Implement Phase 5 tensor graph and ensemble representations

[EmergentMonk]

Summary

Implements Phase 5 as the representation/computational layer above the merged Phase 4 evidence boundary.

Perfect Mathematics Does Not Equal Perfect Biological Reality.

A representation earns scientific interpretation only from explicit evidence and validation.

This PR does not create a V1 biological model, biological validation, molecular-dynamics realism, clinical validity, patient-specific meaning, or a performance claim.

Numerical arrays and true tensors

Adds explicit row-major numerical projections for Cartesian positions and pair-distance-squared values. These remain tensor_claimed=false.

Adds a separately typed centered Cartesian coordinate tensor only because its transformation semantics are declared: component-axis basis permutations, orthogonal Cartesian vector transformations, and translation removal by centroid subtraction.

Typed graphs and hypergraphs

Adds separately named model, model-hypergraph, provenance, tensor-factor, and visualization graph contracts while retaining the existing Phase 3C execution graph as its own namespace.

N-ary model constraints remain hyperedges and are not silently expanded into pairwise biological relationships.

Observables

Adds all 120 unique pair-distance observables for the 16-node V0 fixture, an explicit assumed computational-contact cutoff, and nearest-neighbour geometric accessibility under an explicit assumed clearance threshold.

Every contact/accessibility output remains a computational observable with biological/biochemical claims fixed false.

Ensemble statistics

Uses an explicit deterministic Phase 3B index set [0, 11704, 23408]. Phase 5 first requires the existing fixed Phase 3B residual verification gate, then runs each listed conformation as a one-state slice and computes count/min/max/mean/median/population variance/population standard deviation for pair-distance extrema.

The statistics describe only that explicit synthetic index set. No sampling-population or biological-ensemble inference is made.

Computational uncertainty

Adds explicit Phase 5 uncertainty types:

unknown
interval
distribution
ensemble

Distribution records are metadata-only in this phase and do not silently sample. Computational uncertainty cannot promote evidence strength.

Tensor-network assessment

Adds a fail-closed graph-structured tensor-network assessment. Admission requires both exact reconstruction and material reduction.

The current exact candidate X = X * I3 is deliberately not admitted because it stores more elements than the dense centered tensor. This is a negative result by design, preventing tensor-network sophistication from becoming cargo-cult optimization.

Optional vortex-inspired projection

The reference config keeps vortex-inspired projection disabled. When explicitly enabled it emits only a representational radial-squared / Cartesian phase embedding and fixes biological ontology/scientific interpretation claims false.

Gate and CI

Adds IGM-PHASE5-REPRESENTATION-GATE-V1, deterministic bundle SHA-256 identity, dedicated Phase 5 schemas/validator, igm-represent, and a Phase 5 workflow covering structural validation, Rust tests, deterministic output identity, graph separation, tensor-network rejection, and scientific non-promotion.

Summary by Sourcery

Establish the Phase 5 computational representation layer above the Phase 4 evidence boundary with deterministic, schema-validated, and explicitly non-biological tensor, graph, observable, ensemble, and uncertainty outputs.

New Features:

• Add a deterministic Phase 5 representation bundle covering numerical arrays, a semantically declared centered coordinate tensor, typed graph and hypergraph views, computational observables, ensemble statistics, uncertainty records, and optional vortex-inspired projections.
• Add the igm-represent CLI for schema-admitted representation bundle generation.

Bug Fixes:

• Prevent representations, graph relationships, computational contacts, accessibility results, uncertainty records, and tensor-network candidates from promoting biological, scientific, clinical, or performance claims.

Enhancements:

• Preserve Phase 4 provenance boundaries and Phase 3B execution-profile identity throughout representation generation.
• Add fail-closed tensor-network assessment requiring exact reconstruction and material reduction, rejecting the current non-reducing identity factorization.
• Add deterministic canonical SHA-256 identities and validation for representation bundles, gates, schemas, graph separation, observables, and provenance consistency.

Build:

• Register the Phase 5 representation library and igm-represent binary in Cargo.

