Add evidence-aware Sampled Values profile infrastructure

.github/workflows/ci.yml

@@ -11,8 +11,34 @@ permissions:
   contents: read
 
 jobs:
+  public-content:
+    name: Validate public terminology
+    runs-on: ubuntu-24.04
+    steps:
+      - name: Checkout
+        uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
+
+      - name: Validate public terminology neutrality
+        id: neutrality
+        continue-on-error: true
+        run: python3 scripts/validate-public-neutrality.py
+
+      - name: Upload terminology validation evidence
+        if: always()
+        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
+        with:
+          name: arsvin-public-neutrality-report
+          path: artifacts/public-neutrality-report.txt
+          if-no-files-found: error
+          retention-days: 14
+
+      - name: Enforce terminology validation result
+        if: steps.neutrality.outcome == 'failure'
+        run: exit 1
+
   build-test:
     name: Build, test, and validate site
+    needs: public-content
     runs-on: windows-2025
     steps:
       - name: Checkout
@@ -56,7 +82,7 @@ jobs:
 
       - name: Test with whole-engine and protocol-core coverage gates
         shell: pwsh
-        run: .\scripts\test-with-coverage.ps1 -MinimumWholeEngineLineCoverage 10.5 -MinimumLineCoverage 60 -NoRestore
+        run: .\scripts\test-with-coverage.ps1 -MinimumWholeEngineLineCoverage 13 -MinimumLineCoverage 70 -NoRestore
 
       - name: Upload test and coverage evidence
         if: always()

.github/workflows/pages.yml

@@ -9,6 +9,7 @@ on:
       - 'scripts/build-public-site.py'
       - 'scripts/validate-public-site.ps1'
       - 'scripts/validate-public-site.py'
+      - 'scripts/validate-public-neutrality.py'
       - '.github/workflows/pages.yml'
   workflow_dispatch:
 
@@ -31,6 +32,9 @@ jobs:
       - name: Checkout
         uses: actions/checkout@9c091bb21b7c1c1d1991bb908d89e4e9dddfe3e0 # v7
 
+      - name: Validate public terminology neutrality
+        run: python3 scripts/validate-public-neutrality.py
+
       - name: Build public site and HTML documentation
         run: python3 scripts/build-public-site.py --output artifacts/public-site
 

CHANGELOG.md

@@ -6,14 +6,25 @@ All notable ARSVIN changes are documented here using a lightweight Keep a Change
 
 ### Added
 
-- Added a formal Sampled Values standards and evidence research gate covering authoritative source hierarchy, claim levels, OMICRON benchmark findings, licensed-standard gaps, and implementation acceptance rules.
+- Added a formal Sampled Values standards and evidence research gate covering authoritative source hierarchy, claim levels, licensed-standard gaps, and implementation acceptance rules.
 - Added a conformance and interoperability matrix separating verified, implemented-generic, provisional, unknown, and out-of-scope behavior for Layer-2 SV, 9-2LE-style workflows, IEC 61869-9, timing, scaling, redundancy, and product claims.
 - Added a safe Sampled Values evidence-intake guide for anonymized SCL/PCAP cases, real-device metadata, independent verification, minimum negative tests, and evidence quality levels.
+- Added transport-independent SV frame observations and an accumulator for stable-field checks, rate estimation, and sample-counter wrap detection.
+- Added evidence-aware profile definitions with standards-neutral sampling, dataset, packing, payload, and source metadata.
+- Added an explainable weighted profile detector that reports matches, conflicts, missing evidence, confidence, and score.
+- Added strict and compatible configuration-versus-wire comparison findings without blocking capture or decoding.
+- Added deterministic tests for observation windows, profile detection, evidence explanations, configuration mismatch handling, sparse-evidence confidence, and profile-definition provenance.
+- Added public terminology-neutrality validation to CI and Pages deployment with retained validation evidence.
 
