Merge pull request #234 from acgetchell/feat/222-weighted-proposal-policies

feat(cdt): add weighted move-family proposal policies

Author: acgetchell

.github/workflows/audit.yml

@@ -75,11 +75,11 @@ jobs:
           tool: cargo-audit@${{ steps.cargo_audit_version.outputs.version }}
 
       - name: Run cargo audit
-        run: |
-          cargo audit --json > audit-results.json || true
-          # Show human-readable output as well (reuses cached advisory DB)
-          cargo audit --no-fetch
-        continue-on-error: true
+        run: cargo audit
+
+      - name: Write cargo audit JSON
+        if: always()
+        run: cargo audit --json --no-fetch > audit-results.json
 
       - name: Upload audit results
         uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1

.github/workflows/ci.yml

@@ -61,6 +61,9 @@ jobs:
           cache-bin: false # Avoid restoring stale rustup/cargo shims from cache.
       # toolchain, components, etc. are specified in rust-toolchain.toml
 
+      - name: Check declared Rust target
+        run: cargo check --locked --workspace --target ${{ matrix.target }} --all-features --lib --bins
+
       - name: Set up just
         uses: ./.github/actions/setup-just
 

.github/workflows/dependabot-auto-merge.yml

@@ -104,4 +104,4 @@ jobs:
           APPROVED_HEAD_SHA: ${{ steps.coderabbit-approval.outputs.approved_head_sha }}
           GH_TOKEN: ${{ github.token }}
           PR_URL: ${{ github.event.pull_request.html_url }}
-        run: gh pr merge --squash --match-head-commit "$APPROVED_HEAD_SHA" "$PR_URL"
+        run: gh pr merge --auto --squash --match-head-commit "$APPROVED_HEAD_SHA" "$PR_URL"

.github/workflows/performance.yml

@@ -137,6 +137,10 @@ jobs:
             performance-data-${{ runner.os }}-
             performance-baselines-main-
 
+      - name: Verify deterministic allocation contract
+        if: steps.detect_bench.outputs.bench_found == 'true'
+        run: just allocation-check
+
       # For main branch: save baseline after successful benchmarks
       - name: Run benchmarks and save baseline (main branch)
         if: github.ref == 'refs/heads/main' && github.event_name == 'push' && steps.detect_bench.outputs.bench_found == 'true'

AGENTS.md

@@ -112,6 +112,14 @@ When using the `gh` CLI to view issues, PRs, or other GitHub objects:
 - Keep production module preambles free of test-only imports; place `#[cfg(test)]` imports inside the relevant `tests` module instead
 - Detailed import guidance lives in `docs/dev/rust.md`
 
+### Geometry Validation Ownership
+
+- Delegate structural geometry validation—including incidence, orientation, manifold, embedding, and Euler-topology checks—to `delaunay` through the
+  `src/geometry/` adapter whenever its API supports the check. Do not reimplement those algorithms in `src/cdt/`.
+- CDT validation owns only domain invariants that the geometry backend cannot express, such as foliation, causality, and agreement between CDT metadata and
+  backend geometry. Add a narrow geometry trait/adapter method when CDT needs access to an upstream check.
+- Detailed validation-boundary guidance lives in `docs/dev/rust.md`.
+
 ### Public API Preludes
 
 - Keep `prelude::*` small and focused on common quick-start workflows.

Cargo.lock

@@ -34,6 +34,10 @@ harness = false
 name = "ci_performance_suite"
 harness = false
 
+[[bench]]
+name = "allocation_profile"
+harness = false
+
 [features]
 slow-tests = [  ]
 
@@ -66,6 +70,7 @@ approx = "0.5.1"
 slotmap = "1.1.1"
 assert_cmd = "2.2.2"
 criterion = { version = "0.8.2", features = [ "html_reports" ] }
+dhat = "0.3.3"
 predicates = "3.1.4"
 proptest = "1.11.0"
 

Cargo.toml

@@ -50,7 +50,7 @@ path integral over causally triangulated spacetimes and evaluating it using Mark
 - Resumable serde-backed CDT/MCMC checkpoints for continuation within a compatible build; serialized checkpoint files are
   [version-bound](docs/metropolis.md#serialized-checkpoint-compatibility)
 - Trace CSV simulation output for external analysis workflows; JSON summary/metadata for CLI/config export
-- Volume-profile, Hausdorff-dimension, and spectral-dimension observables for CDT analysis
+- Spatial-vertex input profiles, slab-triangle output profiles, and explicitly finite-window effective dimensional observables
 
 See [CHANGELOG.md](CHANGELOG.md) for release history and [`docs/roadmap.md`](docs/roadmap.md) for current direction, near-term candidates, and non-goals.
 

README.md

@@ -9,6 +9,7 @@ This document describes the Criterion benchmark suites. For regression workflows
 cargo bench
 just bench-ci
 cargo bench --profile perf --bench ci_performance_suite
+cargo bench --profile perf --bench allocation_profile
 ```
 
 Run a focused group:
@@ -35,12 +36,19 @@ that should stay comparable across releases:
 - generating open-boundary and toroidal CDT triangulations;
 - validating generated triangulations;
 - attempting individual ergodic move types;
+- scaling guaranteed-success inverse volume finalization across increasing toroidal CDT meshes;
 - iterating proposal-site candidates through public move-attempt and single-step Metropolis proposal paths;
-- executing ten random-move sweeps, where each sweep attempts one move per current simplex;
+- evaluating state-dependent family policies that inspect the complete offered-site views;
+- executing fixed random-move attempt budgets equal to ten initial sweeps, so reported throughput exactly matches timed attempts;
 - running short Metropolis simulations sized as ten initial sweeps.
 
 Keep this suite stable and release-relevant. Exploratory or noisy benchmarks belong in `cdt_benchmarks.rs`.
 
+`allocation_profile` is a deterministic heap-allocation contract for cached
+observables. It verifies that a cached edge-count read allocates nothing and a
+borrow-to-owned slab-triangle-profile read performs exactly one vector allocation. The
+contract blocks `just ci`, `just bench-ci`, and the performance workflow.
+
 ## Benchmark Groups
 
 ### `triangulation_creation`
@@ -101,7 +109,7 @@ Includes configuration validation, proposal planning, accepted-move application,
 
 ### `simulation_analysis`
 
-Measures post-run analysis such as acceptance rates, average action, and equilibrium measurement extraction.
+Measures post-run analysis such as acceptance rates, average action, and post-thermalization measurement extraction.
 
 Use for output and analysis workflow regressions.