plan: OGAR sink-in + consumer-bridge removal (medcare-bridge → UnifiedBridge → delete)#623
Conversation
Session handover for the next session: what shipped (#616 merged, #617 merged, ruff #29 open), the three-layer thesis, the V3 state (OSINT/FMA/ CPIC all wired as of #618 — none by this session), the built-but-unmerged minter (facet_mint @ serene-mayer 360fc72, 8 tests green), and the brick-3 MedCare probe as the single CONJECTURE→FINDING gate. Includes the burn/owl corrections and the sync-first discipline. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01EYvNjD8M8LMNYbRy3gq2FP
Cross-links the parallel V3-identity/genome session (614/615/618, q2/cpic NARS prototype, CYP2D6 falsifier) as the companion half of the arc. Three corrections their handover surfaced: - mint OGIT->codebook is a CROSS-REPO arc (ogar-vocab + contract CODEBOOK + lance-graph-ogar parity together), per ISS-OGAR-GENETICS-MIRROR-PENDING / E-CODEBOOK-MINT-IS-A-CROSS-REPO-ARC — not a local wire. - Canon:Custom flip must be atomic over the COMPLETE V3 set (OSINT+FMA+ CPIC), now unblocked by #618 (I-LEGACY-API-FEATURE-GATED). - Board-hygiene debt logged: #616/#617 shipped without same-commit EPIPHANIES/PR_ARC/LATEST_STATE updates; fold into the next PR alongside the companion's outstanding 614/615/618 rows. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01EYvNjD8M8LMNYbRy3gq2FP
…edBridge -> delete) Codifies the sink-in layering and sequences the final deletion of the deprecated per-consumer bridges. Grounded finding: the migration is ~90% shipped — all 6 *_bridge.rs (incl. medcare) are already #[deprecated] type FooBridge = UnifiedBridge<FooPort> aliases over one generic harness; CONSUMER-BRIDGE-DEPRECATION.md (OGAR#95) is the live doctrine. Sink-in decision: OGAR owns meaning (PortSpec/codebook/ClassView/ActionDef/ ontology); lance-graph-contract owns the zero-dep wire mirror (parity guard, not a dep); lance-graph-ontology owns the OGIT registry; lance-graph-rbac owns grants keyed on the shared lo-u16 concept; lance-graph-ogar keeps only the UnifiedBridge mechanism + parity. medcare recast: reusable patterns (shared HealthcarePort/codebook surface in OGAR) + domain-specific Ontology Schema enrichment (Healthcare 0x09XX / FMA-V3 / CPIC-V3 in OGIT) — never a bridge in the spine. D-SINK-1..5: ratify -> migrate consumers off MedcareBridge (cross-repo) -> delete the 6 aliases (scope-lock tests ported first) -> reconcile the two WoaBridge identities -> sink the ontology (cross-repo codebook-mint arc). Board-hygiene: INTEGRATION_PLANS.md prepended same-commit. Doc-only, PROPOSED, operator-gated. Extends CONSUMER-BRIDGE-DEPRECATION + ogar-vocab-contract-codebook-migration-v1. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01EYvNjD8M8LMNYbRy3gq2FP
Read all 48 OGAR PRs since the #95 anchor (titles) + #95/#114/#137 bodies. Adds a grounding section + corrects two things: - §2 layering is #95's verbatim (classid hi=APP/render : lo=concept; Medcare patient=0x0005_0901). D-SINK-2 now RIDES #95's wave plan (W0 OpenProject/Redmine -> W1 WoA -> W2 SMB -> W3 Odoo -> W4 q2) instead of inventing its own order. - Records what is ALREADY shipped on the OGAR side so the plan sequences the last mile, not a green field: OGIT 1:1 import (#105), HIRO DO arm + the lossless-DO rule (#114, shape [G] / execution CONJECTURE), FMA skeletal spine (#116/#117), classid-driven envelope parser (#128), polyglot transpiler emit_csharp/emit_python consuming ogar-from-ruff (#131/#138), and the COMPILED-ClassView-is-the-spine correction (#137: not SurrealQL, not a runtime interpreter; Askama = membrane emit). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01EYvNjD8M8LMNYbRy3gq2FP
📝 WalkthroughWalkthroughThree documentation files are added: a four-line entry in the integration plans index registering the OGAR sink-in plan, a 236-line plan document (v1) detailing deliverables, layering boundaries, gates, and medcare recast, and a 210-line session handover capturing the AST-address conjecture state, V3 wiring, minter status, and next actions. ChangesOGAR Sink-in Plan and Session Handover
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
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. Comment |
There was a problem hiding this comment.