 ### Changed
 
 - Documentation now prevents unverified profile constants from being promoted directly into production support claims.
-- The SV roadmap now treats StationScout as a system-engineering workflow benchmark and DANEO 400 as the stronger benchmark for SCL-versus-wire verification, merging-unit analysis, timing statistics, orphan detection, triggered recording, and evidence reporting.
+- Public engineering workflow targets are described generically without proprietary product comparisons or branding.
+- The built-in profile catalog contains only the generic SCL-driven Layer-2 SV fallback until profile-specific requirements are verified.
+- Sparse evidence can no longer produce a `Confirmed` profile result; confirmation requires sufficient evaluated evidence plus matching dataset and sampling behavior.
+- Expanded the deterministic suite from 54 to 74 tests.
+- Raised whole-engine coverage from 10.74% to 13.35% and the enforced floor from 10.5% to 13%.
+- Raised protocol-core coverage from 64.97% to 70.47% and the enforced floor from 60% to 70%.
 
 ## 0.4.0 — 2026-07-12
 ### Added

README.md

@@ -62,6 +62,23 @@ The tools are intentionally separated. Publisher evidence proves what the PC gen
 - `smpCnt` continuity and stream-health diagnostics.
 - Decoded instantaneous values, oscilloscope waveform, phasor, and RMS views.
 - Receiver-side Markdown evidence reports.
+- Shared-engine foundation for evidence-aware profile classification and configuration-versus-wire comparison; live UI integration remains a following milestone.
+
+## Profile intelligence foundation
+
+The shared engine now separates:
+
+```text
+Observed wire facts
+Configured SCL expectations
+Evidence-backed profile definitions
+Explainable profile confidence
+Strict or compatible mismatch findings
+```
+
+Sparse evidence cannot produce a false `Confirmed` result. Unknown or conflicting receive traffic remains observable, and the built-in catalog contains only a generic SCL-driven Layer-2 fallback until named-profile requirements are verified.
+
+See [SV Profile Infrastructure](docs/sv-profile-infrastructure.md) and [SV Standards and Evidence Research Gate](docs/sv-research-gate.md).
 
 ## Release downloads
 
@@ -108,9 +125,11 @@ See [Quick Start](docs/quick-start.md) and [Build and Release](docs/build-and-re
 
 | Area | Current status | Boundary |
 |---|---|---|
-| IEC 61850 Sampled Values APDU | Publisher and subscriber implementation for engineering/lab use | Not certified conformance testing. |
-| IEC 61850-9-2LE-style 4I+4V | Supported lab profile | Validate formal requirements independently. |
-| `nofASDU` | `1`, `2`, `4`, `8` | Publisher and receiver behavior is software/PC timing dependent. |
+| IEC 61850 Sampled Values APDU | Publisher and Subscriber implementation for engineering/lab use | Not certified conformance testing. |
+| IEC 61850-9-2LE-style 4I+4V | Implemented laboratory workflow | Formal profile verification and broader device evidence remain pending. |
+| Generic SCL-driven Layer-2 SV | Dataset-aware engine foundation | Unknown layouts remain visible; unsupported payload elements are diagnosed. |
+| Evidence-aware profile detection | Engine infrastructure implemented | Named-profile definitions await verified source and device evidence. |
+| `nofASDU` | UI workflows emphasize `1`, `2`, `4`, `8` | Publisher and receiver behavior is software/PC timing dependent. |
 | COMTRADE | ASCII, BINARY, BINARY32, FLOAT32 analog replay | Verify scaling and channel mapping before live TX. |
 | PTP / `smpSynch` | Compatibility and lab behavior | Not an IEC 61850-9-3 certified clock. |
 | Windows timing | Best-effort scheduling with visible health metrics | Not deterministic real-time execution. |
@@ -123,7 +142,7 @@ For users:
 - Windows 10 or Windows 11, x64.
 - Npcap for live capture or transmission.
 - Administrator permission when required by the local Npcap/network configuration.
-- Wireshark or another independent packet tool is recommended for verification.
+- An independent packet dissector or process-bus analyzer is recommended for verification.
 
