Research: five tested structural-frontier extensions after PR #17

.github/workflows/jev-multicall.yml

@@ -66,7 +66,8 @@ jobs:
           python -B -m graph_synthesis.source_structural.report --update-paper
           python -B -m graph_synthesis.assumption_aware.report --update-paper
           python -B -m graph_synthesis.uncertainty.report --update-paper
-          git diff --exit-code -- experiments/jev-multicall-20260918/REPORT.md manuscript/paper-current.md CURRENT_RESULTS.md graph_synthesis/risk_control graph_synthesis/reliability graph_synthesis/structural graph_synthesis/source_structural graph_synthesis/assumption_aware graph_synthesis/uncertainty
+          python -B -m graph_synthesis.frontier.report --update-paper
+          git diff --exit-code -- experiments/jev-multicall-20260918/REPORT.md manuscript/paper-current.md CURRENT_RESULTS.md graph_synthesis/risk_control graph_synthesis/reliability graph_synthesis/structural graph_synthesis/source_structural graph_synthesis/assumption_aware graph_synthesis/uncertainty graph_synthesis/frontier
           python -B -m graph_synthesis.render_current_paper --output "$RUNNER_TEMP/multicall-paper.pdf"
           git diff --exit-code -- manuscript/paper-current.html
       - uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02

