OSINT cockpit: dissolve property-node islands → tenant, reason dual-use as a Person×Situation causal chain#69
Merged
Conversation
The basin belongs in the node's own detail (GUID family byte + EdgeBlock mixin adapters), not in an outsourced helper node. A materialized hub — whether a SOA_HUB_CLASS basin node or a SchemaValue property-node — cannot dock as an edge; it only invites 'look up a property as a node', which fragments the graph into islands. Backend (osint_soa_bytes): - Stop emitting per-basin SOA_HUB_CLASS hub nodes; node_count = members only. - member-of / interfaces now dock to the basin's ANCHOR ENTITY (the real top-degree node that names the basin) via anchor_idx, never a synthetic hub. No self-loop on the anchor. - Drop the basin-hub label loop; remove the now-dead SOA_HUB_CLASS const. Frontend (OsintGraph.tsx): - Render entity<->entity relations only: real relations (2..7,9) PLUS basin tissue (member-of 0 / interfaces 1, now real anchor edges). Schema property-nodes (cls 5/6) and their facet spokes (VALID_FOR 8, facets 10..20) are no longer rendered — a dimension is a prefix carried ON the node (read live by the facet lens), never a node to spoke into. - Give member-of/interfaces visible tissue colours. Retrieval stays explicit-prefix: to gather a basin you filter the family prefix, you don't chase a hub. The facet lens (tenant colouring + raw-edge naming) is unaffected.
…lette Wire the dual-use dimensions ON the node into the OSO1 wire and make them navigable: the explicit prefix you select is the query, and the graph filters to matches immediately. Backend (osint_soa_bytes): - Emit an additive tenant tail: node_count x 11 facet bytes = value[1..=11] (militaryUse, civicUse, airo:type, MLType, purpose, capacity, currentStatus, type, output, impact, stakeholder). No schema node — the dimension is a prefix carried on the node. Frontend (OsintGraph.tsx): - decode picks the tenant stride that fits (11 current / 6 legacy), exposed as soa.tenantStride; all tenant reads are stride-aware. - FACET_AXES_UI widened 6 -> 11 to match the backend axis order. - New lower-right EXPANDABLE PROPERTY PALETTE: per-axis value catalogue (value -> label + count), click a value to add/remove it from the filter. A node survives if, for every axis carrying >=1 selected value, its code is selected (AND across axes, OR within an axis); non-matches dim, edges survive only between two matches. Live '<n> match' count + clear. - This subsumes the single-axis facet legend and expresses the quid-pro-quo query directly: militaryUse=* + stakeholder=* + a purpose value. Note: McClelland motive (GUID2) is not yet populated in the bake, so 'filter by motivation' uses purpose:vair (the AIRO purpose dim) — the real 'why' signal in this data. The '<> dimensions' toggle is now inert (schema nodes no longer rendered); the palette replaces it.
…+ McClelland/Freud power flow Measures the model as distance during reasoning, never materialized: - CAUSALITY axis (intent->impact drift) = divergence, per capability the Jaccard distance of the impact sets between the militaryUse branch and the civicUse branch of the SAME offer (the shared pivot that makes the fork measurable). - DEMAND fork = mil vs civ count for that offer. - POWER level read straight from airo:type bits (the adjacency): P1 oral consume (Subject) / P3 phallic control-others (Deployer) / P4 genital empower-others (Developer/Provider) — McClelland nPow as Freud's gradient. nAch/nAff from intent/use labels (keyword heuristic). '◆ dual-use' lens paints every node by the causality distance of the capability it offers (cool->hot) and streams the ranked capabilities (demand fork · Δimpact · power level · motive) into the readout, with a Macht-adjacency % (how much high-divergence skews to power/nPow). Codebooks stay separate (no merge — distance measured at reason time); nothing materialized as a node or edge. Replaces the now-inert dimensions toggle.
Reframe the divergence readout as the CAUSAL CHAIN AIwar builds to prove the
harm companies deny ('as long as dual-use isn't proven harmful, we continue'):
SITUATION (4 outside factors, the two orthogonal axes):
capability → [mil/civ demand] → explicit purpose ⟹ implicit impact
— the intent→impact drift, chained end to end.
PERSON (personality trait):
POWER_LEVEL from airo:type (Freud gradient P1 consume / P3 control-others /
P4 empower), else the McClelland motive (nPow/nAch/nAff), reasoned AGAINST
the situational chain.
Per capability the readout now shows the chain + the driving trait, and a
'Macht-driven %' = how much of the high-divergence is carried by a power trait
(P3/P4 or nPow) — the Person→Situation attribution that closes the causal chain
the 'can't prove it's harmful' defense relies on staying open.
Model additions: per-capability dominant explicit purpose + implicit impact
(the chain's two ends), collected from the tenant purpose/impact bytes. No
demand-axis distance fabricated (no semantic table in the cockpit — the honest
gap); the causality axis is what's computed. Lewin/Atkinson/Rheinberg grounding.