 For developers:
 
@@ -145,34 +164,35 @@ NuGet versions are centrally managed and committed `packages.lock.json` files lo
 Build all release artifacts except the installer:
 
 ```powershell
-.\scripts\publish-release.ps1 -Version 0.3.1
+.\scripts\publish-release.ps1 -Version 0.4.0
 ```
 
 Compatibility wrapper:
 
 ```powershell
-.\publish-win-x64.ps1 -Version 0.3.1
+.\publish-win-x64.ps1 -Version 0.4.0
 ```
 
 Run tests with both repository coverage gates and retain TRX/Cobertura evidence:
 
 ```powershell
-.\scripts\test-with-coverage.ps1 -MinimumWholeEngineLineCoverage 10.5 -MinimumLineCoverage 60
+.\scripts\test-with-coverage.ps1 -MinimumWholeEngineLineCoverage 13 -MinimumLineCoverage 70
 ```
 
-The current suite contains 54 deterministic tests. The complete shared `ARSVIN.Engine` baseline measures 10.74% line coverage across 15,742 instrumented production lines, with 1,691 covered lines. The protocol-core regression surface measures 64.97% across 1,550 lines, with 1,007 covered lines. CI enforces floors of 10.5% for the whole engine and 60% for protocol core.
+The current suite contains 74 deterministic tests. The complete shared `ARSVIN.Engine` baseline measures 13.35% line coverage across 16,312 instrumented production lines, with 2,178 covered lines. The protocol-core regression surface measures 70.47% across 2,120 lines, with 1,494 covered lines. CI enforces floors of 13% for the whole engine and 70% for protocol core.
 
 Generate a CycloneDX SBOM after restoring the solution:
 
 ```powershell
-.\scripts\generate-sbom.ps1 -Version 0.3.1
+.\scripts\generate-sbom.ps1 -Version 0.4.0
 ```
 
 ## Repository structure
 
 ```text
 src/ARSVIN.Engine/                     Shared production engine project and source ownership
 src/ARSVIN.Engine/AR.Iec61850/         IEC 61850, SCL, SV, MMS, capture, diagnostics, and protocol code
+src/ARSVIN.Engine/AR.Iec61850/SampledValues/Profiles/  Profile observation, detection, and comparison engine
 src/ARSVIN.Engine/AR.Iec61850.Transports.Npcap/  Npcap transport implementation
 src/ARSVIN/                            Publisher application
 src/ARSVIN.Subscriber/                 ArSubsv subscriber and visualization companion
@@ -191,6 +211,10 @@ The shared engine is compiled once as `ARSVIN.Engine`, and its source is physica
 
 - [Documentation index](docs/index.md)
 - [Quick start](docs/quick-start.md)
+- [SV standards and evidence research gate](docs/sv-research-gate.md)
+- [SV conformance and interoperability matrix](docs/sv-evidence-matrix.md)
+- [SV profile infrastructure](docs/sv-profile-infrastructure.md)
+- [Profile detection output contract](docs/profile-detection-output.md)
 - [Subscriber verification app](docs/subscriber-verification-app.md)
 - [ArSubsv SV scout companion](docs/arsubsv-sv-scout-companion.md)
 - [SV profile support](docs/sv-profile-support.md)

docs/arsubsv-sv-scout-companion.md

@@ -2,7 +2,7 @@
 
 ArSubsv is the receiver-side companion application for ARSVIN. It is designed to make IEC 61850 Sampled Values visible during lab publishing, commissioning preparation, and publisher verification.
 
-ArSubsv is not an OMICRON product and does not copy OMICRON branding or UI. Its feature target is the same engineering class: discover SV streams, subscribe to one or more streams, visualize current/voltage waveforms, calculate RMS/phase indicators, compare stream timing and counters, and export receiver-side evidence.
+ArSubsv uses its own branding and interface. Its engineering purpose is to discover SV streams, subscribe to one or more streams, visualize current and voltage waveforms, calculate RMS and phase indicators, compare stream timing and counters, and export receiver-side evidence.
 
 ## Feature set
 
@@ -21,7 +21,7 @@ ArSubsv is not an OMICRON product and does not copy OMICRON branding or UI. Its
 
 ArSubsv verifies that this PC and selected NIC receive and decode Sampled Values. It cannot prove that another relay, protection IED, BCU, or fault recorder has consumed the same multicast stream.
 
-SCL binding improves decoding and validation quality. Without SCL, ArSubsv can still discover and count streams, but waveform/phasor/value visualization is limited because the payload layout is not known.
+SCL binding improves decoding and validation quality. Without SCL, ArSubsv can still discover and count streams, but waveform, phasor, and value visualization is limited because the payload layout is not known.
 
 ## Capture file support