GNN v3.0.0 — Long-Running Orchestration

GNN v3.0.0 — Long-Running Orchestration

GeneralizedNotationNotation v3.0.0 adds safe-by-design long-running orchestration on top of the GNN text language and 25-step processing pipeline. Every new module generates, validates, replays, or plans data only — no live infrastructure is ever mutated (no container is executed, no cluster or device is touched).

Highlights — three orchestration contracts (src/pipeline/)

• Durable observation streams (durable_streams.py) — content-checksummed file/array StreamManifests and replayable ExecutionTraces, so an extended run can be observed, paused, and re-derived deterministically before any live sensor or device-backed stream is introduced. Cross-architecture-deterministic checksums, injective stream ids, and trace meaning re-bound to ground truth on verify.
• Resumable run sessions (run_session.py) — RunSession manifests with atomic checkpoint/resume, status inspection, and path-escape-safe, idempotent cancellation cleanup, so an extended model-family acceptance run can be interrupted and resumed without corrupting partial state.
• Auditable container plans (container_plan.py) — hardened plan generation (non-root, read-only rootfs, cap-drop ALL, digest-pinned images, resource limits), a static security_review (privileged / root / unpinned-image / plaintext-secret / sensitive-host-mount / host-namespace / dangerous-capability findings), rollback descriptors, and deterministic plan hashes — without ever touching a real cluster.

Additive live wiring (no change to the 25-step critical path)

• session_acceptance.py — resumable, checkpoint-after-each model-family acceptance.
• run_manifest.py — emit durable StreamManifests + a replayable ExecutionTrace from a completed run's output/ (verified on a real run: 105 manifests + a 25-event trace).
• pipeline_container_plan.py — a security_review-clean container plan generated from the real input/config.yaml.
• A strict, fail-closed acceptance gate: scripts/run_v3_orchestration_acceptance.py (--inject-defect exits non-zero).
• Three new Model Context Protocol tools (tool total 137 to 140).

Release evidence

The standing release gates were re-run green for all 9 model families: semantic fidelity (run_semantic_fidelity_gate.py), cross-framework reliability (run_cross_framework_reliability.py, GridWorld compared across PyMDP, RxInfer, and ActiveInference.jl), and model-family acceptance (run_model_family_acceptance.py). The orchestration foundation ships without regressing the v2.0.0 reliability surface. The CI matrix is green (ruff format/check, four documentation audits, and mypy src), and the long-standing tool-version drift in the matrix was cleaned up as part of this release.

Documentation, manuscript & visualizations

• README, AGENTS, DOCS, ARCHITECTURE, and the doc/ entry points brought current to v3.0.0; new reference page doc/pipeline/v3_orchestration.md.
• The auto-injected manuscript gained a "Long-Running Orchestration Contracts" methods section and a new architecture figure; every quantitative value is a token resolved from the live repository.
• Visualizations improved: a clean layered pipeline DAG and a new orchestration-architecture diagram.

Next

v4.0.0 — bounded autonomy and reviewed self-editing workflows, gated behind this orchestration work.

Full details: see CHANGELOG.md and TO-DO.md.

v2.0.0

Added

• Semantic fidelity release gate: scripts/run_semantic_fidelity_gate.py writes gnn_semantic_fidelity_ledger_v1 artifacts for maintained model families.
• Strict semantic contracts: representative fixtures now preserve model identity, variables, edges, dimensions, parameter shapes, equations, time, and ontology mappings across JSON parse/serialize/parse checks.
• Cross-framework reliability release gate: scripts/run_cross_framework_reliability.py writes gnn_cross_framework_reliability_ledger_v1 artifacts with compatible, required, and unsupported backend statuses.
• GridWorld three-backend comparison: GridWorld is profiled for PyMDP, RxInfer, and ActiveInference.jl, including seed, trace length, matrix-shape, and matrix-provenance parity.

Changed

• GridWorld model-family acceptance now requests PyMDP, RxInfer, and ActiveInference.jl for the v2 comparison fixture instead of a PyMDP-only profile.
• Roadmap next target moves to v3.0.0 for durable streams, long-running sessions, and auditable container plans.