CI:

• Add a Phase 5 GitHub Actions workflow covering structural governance, Rust tests, deterministic output, tamper rejection, graph namespace separation, tensor-network rejection, and non-promotion checks.

Documentation:

• Document Phase 5 representation contracts, assumptions, computational boundaries, tensor semantics, ensemble statistics, uncertainty handling, and CLI usage.

Tests:

• Add Phase 5 Rust and validator coverage for tensor semantics, graph separation, observables, ensemble gating, uncertainty kinds, optional projections, deterministic identity, and non-promoting admission.

Chores:

• Add Phase 5 configuration and representation-bundle JSON Schemas plus the reference V0 representation profile.

[sourcery-ai]

Sorry @EmergentMonk, you have reached your weekly rate limit of 500000 diff characters.

Please try again later or upgrade to continue using Sourcery

[sourcery-ai]

Reviewer's Guide

Implements the Phase 5 representation layer on top of the existing Phase 4 boundary, adding Rust runtime support, a new CLI, typed tensor/graph/ensemble structures, deterministic bundle identity, governance gates, and CI validation for the new contracts without promoting biological/clinical claims.

Sequence diagram for deterministic Phase 5 bundle generation

sequenceDiagram
    participant CLI as igm-represent
    participant Engine as Phase5Engine
    participant Model as Phase4 validated profile
    participant P3B as Phase3B engine
    participant Gate as Phase5 representation gate
    participant Bundle as JSON bundle

    CLI->>Engine: load(model_path, config_path, repository_root)
    Engine->>Model: load_profile(model_path)
    Engine->>P3B: PentaCrtEngine::load(model_path, execution_profile_path)
    CLI->>Engine: bundle()
    Engine->>P3B: verify_penta_crt(penta, verification_samples)
    P3B-->>Engine: accepted residual verification
    Engine->>P3B: run_penta_crt(index, one-state slice)
    P3B-->>Engine: pair-distance extrema
    Engine->>Gate: make_gate(config, centered_tensor, assessment, vortex)
    Gate-->>Engine: accepted non-promoting gate receipt
    Engine->>Bundle: compute bundle_sha256
    Bundle-->>CLI: deterministic representation bundle

Loading

File-Level Changes

Change	Details	Files
Add Phase 5 representation engine and data model in Rust, including tensors, graphs, observables, ensemble statistics, uncertainty, tensor-network assessment, and deterministic bundle identity.	Introduce phase5.rs with Phase5Engine that loads model/profile/config, builds geometry, and emits a RepresentationBundle JSON structure Define explicit numerical array projections for positions and pair-distance-squared with tensor_claimed=false Add a declared centered Cartesian tensor with explicit transformation semantics and tensor_claimed=true Add typed model graph, model hypergraph, provenance graph, tensor-factor graph, visualization graph, and graph separation receipt to keep namespaces distinct and non-promoting Compute pair-distance and accessibility observables under explicit assumed thresholds with biological/biochemical claims fixed false Compute ensemble statistics over a deterministic Phase 3B index set after residual-gate verification, using population variance only for that explicit set Define computational uncertainty kinds (unknown, interval, distribution, ensemble) and ensure they cannot promote evidence strength Implement tensor-network assessment that requires exactness and material reduction, deliberately rejecting the identity factorization Make vortex-inspired projection optional, representational-only, and non-ontological Compute fail-closed gate receipt (Phase5GateReceipt) and bundle SHA-256 identity via canonical JSON, enforcing non-promotion of scientific/biological/clinical claims	runtime/rust/src/phase5.rs runtime/rust/src/lib_v6.rs
Add igm-represent CLI to emit Phase 5 bundles from model and config.	Introduce representation_main.rs binary that wires Phase5Engine::load + bundle() to a bundle MODEL.json CONFIG.json command Provide usage/help text that clarifies the bundle is representational/computational evidence only Ensure failures are reported via RuntimeError and non-zero exit codes	runtime/rust/src/representation_main.rs Cargo.toml
Define Phase 5 JSON schemas, reference config/profile, and documentation describing contracts, semantics, and non-promotion rules.	Add PHASE5_REPRESENTATIONS.md documenting all Phase 5 contracts, tensor/graph/ensemble semantics, uncertainty vocabulary, tensor-network admission criteria, and the representation gate Introduce phase5-representation-config.schema.json that pins config identity, V0 validation level, and non-promotion of scientific/biological/clinical/performance claims Introduce phase5-representation-bundle.schema.json that pins the representation contract, gate text, and requires scientific/biological/clinical/performance claims to be false Add runtime/profiles/igm-phase5-v0.json as the reference Phase 5 representation configuration with explicit thresholds, ensemble indices, tensor-network assessment settings, and claim flags fixed false	docs/PHASE5_REPRESENTATIONS.md schemas/phase5-representation-config.schema.json schemas/phase5-representation-bundle.schema.json runtime/profiles/igm-phase5-v0.json
Add fail-closed Python validation tooling and CI workflow for Phase 5 representations.	Introduce tools/validate_phase5.py to validate repository-level Phase 5 contracts and optional bundles, enforcing invariants like non-promotion, graph separation, hypergraph non-expansion, ensemble assumptions, tensor-network rejection, and deterministic SHA-256 identity Wire validate_phase5.py, JSON-schema validation, Phase 4/profile/docs checks, Rust tests, release build, deterministic bundle emission, tensor-network rejection check, and graph-separation/non-claims checks into a new GitHub Actions workflow Ensure the workflow runs on pushes and PRs that touch Phase 5-relevant files and verifies that Phase 5 bundles satisfy the representation gate	tools/validate_phase5.py .github/workflows/phase5-representations.yml