…LType Verified every tenant position matches between the bake (osint_gotham.rs value[1..=11]) and the frontend reasoning (OsintGraph.tsx AX 0..10), plus the airo:type bitset bit-values (P4=Dev|Prov|Supp, P3=Deployer|Operator, P1=Subject) and stride=11. One naming blemish fixed: byte 4 is fed from MLTask/MLTasks (the NEED per the need/offer typing) but was labeled MLType (the OFFER) on both sides. Renamed FACET_MLTYPE→FACET_MLTASK, REL_FACET_MLTYPE→REL_FACET_MLTASK (bake) and AX.mlType→mlTask, FACET_AXES_UI[3]/REL_NAME[13] 'MLType'→'MLTask' (reasoning). The codebook stays ML_TYPE (the value register) — dimension name ≠ codebook name.
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 31348d26c0
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| // DEMAND fork = mil vs civ count for that offer. | ||
| // POWER level = airo:type bits (Freud gradient), the adjacency to nPow. | ||
| const duModel = useMemo(() => { | ||
| if (!soa || !soa.tenants || !soa.tenantStride || !view) return null; |
There was a problem hiding this comment.
When /osint.soa is a legacy 6-wide tenant asset (which decodeSoa explicitly accepts, and the committed fallback crates/cockpit-server/assets/osint_scene.soa is 6-wide), this guard still enables the new dual-use model. The code below then indexes axes such as AX.impact/AX.stakeholder with i * stride + ax; with stride === 6 those reads come from the next node's tenant bytes rather than missing columns, so the ◆ dual-use heatmap/readout is computed from unrelated data. Return null unless tenantStride >= 11 (or avoid the 11-wide-only axes) before enabling this lens.
Useful? React with 👍 / 👎.
|
Caution Review failedThe pull request is closed. ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (2)
📝 WalkthroughWalkthroughThe PR changes the OSINT SoA wire format and facet encoding on the server (renaming the MLTask facet, emitting members-only nodes with anchor-docked edges, and appending a per-node facet tenant tail), and rewrites the cockpit OsintGraph component to decode that tail, filter/color nodes by tenant facets, and drive a new dual-use divergence heat lens. ChangesServer-side SoA and facet encoding
OsintGraph cockpit rewrite
Estimated code review effort: 4 (Complex) | ~60 minutes Sequence Diagram(s)sequenceDiagram
participant User
participant OsintGraph
participant duModel
participant HeatOverlay
User->>OsintGraph: click dual-use button
OsintGraph->>duModel: compute causality/divergence from tenant signals
duModel-->>OsintGraph: ranked Person x Situation rows and heat map
OsintGraph->>HeatOverlay: heatNodes(heat)
OsintGraph->>OsintGraph: stream ranked chain lines to readout
Possibly related PRs
Poem
✨ Finishing Touches📝 Generate docstrings
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 |
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.
What
Follow-up to the merged #68. Takes the OSINT cockpit from "11 dims stacked as tenant bytes + materialized schema hubs" to the substrate as intended: the dimension is a prefix carried on the node, and the cross-cutting structure is reasoned as logical edges — never a property-as-node hub. Then wires a real analysis on top: dual-use divergence as a Person × Situation causal chain.
Why
The rendered graph was "still islands": each dimension value (
MLTask,SpatialReasoning, a basin) was materialized as a node with spokes into it — a category error (key: valueis a tenant slot, not two nodes + an edge). A materialized property-node has no logic to dock as an edge; it just invites "look up a property as a node." This PR removes that and replaces it with reasoning over the tenant.Changes (5 commits)
1 — dissolve the hubs into anchor-entity tissue (
osint_gotham.rs,OsintGraph.tsx)SOA_HUB_CLASShubs;member-of/interfacesnow dock to the basin's anchor entity (a real node with its own edges), not a synthetic hub.SchemaValue/SchemaAxisproperty-nodes + their facet spokes are no longer rendered — the dimension lives on the node.2 — 11-wide facet tenant tail + expandable property palette (
osint_gotham.rs,OsintGraph.tsx)node_count × 11facet bytes (value[1..=11]); client decodes stride 11 (legacy 6 still decodes).3 — dual-use divergence lens (
OsintGraph.tsx)airo:typebits (P1 consume · P3 control-others · P4 empower).4 — the causal chain: Person × Situation (
OsintGraph.tsx)capability → [mil/civ] → explicit purpose ⟹ implicit impact; Person = the power trait reasoned against it.Macht-driven %= the Person→Situation attribution (Lewin/Atkinson/Rheinberg;Tendency = Motive × Incentive).5 — align bake ⟷ reasoning schema
MLTask(the NEED), was mislabeledMLType(the OFFER) on both sides — renamed consistently. Codebook staysML_TYPE(value register ≠ dimension name).Test on Railway
Deploy this branch → the default OSINT graph:
N match.◆ dual-uselens — nodes painted cool→hot by causality divergence; readout streams the causal chain per capability (capability → [mil/civ] → explicit ⟹ implicit │ trait) withMacht-driven %.Honest gaps
cargo/npm-verify locally (session network-gated); edits are structurally balanced and mirror existing patterns — Railway's build is the verification.🤖 Generated with Claude Code
Generated by Claude Code
Summary by CodeRabbit
New Features
Bug Fixes