Refactor example_db: stored ground truth, human-friendly JSON, no static fixture

.claude/skills/final-review/SKILL.md

@@ -150,12 +150,12 @@ Use `AskUserQuestion` with your recommendation:
 
 Scan the PR diff for dangerous actions:
 
-- **Blacklisted files**: If the diff touches `docs/src/reductions/reduction_graph.json` or `docs/src/reductions/problem_schemas.json`, **block merge**. These files are auto-generated and must not be committed in PRs — they are rebuilt by CI/`make doc`. Flag immediately and recommend OnHold.
+- **Blacklisted files**: If the diff touches `docs/src/reductions/reduction_graph.json`, `docs/src/reductions/problem_schemas.json`, or `src/example_db/fixtures/examples.json` (legacy, no longer exists), **block merge**. These files are auto-generated and must not be committed in PRs — they are rebuilt by CI/`make doc`/`make paper`. Flag immediately and recommend OnHold.
 - **Removed features**: Any existing model, rule, test, or example deleted?
 - **Unrelated changes**: Files modified that don't belong to this PR (e.g., changes to unrelated models/rules, CI config, Cargo.toml dependency changes not needed for this PR)
 - **Force push indicators**: Any sign of history rewriting
 - **Broad modifications**: Changes to core traits, macros, or shared infrastructure that could affect other features
-- **`examples.json` one-line rule**: Each PR should add exactly one model or one rule entry to `examples.json`. Check the **PR diff** (`gh pr diff`), not the post-merge-with-main diff — after merging main, fixture regeneration may touch many lines from other PRs. Only flag if the PR's own changes add or modify more than one entry.
+- **No committed `examples.json`**: The example database is generated on demand by `make paper` (via `export_examples`). PRs should not commit `src/example_db/fixtures/examples.json` (legacy path, deleted) or `docs/paper/data/examples.json` (current output path) — both are gitignored build artifacts.
 
 Report findings with fix options for each concern:
 
@@ -217,10 +217,10 @@ Verify the PR includes all required components:
 
 **Paper-example consistency check (both Model and Rule PRs):**
 
-The paper example must use data from the canonical fixture JSON (`src/example_db/fixtures/examples.json`), not hand-written data. To verify:
-1. If the PR changes example builders/specs, run `make regenerate-fixtures` on the PR branch.
+The paper example must use data from the canonical example database (generated on demand by `make paper` via `export_examples`), not hand-written data. To verify:
+1. If the PR changes example specs, run `make paper` to regenerate `docs/paper/data/examples.json`.
 2. For **[Rule] PRs**: the paper's `reduction-rule` entry must call `load-example(source, target, ...)` (defined in `reductions.typ`) to load the canonical example from `examples.json`, and derive all concrete values from the loaded data using Typst array operations — no hand-written instance data.
-3. For **[Model] PRs**: read the problem's entry in `examples.json` under `models` and compare its `instance` field against the paper's `problem-def` example. The paper example must use the same instance (allowing 0-indexed JSON vs 1-indexed math notation). If they differ, flag: "Paper example does not match `example_db` canonical instance in `examples.json`."
+3. For **[Model] PRs**: run the export and read the problem's entry in the generated `examples.json` under `models`, compare its `instance` field against the paper's `problem-def` example. The paper example must use the same instance (allowing 0-indexed JSON vs 1-indexed math notation). If they differ, flag: "Paper example does not match `example_db` canonical instance."
 
 **Issue–test round-trip consistency check (both Model and Rule PRs):**
 
@@ -338,13 +338,32 @@ Use `AskUserQuestion`:
    ```bash
    gh pr review <number> --approve || true
    ```
-5. Present the PR link for the reviewer to merge:
-   > CI green, commits pushed, PR approved. Please merge when ready:
+5. Post a community call validation checklist as a comment on the **linked issue** (not the PR). All CLI commands must be copy-pastable — substitute actual problem names from the PR diff (no angle-bracket placeholders). Example for a rule PR adding `Satisfiability` → `MaximumIndependentSet`:
+   ````bash
+   COMMENT_FILE=$(mktemp)
+   cat > "$COMMENT_FILE" <<'EOF'
+   Please kindly check the following items (PR #123):
+   - [ ] **Paper** ([PDF](https://github.com/CodingThrust/problem-reductions/blob/main/docs/paper/reductions.pdf)): check definition, proof sketch, and example figure
+   - [ ] **CLI demo** (build from source: `cargo install --path problemreductions-cli`):
+     ```bash
+     pred show Satisfiability
+     pred create --example Satisfiability -o instance.json
+     pred reduce instance.json Satisfiability MaximumIndependentSet -o reduced.json
+     pred solve reduced.json MaximumIndependentSet
+     ```
+   - [ ] **Implementation (Optional)**: spot-check the source files changed in this PR for correctness
+   EOF
+   gh issue comment <ISSUE_NUMBER> --body-file "$COMMENT_FILE"
+   rm -f "$COMMENT_FILE"
+   ````
+   For model PRs, omit the `pred reduce` / `pred solve` lines. If there is no linked issue, post the checklist as a PR comment instead.
+6. Present the PR link for the reviewer to merge:
+   > CI green, commits pushed, PR approved. Community call checklist posted on #<ISSUE_NUMBER>. Please merge when ready:
    > **<PR URL>**
-6. After the reviewer merges, use `AskUserQuestion` to confirm:
+7. After the reviewer merges, use `AskUserQuestion` to confirm:
    > **Merged? (continue to move card & cleanup worktree)** Once confirmed, I will move the board item to Done and clean up the worktree.
    > - "Yes" — proceed with cleanup
-7. Move the project board item to `Done` and clean up:
+8. Move the project board item to `Done` and clean up:
    ```bash
    python3 scripts/pipeline_board.py move <ITEM_ID> done
    cd "$REPO_ROOT"