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[EmergentMonk]

Phase 5 implementation is now green on the exact current head 087153a4ad8170b733bd3a6abc81bcb8fc344e4f.

Implemented contracts include explicit non-tensor arrays, a transformation-declared centered coordinate tensor, typed model/hypergraph/provenance/tensor-factor/visualization namespaces, 120 pair-distance/contact observables, geometric accessibility, Phase-3B-gated explicit-index ensemble statistics, computational uncertainty (unknown|interval|distribution|ensemble), optional non-ontological vortex-inspired projection, fail-closed tensor-network assessment, deterministic bundle identity, and IGM-PHASE5-REPRESENTATION-GATE-V1.

The governance policy also now carries normative INV-REP-001: A representation earns scientific interpretation only from explicit evidence and validation.

Current exact-head workflows are all green: documentation foundation, visual laboratory, Rust runtime, Phase 4 evidence adapters, and the dedicated Phase 5 representation suite. No unresolved review threads are present at this point.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 171195d1bb

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".

[EmergentMonk]

Addressed all 7 Codex review findings on the current Phase 5 head 7d63b073586b8ce4b1406bd20f60b6c4e8b764bb.

Fixes include:

• recompute and verify the domain-separated canonical bundle_sha256 and gate identity rather than only checking digest shape;
• retain the exact Phase 3B execution-profile ID and canonical SHA-256 in ensemble output and bind them into the final bundle identity;
• run the mandatory Phase 4 source validators and executable evidence CLI checks inside the Phase 5 workflow;
• route the public Phase 5 API through phase5_v2.rs, which applies the declared model/config JSON-Schema gates before projection;
• replace permissive nested bundle objects with fail-closed JSON-Schema definitions, required fields/constants/shapes, and additionalProperties=false;
• independently recompute pair coverage, distances, contact predicates, nearest-neighbour accessibility, ensemble/statistical consistency, uncertainty constraints, and tensor-network accounting;
• validate actual model/provenance/tensor-factor/visualization graph contracts, namespaces, endpoints, hyperedge participants, and non-promotion semantics rather than trusting only the separation receipt.

The tamper suite now also checks stale bundle identities, a recomputed-digest graph semantic promotion, and a recomputed-digest inconsistent observable. The first version of that regression targeted the visualization graph, which legitimately has zero edges under the current V0 cutoff, so it was corrected to tamper a populated model-graph edge instead.

All five workflows are green on this exact head: Phase 5 representation checks, Rust runtime checks, Phase 4 evidence adapter checks, documentation foundation, and visual laboratory checks. All 7 Codex review threads are resolved.