.github/workflows/structural-frontier.yml

@@ -0,0 +1,124 @@
+name: Structural frontier research
+on:
+  push:
+    branches: [main, research/structure-aware-next-five]
+  pull_request:
+permissions:
+  contents: read
+concurrency:
+  group: frontier-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+jobs:
+  execute-and-publish:
+    if: github.event_name == 'push' && github.ref == 'refs/heads/research/structure-aware-next-five'
+    runs-on: ubuntu-latest
+    timeout-minutes: 20
+    permissions:
+      contents: write
+    env:
+      PYTHONDONTWRITEBYTECODE: '1'
+    steps:
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262
+        with:
+          ref: ${{ github.sha }}
+          fetch-depth: 0
+      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065
+        with:
+          python-version: '3.12'
+      - run: python -m pip install -r graph_synthesis/assumption_aware/requirements.txt
+      - name: Execute frozen hypotheses, adversarial tests and full integrated regressions
+        shell: bash
+        run: |
+          set -euo pipefail
+          python -B -m graph_synthesis.frontier.integrate_main
+          python -B -m graph_synthesis.frontier.audit_baseline
+          python -B -m graph_synthesis.frontier.run | tee graph_synthesis/frontier/execution.log
+          python -B -m unittest discover -s graph_synthesis/frontier/tests -v 2>&1 | tee graph_synthesis/frontier/tests.log
+          python -B -m graph_synthesis.frontier.run --check | tee graph_synthesis/frontier/replay.log
+          python -B -m graph_synthesis.verify --output "$RUNNER_TEMP/frontier-original-inventory.json"
+          cp "$RUNNER_TEMP/frontier-original-inventory.json" graph_synthesis/frontier/original-inventory.json
+          python -B -m graph_synthesis.frontier.report --figures --update-paper
+          python -B -m graph_synthesis.render_current_paper --output manuscript/paper-current.pdf
+          python -B -m graph_synthesis.frontier.regression 2>&1 | tee graph_synthesis/frontier/all-regressions.log
+          python -B -m graph_synthesis.frontier.report
+          python -B -m graph_synthesis.frontier.verify
+          git diff --exit-code 6530321dd0060c9c7f13bb69d6f1c88346a8e1e2 -- graph_synthesis/structural graph_synthesis/source_structural graph_synthesis/assumption_aware
+      - name: Publish generated research artifacts on this branch only
+        shell: bash
+        run: |
+          set -euo pipefail
+          git config user.name 'github-actions[bot]'
+          git config user.email '41898282+github-actions[bot]@users.noreply.github.com'
+          git add graph_synthesis/frontier CURRENT_RESULTS.md manuscript/paper-current.md manuscript/paper-current.pdf manuscript/paper-current.html manuscript/paper-current.build.json
+          if ! git diff --cached --quiet; then
+            git commit -m 'research: record frontier replay, all integrated tests and complete additive paper'
+            git push origin HEAD:research/structure-aware-next-five
+          fi
+      - uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02
+        with:
+          name: structural-frontier-review-package
+          path: |
+            graph_synthesis/frontier/
+            manuscript/paper-current.md
+            manuscript/paper-current.pdf
+            manuscript/paper-current.html
+            manuscript/paper-current.build.json
+            CURRENT_RESULTS.md
+          retention-days: 14
+      - name: Package complete integrated source
+        run: git archive --format=zip --output="$RUNNER_TEMP/frontier-integrated-source.zip" HEAD
+      - uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02
+        with:
+          name: structural-frontier-integrated-source
+          path: ${{ runner.temp }}/frontier-integrated-source.zip
+          retention-days: 14
+  replay:
+    if: github.event_name == 'pull_request' || github.ref == 'refs/heads/main'
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [ubuntu-latest, windows-latest]
+    runs-on: ${{ matrix.os }}
+    timeout-minutes: 15
+    defaults:
+      run:
+        shell: bash
+    env:
+      PYTHONDONTWRITEBYTECODE: '1'
+    steps:
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262
+        with:
+          ref: ${{ github.event.pull_request.head.sha || github.sha }}
+          persist-credentials: false
+      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065
+        with:
+          python-version: '3.12'
+      - run: python -m pip install -r reproduction/requirements-jev.txt
+      - name: Regression, numerical replay and source/manuscript integrity
+        run: |
+          python -B -m unittest discover -s graph_synthesis/frontier/tests -v
+          python -B -m graph_synthesis.frontier.run --check
+          python -B -m graph_synthesis.frontier.verify
+          python -B -m graph_synthesis.verify --output "$RUNNER_TEMP/frontier-original.json"
+          git diff --exit-code
+  figures-and-paper:
+    if: github.event_name == 'pull_request' || github.ref == 'refs/heads/main'
+    runs-on: ubuntu-latest
+    timeout-minutes: 15
+    env:
+      PYTHONDONTWRITEBYTECODE: '1'
+    steps:
+      - uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262
+        with:
+          ref: ${{ github.event.pull_request.head.sha || github.sha }}
+          persist-credentials: false
+      - uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065
+        with:
+          python-version: '3.12'
+      - run: python -m pip install -r graph_synthesis/requirements-figures.txt
+      - name: Byte-identical figure and complete manuscript regeneration
+        run: |
+          python -B -m graph_synthesis.frontier.report --figures --update-paper
+          git diff --exit-code -- graph_synthesis/frontier manuscript/paper-current.md CURRENT_RESULTS.md
+          python -B -m graph_synthesis.render_current_paper --output "$RUNNER_TEMP/frontier-paper.pdf"
+          git diff --exit-code -- manuscript/paper-current.html

CURRENT_RESULTS.md

@@ -58,6 +58,14 @@ Marginal-only lineage bounds tighten 4.96%, missing the frozen 10% target. Query
 
 <!-- UNCERTAINTY_RESEARCH_END -->
 
+<!-- FRONTIER_RESEARCH_START -->
+
+## Five structural-frontier extensions after PR #17
+
+[Executed report](graph_synthesis/frontier/RESULTS.md), [pre-execution protocol](graph_synthesis/frontier/PROTOCOL.md), [novelty audit](graph_synthesis/frontier/NOVELTY.md), [five figure pairs](graph_synthesis/frontier/figures/) and [complete updated paper](manuscript/paper-current.pdf). H1/H3/H4/H5 meet controlled supplied-input targets; H2 fails its frozen review-quality target. Source-conditioned probabilities prevent nine independent-marginal false admissions on 96 controlled fixtures; all three review selectors tie at 13.043478 expected contaminated groups. Ten large lineage formulas and nine bipartite conflict fixtures are exact. Connected-path summary transitions fall by 96.47%, excluding linear witness decoding. No new Jev calls, independent semantic-accuracy finding, worldwide novelty claim or production-policy change. All earlier studies and negative results remain intact.
+
+<!-- FRONTIER_RESEARCH_END -->
+
 Read the [updated full paper](manuscript/paper-current.md), [PDF](manuscript/paper-current.pdf), or [HTML](manuscript/paper-current.html).
 
