced against the held main integration fc7a4645911fe539f84f585e7b83fbf699c2ace8 + d227d23913ff3d99e2a369e79a2194e80601d71e, with the independently verified paired graph candidate receipt SHA 1a76cd759871a0b04194f0223629d3d9bf40796bce8fcd85515a247b80dd19eb:
pages/umap.html says “607 METPO trait embeddings”; pages/graph.html makes the same population claim.- Both tooltips show only the point label, identifier, and category. Neither generated page exposes
match_method, kgm_nodes, or a parent-proxy explanation.
- Both JSON arrays already contain the correct
match_method and kgm_nodes fields. The ledger and input vectors are valid; the gap is presentation.
Acceptance:
- Update both maintained templates (
src/traitmech/templates/umap.html and graph.html) to display direct-match and parent-proxy counts, derived from the verified current data rather than hard-coded totals.
- Clearly explain that parent proxies place a record using the named parent node's graph embedding. Show each hovered point's match method and source-node IDs using safe text/DOM insertion.
- Preserve the common BGE view as primary and retain both specialist graph views. Do not change coordinates, populations, neighbors, source receipts, or numerical runtime.
- Add focused tests that distinguish direct and parent-proxy points, verify source labels and escaping, and reject missing/incorrect disclosures. Rerender the maintained site and verify every route, then complete the existing independent site review.
This is a focused presentation follow-up to #917 and #942. No new model inference or graph source stream is required. Existing-issue search included open and closed issues for proxy, embedding provenance, and graph tooltip; no issue specifically addressed this UI omission.
The freshly rendered 607-record graph pages in draft #924 present every point as a METPO trait embedding, while the verified data contains 375 direct matches and 232 parent proxies. The two new records (
traitmech:000211DISARM andtraitmech:000212CBASS) both use the parent nodeMETPO:1000188. Their metadata is correct, but a reader cannot see that distinction in the graph UI.Reproduced against the held main integration
fc7a4645911fe539f84f585e7b83fbf699c2ace8+d227d23913ff3d99e2a369e79a2194e80601d71e, with the independently verified paired graph candidate receipt SHA1a76cd759871a0b04194f0223629d3d9bf40796bce8fcd85515a247b80dd19eb:pages/umap.htmlsays “607 METPO trait embeddings”;pages/graph.htmlmakes the same population claim.match_method,kgm_nodes, or a parent-proxy explanation.match_methodandkgm_nodesfields. The ledger and input vectors are valid; the gap is presentation.Acceptance:
src/traitmech/templates/umap.htmlandgraph.html) to display direct-match and parent-proxy counts, derived from the verified current data rather than hard-coded totals.This is a focused presentation follow-up to #917 and #942. No new model inference or graph source stream is required. Existing-issue search included open and closed issues for proxy, embedding provenance, and graph tooltip; no issue specifically addressed this UI omission.