🧹 Nitpick comments (2)
.claude/plans/ogar-sink-in-and-consumer-bridge-removal-v1.md (2)
67-68: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick winClarify medcare-rs placement in the migration wave.
The wave sequence (W0–W4) lists OpenProject, Redmine, WoA, SMB, Odoo, and q2, but
medcare-rsis cited as a consumer in D-SINK-2 without a wave assignment. If Medcare is already migrated or follows a separate track, state that explicitly; if it belongs in the sequence, add it to the wave list to match OGAR#95's ordering.Also applies to: 160-162
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In @.claude/plans/ogar-sink-in-and-consumer-bridge-removal-v1.md around lines 67 - 68, Clarify where medcare-rs belongs in the migration plan so D-SINK-2 matches the wave ordering described in the document. Update the wave sequence in the planning section to either explicitly place medcare-rs in W0–W4 or state that it is already migrated or tracked separately, and make sure the D-SINK-2 consumer reference for medcare-rs is consistent with that decision.
162-162: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low valueExpand "Iron Rule 5" on first use or add a footnote.
The document cites "Iron Rule 5" as a cross-repo approval gate without immediate context. A brief parenthetical or link to
.claude/knowledge/ogar-consumer-preflight.md(which mentions the "pull-don't-reconstruct iron rule") would help readers who don't have the full rule set memorized.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In @.claude/plans/ogar-sink-in-and-consumer-bridge-removal-v1.md at line 162, The first mention of “Iron Rule 5” is too terse and lacks context for readers. Update the “Cross-repo approval gate” reference in this plan to expand the rule inline with a short parenthetical or add a footnote/link to .claude/knowledge/ogar-consumer-preflight.md, and keep the “Iron Rule 5” wording tied to that section so it’s easy to find if the surrounding text changes.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Nitpick comments:
In @.claude/plans/ogar-sink-in-and-consumer-bridge-removal-v1.md:
- Around line 67-68: Clarify where medcare-rs belongs in the migration plan so
D-SINK-2 matches the wave ordering described in the document. Update the wave
sequence in the planning section to either explicitly place medcare-rs in W0–W4
or state that it is already migrated or tracked separately, and make sure the
D-SINK-2 consumer reference for medcare-rs is consistent with that decision.
- Line 162: The first mention of “Iron Rule 5” is too terse and lacks context
for readers. Update the “Cross-repo approval gate” reference in this plan to
expand the rule inline with a short parenthetical or add a footnote/link to
.claude/knowledge/ogar-consumer-preflight.md, and keep the “Iron Rule 5” wording
tied to that section so it’s easy to find if the surrounding text changes.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro Plus
Run ID: 2c35b62f-42c5-470c-bca0-1478807e09cf
📒 Files selected for processing (3)
.claude/board/INTEGRATION_PLANS.md.claude/handovers/2026-06-26-ast-address-to-next-session-capstone.md.claude/plans/ogar-sink-in-and-consumer-bridge-removal-v1.md
What
A doc-only plan that codifies the sink-in layering (what lives in OGAR / lance-graph-ogar vs lance-graph-contract / lance-graph-rbac) and sequences the final removal of the deprecated per-consumer bridges — recasting the medcare domain from a bridge into reusable patterns + domain-specific Ontology Schema enrichment. Plus two session handovers (rebased onto current main).
Grounded against the actual OGAR PR arc #95→#142 (all 48 read at title level; #95/#114/#137 read in full).