Fixed

• JSON serialization now emits equation objects instead of lossy stringified dataclasses, preventing silent semantic round-trip drift.
• Cross-framework reliability no longer certifies aggregate Step 12 success without successful non-skipped execution-detail rows and current simulation payloads for required backends.

---

v1.9.0

[1.9.0] — 2026-06-12

Added

• Model-family acceptance release gate: manifest-driven all-family strict acceptance for basics, discrete, continuous, hierarchical, multi-agent, precision, structured, gridworld, and scaling-study fixtures.
• Cross-step evidence ledger: release ledger now links Step 3/5/6/11/12/15/16/23 statuses, artifact links, telemetry presence, renderer/execution status, and concrete skip reasons per family.
• Interpretability summaries: per-family summaries now include variable/edge inventories, matrix-shape tables, telemetry presence, optional trace previews, renderer/execution status, and artifact links.

Changed

• Continuous and hierarchical Step 11/12 outcomes are explicit profiled unsupported skips with concrete reasons, not raw render/execute failures accepted by profile math.
• v1.7.0 is retired as a foundation-only track; unfinished runtime-depth ambitions move forward into v2+ reliability and orchestration milestones.
• Current test evidence updated to 2,399 collected tests; final full-suite release evidence is recorded in TO-DO.md, README.md, and test documentation after the v1.9 release gate rerun.

Fixed

• Removed the model-family acceptance reason-pattern fallback that could reclassify failed renderer/executor steps as unsupported success.
• Hardened strict acceptance so profiled unsupported steps must be skipped before execution and failed Step 11/12 summaries fail closed.
• Prevented cross-framework analysis from reading stale repo-tracked output/ artifacts during isolated /tmp acceptance runs.
• Relaxed an environment performance smoke threshold to match other slow module smoke tests and avoid full-suite load false negatives.

Validation

• PR #12 checks passed: Bandit, CodeQL, analyze (python), dependency-review, markdown-audit, security, test (3.11), test (3.12), test (3.13).
• Local full suite: 2381 passed, 17 skipped, 1 xfailed with Ollama integration tests ignored.
• Local collect-only: 2,399 collected tests with Ollama integration tests ignored.
• All-family strict acceptance: passed for 9 manifest families; continuous and hierarchical Step 11/12 are explicit profiled unsupported skips with concrete reasons and no raw failed Step 11/12 counts.

v1.8.0: Developer Kit & Safe Local MCP Orchestration

v1.8.0 — Developer Kit & Safe Local MCP Orchestration

Added

• Template developer kit: packaged template index, package-data template assets, gnn templates list, gnn templates show NAME, and gnn pull NAME --output-dir ... --dry-run --overwrite.
• Template safety contracts: checksum verification, collision handling, symlink/path traversal rejection, unknown-template failures, and installed-wheel smoke coverage outside the repo checkout.
• MCP local HTTP orchestration: bearer-token auth through GNN_MCP_TOKEN, localhost default binding, explicit insecure local opt-in with GNN_MCP_ALLOW_INSECURE_LOCAL=1, safe-tool filtering, optional rate limiting, and default-denied resource reads unless explicitly allowlisted.
• Capability-contract verifier for template package data, MCP auth/resource safety, acceptance-command isolation, roadmap ordering, count drift, and autonomy non-mutation claims.

Changed

• TO-DO.md now treats v1.8.0 as the developer-kit release and v1.9.0 as the next model-family reliability target.
• Developer documentation advertises verified template and MCP commands only, with /tmp output directories in acceptance smokes to avoid tracked output/ churn.
• Current evidence: 2,397 collected tests; latest full local suite evidence with Ollama integration excludes is 2,379 passed, 17 skipped, 1 xfailed.

Fixed

• Removed release-facing false-certification paths around optional framework fallback, stale counts, MCP unauthenticated HTTP, unsafe MCP resource exposure, and template assets that only work from a source checkout.
• Fixed CI-sensitive optional Julia package gating so missing optional backend packages skip instead of failing while preserving failures for real Julia errors.