.claude/skills/review-pipeline/SKILL.md

@@ -125,11 +125,15 @@ The two expensive context calls are allowed exactly once each per top-level `rev
 Branch from the review-pipeline report:
 - `Bundle status: empty` => the selected PR is no longer eligible; run `cd "$REPO_ROOT" && python3 scripts/pipeline_worktree.py cleanup --worktree "$WORKTREE_DIR"`, then for untargeted runs return to Step 0a, for explicit `PR` runs STOP
 - `Bundle status: needs-user-choice` => run `cd "$REPO_ROOT" && python3 scripts/pipeline_worktree.py cleanup --worktree "$WORKTREE_DIR"`, STOP and ask the user which PR is intended
-- `Bundle status: ready` => continue
+- `Bundle status: ready` => claim the item and continue
 
-The bundle already handled the mechanical claim step (moved to Under review). Use the identifiers from the report for all subsequent operations.
+**Claim the item** (move to Under review) only after confirming `Bundle status: ready`:
 
-All subsequent steps run inside the worktree and should read facts from the reports instead of re-fetching them.
+```bash
+python3 scripts/pipeline_board.py move <ITEM_ID> under-review
+```
+
+Use the identifiers from the report for all subsequent operations. All subsequent steps run inside the worktree and should read facts from the reports instead of re-fetching them.
 
 ### 1. Run Three Sub-Reviews (Parallel)
 

.github/ISSUE_TEMPLATE/problem.md

@@ -95,6 +95,9 @@ This example will be shown in our paper, where you could find some references.
 <!--
 Optimization: provide one optimal configuration and its objective value.
 Satisfaction: provide one valid / satisfying configuration and a brief justification.
+
+This will be stored as ground truth in the example database and cross-validated
+by an ILP solver.
 -->
 
 ## BibTeX

.gitignore

@@ -78,13 +78,13 @@ pkgref/
 
 # Generated example outputs
 docs/paper/examples/
+docs/paper/data/
 
 # Claude Code logs
 claude-output.log
 .worktrees/
 .worktree/
 *.json
-!src/example_db/fixtures/*.json
 .claude/worktrees/
 docs/test-reports/
 docs/superpowers/

Cargo.toml

@@ -46,8 +46,8 @@ name = "solver_benchmarks"
 harness = false
 
 [[example]]
-name = "regenerate_fixtures"
-path = "examples/regenerate_fixtures.rs"
+name = "export_examples"
+path = "examples/export_examples.rs"
 required-features = ["example-db"]
 
 [profile.release]

Makefile

@@ -1,6 +1,6 @@
 # Makefile for problemreductions
 
-.PHONY: help build test mcp-test fmt clippy doc mdbook paper clean coverage rust-export compare qubo-testdata export-schemas release run-plan run-issue run-pipeline run-pipeline-forever run-review run-review-forever board-next board-claim board-ack board-move issue-context issue-guards pr-context pr-wait-ci worktree-issue worktree-pr diagrams jl-testdata cli cli-demo copilot-review regenerate-fixtures
+.PHONY: help build test mcp-test fmt clippy doc mdbook paper clean coverage rust-export compare qubo-testdata export-schemas release run-plan run-issue run-pipeline run-pipeline-forever run-review run-review-forever board-next board-claim board-ack board-move issue-context issue-guards pr-context pr-wait-ci worktree-issue worktree-pr diagrams jl-testdata cli cli-demo copilot-review
 
 RUNNER ?= codex
 CLAUDE_MODEL ?= opus
@@ -24,7 +24,6 @@ help:
 	@echo "  check        - Quick check (fmt + clippy + test)"
 	@echo "  rust-export  - Generate Rust mapping JSON exports"
 	@echo "  compare      - Generate and compare Rust mapping exports"
-	@echo "  regenerate-fixtures - Recompute example DB fixtures (BruteForce/ILP, slow)"
 	@echo "  export-schemas - Export problem schemas to JSON"
 	@echo "  qubo-testdata - Regenerate QUBO test data (requires uv)"
 	@echo "  jl-testdata  - Regenerate Julia parity test data (requires julia)"
@@ -113,16 +112,14 @@ mdbook:
 	python3 -m http.server 3001 -d book &
 	@sleep 1 && (command -v xdg-open >/dev/null && xdg-open http://localhost:3001 || open http://localhost:3001)
 
-# Regenerate example DB fixtures from code (runs BruteForce/ILP — slow)
-regenerate-fixtures:
-	cargo run --release --features "ilp-highs example-db" --example regenerate_fixtures
 
 # Export problem schemas to JSON
 export-schemas:
 	cargo run --example export_schemas
 
-# Build Typst paper (reads canonical examples directly from fixtures)
+# Build Typst paper (generates example data on demand)
 paper:
+	cargo run --features "example-db" --example export_examples
 	cargo run --example export_petersen_mapping
 	cargo run --example export_graph
 	cargo run --example export_schemas