 The fresh full Jev run does **not** reproduce the original interval-based entity improvement: its selected-minus-baseline macro-F1 interval includes zero. Entity accuracy still rises descriptively from 405/413 to 408/413; relation accuracy rises from 286/339 to 290/339, also without a resolved macro-F1 effect. This repeats previously inspected data; it is not independent validation.

graph_synthesis/frontier/IMPLEMENTATION.md

@@ -0,0 +1,19 @@
+# Implementation and execution record
+
+The protocol was committed before implementation/execution as `ef8fc308d0a50cd37a6cdd3b7546e050093349c8`, against main `a62a3257645d8e35cd4e45be53bfa9511d27724b`. Seed, populations, comparison methods, budgets and primary endpoints were not retuned after results were observed. These are exploratory extensions informed by published outcomes, not independent preregistration.
+
+The first local execution produced H1=true, H2=false, H3=true, H4=true, H5=true. Subsequent validation tightened malformed atom-ID handling to reject mixed string/nonstring primitive IDs with ValueError before sorting. This does not change any valid-input benchmark output. All 75 new regression tests pass. Complete numerical replay and archived-inventory verification are performed separately; their exact outputs are in the committed logs.
+
+H2 is retained as a negative result even though the modeled optimization objective improves over one comparator. No threshold, detector assumption or cost mapping was chosen using its evaluation outcomes. Controls retain errors under misspecified source metadata, incorrectly prioritized review risks, false high-priority graph assertions, unsupported graph topologies and exhausted state budgets.
+
+`execution.log` is actual execution output. `tests.log` is actual unittest output, including its environment-dependent elapsed time. `replay.log` is the recorded deterministic numerical replay output. `original-inventory.json` is the archived-evidence verifier result. `artifact-manifest.json` binds extension bytes; the current manuscript uses its independent `.build.json` source/renderer/PDF record. The result journal binds the full generating specifications and source/code hashes; it does not claim provider-signed provenance.
+
+No new Jev requests, credentials, network model execution, production graph writes or default compiler changes are made. The renderer uses repository-local image assets only. Python/numpy numeric replay uses the repository's existing explicit float tolerances; SVG/PNG and Markdown regeneration is checked byte-for-byte under pinned rendering dependencies.
+
+Concurrent main integration pinned `a21314d17c33c639b75522ecce120586ef8dab35`. Both intervening studies, their evidence and protocols are preserved. Shared manuscript and renderer conflicts are resolved additively; H4 overlap is explicitly disclosed in the novelty audit and paper. Benchmark code, frozen protocol, fixtures and target indicators remain unchanged. The expanded combined regression count is recorded by all-regressions.json after integration.
+
+The first integrated execution passed all 726 regressions but its CI-token push was rejected for workflow permission. The authorized connected GitHub action then applied the workflow union and genuine two-parent merge; subsequent CI publishing changes research artifacts only. No credential permissions were escalated.
+
+Final integration preserves PRs #18, #19, #24 and #21 at `114450ef08736c21c97c8725de700d8d2c6ca94d`, including their requirements and all evidence. The full regression runner includes all their tests. The reporting anchor follows the latest supported refinement study, before the historical follow-up boundary. The renderer from current main is preserved unchanged because its generic marker-removal helper already handles the frontier section. Frozen benchmark implementations, protocol, fixture seeds and target indicators remain unchanged.
+
+Integration validation recorded 772 passing tests and one manuscript-regeneration failure in the first 773-test run. Choosing the last physical research marker incorrectly placed the new section after the historical follow-up, whose unchanged writer relocates itself immediately before the original-study boundary. The correction selects the latest refinement predecessor, preserving the historical follow-up position. The existing failing regression remains unchanged and must pass before publication. No research result or acceptance threshold changed.