Validation

• PR #11 checks: 9 passed, 0 failed.
• Local gates: git diff --check, Ruff check/format, mypy, docs audit, GNN doc patterns, maintained-doc terms, repo terminology, capability contracts, and just lint.
• Acceptance smokes: gnn templates list, gnn templates show pomdp-gridworld-3x3, gnn pull pomdp-gridworld-3x3 --output-dir /tmp/gnn-pull --dry-run, and token-enabled MCP HTTP tests.
• Full local suite: uv run --extra dev python -m pytest src/tests/ -q --tb=no --ignore=src/tests/llm/test_llm_ollama.py --ignore=src/tests/llm/test_llm_ollama_integration.py -> 2,379 passed, 17 skipped, 1 xfailed.

v1.6.0: Zero-Mock Testing Stabilization & Documentation Hardening

Fixed

• Testing Constraints: Entirely removed dependency on internal hypothesis.internal.conjecture.optimiser arrays mitigating python 3.13 crashes, transitioning to native parameterized fuzzers.
• HTML Assertion Accuracy: Relaxed legacy test_report_functional strings to mathematically accept dynamic semantic UI tags (<html lang="en"> instead of <html>).

Added

• Global Documentation Guardrails: Enforced 100% adherence to zero-mock documentation mandates. Provisioned AGENTS.md, README.md, and SPEC.md components natively within transient boundaries (output/, .benchmarks/, scripts/) and defined the master environment SPEC.md at root.
• Type checker visual dashboards: Baseball-card style model summary PNGs (dark neon theme), validity mosaics, issue distribution charts, and type category pie charts generated in output/5_type_checker_output/visualizations/
• Resource estimation integration: estimate_file_resources() now delegates to estimation_strategies.py for real FLOPS, memory, and complexity scoring instead of naive heuristics
• Documentation hub sync: Automated doc/gnn/modules/[00-24].md reconciliation from src/*/AGENTS.md source-of-truth

Changed

• Zero-Mock Policy Enforcement: Systematically excised all mentions and framework hooks relating to pytest-mock allowing completely isolated functional testing constraints natively.
• Type checker consolidation: Deleted redundant src/type_checker/checker.py; all logic unified in processor.py (GNNTypeChecker)
• Test suite alignment: test_type_checker_overall.py rewired to target production processor.py orchestrator (zero-mock)
• Deprecated marker removed: safe_to_fail marker replaced with standard xfail in pyproject.toml and pytest.ini
• Default local LLM: Ollama default tag is smollm2:135m-instruct-q4_K_S (llm.defaults.DEFAULT_OLLAMA_MODEL); override with OLLAMA_MODEL or input/config.yaml llm.model.
• Core dependencies: openai, ollama, python-dotenv, and aiohttp are installed with the base package (LLM step and OpenRouter/Perplexity providers work without uv sync --extra llm).

v1.3.0 — The Pipeline Coherence & Agentic Integration Release

v1.3.0 — The Pipeline Coherence & Agentic Integration Release

This release establishes the Generalized Notation Notation (GNN) repository as a pristine, fully integrated, agent-ready ecosystem. With over 756 documentation files audited, deep MCP integration stabilized, and 100% test passing ratios across 1,522 unit tests and 25 pipeline steps, this update sets a new gold standard for repository coherence.

🌟 Major Highlights

Agent-Ready SKILL.md Capability Framework

• Deployed 28 Anthropic skills standard SKILL.md files covering every src/ module.
• Mapped precise AI capability boundaries, workflow patterns, and key API commands.
• Validated 100% of API references natively against runtime __all__ list exports, securing accuracy for AI-driven multi-agent orchestration.

Pipeline Coherence & Zero Dependencies

• The 25-step GNN Processing Pipeline achieves perfect coherence, producing a 100.0 Health Score via Intelligence Analysis.
• Resolved all structural and integration warnings, specifically remediating asynchronous Model Context Protocol (MCP) registration bugs inside the llm and api modules (initialize_llm_module and register_tools properly initialized).
• The pipeline now runs with 0 errors, 0 failures, and 0 warnings under automated orchestration.

Massive Repo-Wide Documentation Audit

• Audited 756 total .md documentation files.
• Validated 5,065 internal reference links, cleanly resolving 124 dead links/legacy artifacts.
• Normalized stale test count claims across 44 files, explicitly confirming 1,522 passing tests.
• Enforced 100% Triad structural completeness (README, AGENTS, SPEC, SKILL) across all 28 project modules.
• Achieved Zero YAML parsing errors, Zero missing required sections, and Zero TODO/FIXME gaps in the core documentation hub.