Grounded findings (§1.5)
The migration is ~90% shipped, on both sides:
medcare_bridge.rs) are already#[deprecated] type FooBridge = UnifiedBridge<FooPort>aliases over one generic harness.APP-CLASS-CODEBOOK-LAYOUT(feat: BGE-M3 BF16 HDR lens + multi-lens voting — 99 tests bge_m3_lens.rs: second precision lens from BGE-M3 BF16 GGUF (dtype=30) - BF16→f32 via one shift: f32::from_bits((u16 as u32) << 16) - 256×256 HDR table std=73.6, CLAM 256 centroids - vote_distance(): compare Jina vs BGE-M3, return agreement 0.0-1.0 - 5 tests (size, diagonal, variance, vote) data/bge-m3-hdr/: 64 KB table + 488 KB index baked in Both lenses from same XLM-RoBERTa base, different training: Jina F16: cos[-0.067, 0.234], std=73.6 BGE-M3 BF16: cos[-0.090, 0.248], std=73.6 Multi-lens agreement → NARS confidence boost. Jina reranker v3 BF16 downloading for relevance precision lens. Reranker = cross-encoder relevance score, not embedding distance. Could gate cascade transitions: "is this pair actually relevant?" 99 tests pass. https://claude.ai/code/session_01ChLvBfpJS8dQhHxRD4pYNp #95, classid[hi APP/render][lo concept], Medcarepatient = 0x0005_0901), the wave-ordered consumer migration (W0 OpenProject/Redmine → W4 q2), OGIT 1:1 import (feat: ocr_pipeline example — full tesseract→thinking→AriGraph→NARS→export #105), the HIRO DO arm + lossless-DO rule (docs: calibration session handover — H1-H5 hypotheses, Cronbach α, full protocol 5 testable hypotheses: H1: BF16 truncation flips ~5% of ranks (bucket boundary effect) H2: γ+φ encoding preserves more rank order than linear CDF H3: i8 signed > u8 unsigned for gate-heavy roles specifically H4: ICC profile correction brings ALL encoding paths to ρ > 0.998 H5: Cronbach α reveals which tasks need multi-lens vs single lens Testing protocol: Phase 1: ONNX f32 ground truth (rten, Jina v5, 1000 pairs) Phase 2: BF16 baseline (stream GGUF, same CLAM) Phase 3: 5 encoding paths (linear, γ+φ, i8, γ+φ signed, spiral) Phase 4: Spearman ρ before/after ICC per path Phase 5: Cronbach α across 6 lenses Synthesis matrix: which encoding × which role × ICC or not. Estimated: 3-4 hours. Validates everything built in 67+ commits. https://claude.ai/code/session_01ChLvBfpJS8dQhHxRD4pYNp #114, shape[G]/ execution CONJECTURE), FMA skeletal spine (feat: pooling strategies + builder pattern + commit sinks (EmbedAnything patterns) #116/feat: 6 EmbedAnything patterns — auto-detect, chunker, tensor bridge + practices #117), the classid-driven envelope parser (feat: Phase 3 BF16 wiring + φ-spiral reconstruction theory #128), the polyglot transpileremit_csharp/emit_pythonconsumingogar-from-ruff(feat: Jina v5 ONNX + ModernBERT ONNX downloaded (gitignored, local only) #131/feat: FORWARD PASS WORKS + playground + 4096×16 branch architecture (no KV cache) #138), and the correction that the spine is the COMPILED ClassView, not SurrealQL (feat: resonance thinking + i8 cognitive markers + gate stacking + meta-awareness #137).The decision (§2 — #95's layout, verbatim)
ActionHandler/auth_store 0x0B01).UnifiedBridge<P>mechanism +parity.The medcare recast (§3)
The medcare domain dissolves into reusable patterns (the shared
HealthcarePort::class_id/ convergent codebook surface in OGAR) + domain-specific Ontology Schema enrichment (Healthcare0x09XX/ FMA-V30x1000_0A01/ CPIC-V30x1000_0E00in OGIT) — never a bridge in the spine.Deliverables (operator-gated)
D-SINK-1 ratify layering → D-SINK-2 migrate consumers off the aliases, riding #95's wave plan (cross-repo, Iron Rule 5) → D-SINK-3 delete the 6 aliases (scope-lock tests ported first, per
I-LEGACY-API-FEATURE-GATED) → D-SINK-4 reconcile the twoWoaBridgeidentities → D-SINK-5 sink the ontology (the cross-repo codebook-mint arc).Gates
No deletion before consumers migrate (OGAR#95: "nothing removed; deletion lands later"). Intersects the operator-gated classid realign (D-OVC) + Canon:Custom. Extends
CONSUMER-BRIDGE-DEPRECATION.md+ogar-vocab-contract-codebook-migration-v1; does not supersede them.Scope
Doc-only — the plan + an
INTEGRATION_PLANS.mdprepend (board-hygiene, same commit) + two handover docs. No code, no contract surface touched. PROPOSED.🤖 Generated with Claude Code
Generated by Claude Code
Summary by CodeRabbit