🛠️ Detailed Engineering Fixes

• src.api.mcp: Bridged FastAPI backend jobs with MCP wrapper functions, surfacing 5 fully functional API coordination tools over the protocol.
• src.llm.mcp: Handled coroutine lifecycle within the MCP synchronous initialization space, preventing an unawaited processor loop and achieving robust default context window loading.
• render/ framework mappings: Successfully verified the continuous parsing engine across targeted dynamic modules (rxinfer, activeinference_jl, jax, pymdp, discopy) matching precisely to their theoretical documentation profiles.

Summary Profile

• Tests Passed: 1,522 / 1,522
• Script Pipeline State: 25 Steps Clean (0 Warnings)
• MCP Tools Registered: 100+ over 29 module servers
• Documentation Health: Perfect (0 broken links, completely synced)

This represents the final polished milestone of Phase 7 normalization.

v1.2.0 — Data Accuracy & Documentation Sync

🎯 Release Highlights

v1.2.0 is a data-integrity release that fixes 5 pipeline bugs and unifies all documentation metadata across 635 files. No new features — only correctness and consistency.

---

🐛 Pipeline Bug Fixes

Critical

• ActiveInference.jl renderer typo — POLICY_LENGTH → POLICY_LEN caused UndefVarError in generated Julia scripts (src/render/activeinference_jl/activeinference_renderer.py)
• ActiveInference.jl validation ordering — validation_status referenced before definition, causing runtime errors during code generation

Medium

• LLM connection extraction — extract_connections() in src/llm/analyzer.py now correctly parses GNN-specific connection operators (>, -, <) from ## Connections sections, increasing detected connections from 0 to 11

Low

• Website analysis scan — src/website/generator.py now scans root-level JSON files in 16_analysis_output/, not just the analysis_results/ subdirectory (was reporting 0 analysis results)
• Pipeline summary status — Preliminary summary in src/main.py changed from IN_PROGRESS to SUCCESS to accurately reflect completed steps

---

📖 Documentation Sync

This release eliminates all metadata inconsistencies across the entire repository:

Metric	Before (v1.1.4)	After (v1.2.0)
Test count	Mixed: 1,083 / 1,319	1,319 everywhere (35+ files)
Version string	Mixed: v1.1.0 / v1.1.4	v1.2.0 everywhere (40+ files)
Dates	Mixed: Feb 9 / Feb 20 / Feb 21	2026-02-23 everywhere (50+ files)
Execution time	Mixed: 40s / 90s / 120s / 3min	~5 min (with LLM step) (6 files)
Module coverage %	Divergent tables in root vs src	Unified (2 files)
Architecture version	2.2.1	2.3.0
Total files changed	—	635

---

📋 New: Project Roadmap

Added TO-DO.md with planned milestones:

Release	Theme
v1.2.1	LLM step performance (profile, batch, cache)
v1.3.0	Testing & CI improvements
v1.4.0	GNN language & parser enhancements
v1.5.0	Website & reporting upgrades
v2.0.0	Multi-model composition, LSP, cloud execution

---

✅ Verification

• Full pipeline: 25/25 steps SUCCESS (299s)
• Test suite: 1,303 passed, 16 skipped, 0 failed (621s)
• Stale references: 0 remaining
• Python version: 3.11+
• Package manager: UV recommended

📦 Installation

git clone https://github.com/ActiveInferenceInstitute/GeneralizedNotationNotation.git
cd GeneralizedNotationNotation
python3 src/1_setup.py --verbose
python3 src/main.py --target-dir input/gnn_files --verbose

🔗 Key Documentation

• README.md — Project overview
• ARCHITECTURE.md — System architecture (v2.3.0)
• SETUP_GUIDE.md — Installation guide
• TO-DO.md — Release roadmap
• doc/gnn/README.md — GNN documentation index
• CONTRIBUTING.md — Contribution guidelines

---

Full Changelog: v1.1.4...v1.2.0

GNN v1

Version 1 of GNN.