Rename hard subset metadata to metadata/hardset

Move the five hard subset metadata files, rename hard.json to hardset.json, and update the reproduction script and dataset documentation. Task IDs and ranking data are unchanged.

README.md

@@ -52,7 +52,7 @@ paper's OR problem into runnable, verifiably-correct optimization code.
 | `metadata/` | Paper metadata, Gurobi reference summaries, and benchmark split metadata |
 | `metadata/paper_meta_info.json` | Metadata for all 180 canonical papers, including optimization direction |
 | `metadata/gurobi_references.parquet` / `.csv.gz` | Canonical Gurobi objectives, runtimes, statuses, and related reference fields |
-| `metadata/splits/` | Hard-set ids, ranking inputs, methodology, and reproduction script |
+| `metadata/hardset/` | Hard-set ids, ranking inputs, methodology, and reproduction script |
 | `tasks/<task_id>/` | Self-contained task package for one paper |
 
 ## What's in each task folder

RELEASE_NOTES.md

@@ -10,10 +10,10 @@ This update consolidates the dataset, reference results, validation logic, and e
 - Paper-level metadata, including optimization direction, is consolidated in:
   - `metadata/paper_meta_info.json`
 - The canonical 50-task hard subset is published in:
-  - `metadata/splits/hard.json`
+  - `metadata/hardset/hardset.json`
 - Its selection criteria, frozen input snapshot, full 180-task ranking, and
-  reproduction script are published in `metadata/splits/` (see
-  `metadata/splits/hard_methodology.md`).
+  reproduction script are published in `metadata/hardset/` (see
+  `metadata/hardset/hard_methodology.md`).
 
 
 ### Canonical Gurobi references
@@ -61,7 +61,7 @@ Users of earlier FrontierOR snapshots should:
 
 - update task paths to `tasks/<task_id>/`;
 - use `metadata/paper_meta_info.json` for paper metadata;
-- use `metadata/splits/hard.json` for the hard subset;
+- use `metadata/hardset/hardset.json` for the hard subset;
 - use `metadata/gurobi_references.parquet` or `.csv.gz` for Gurobi references;
 - rerun feasibility checks when relying on results produced by older checker versions.
 

metadata/{splits → hardset}/hard_methodology.md

@@ -1,19 +1,19 @@
 # FrontierOR hard-50 selection
 
-The published `hard.json` is a **frozen 50-task subset** of the 180 canonical
+The published `hardset.json` is a **frozen 50-task subset** of the 180 canonical
 tasks. It was selected using task metadata, instance-size summaries, and Gurobi
 reference performance.
 
 ## Published evidence
 
-- `hard.json`: authoritative selected task IDs in descending hard-rank order.
+- `hardset.json`: authoritative selected task IDs in descending hard-rank order.
 - `hard_selection_inputs_180.csv`: frozen, one-row-per-task input snapshot used
   for this selection. It includes the relevant paper metadata, mean instance
   size fields, and historical five-large-instance Gurobi summary statistics.
 - `hard_ranking_180.csv`: derived scores, rule flags, rank, and membership for
   all 180 tasks, including the 130 tasks not selected.
 - `reproduce_hard.py`: recomputes the ranks from the frozen inputs and checks
-  the selected set against `hard.json`.
+  the selected set against `hardset.json`.
 
 The frozen Gurobi summary fields came from the benchmark's historical
 `gurobi_results_11/21/31/41/51.csv` analysis. In that analysis, `time_out`
@@ -86,14 +86,14 @@ scale score (all descending), then by `paper_id` (ascending); take the first
 With Python, `pandas`, and `numpy` installed, run from the dataset root:
 
 ```bash
-python metadata/splits/reproduce_hard.py
+python metadata/hardset/reproduce_hard.py
 ```
 
 To materialize the computed ranking for comparison:
 
 ```bash
-python metadata/splits/reproduce_hard.py --ranking-out /tmp/hard_ranking_180.csv
+python metadata/hardset/reproduce_hard.py --ranking-out /tmp/hard_ranking_180.csv
 ```
 
 The script verifies that the computed top 50 IDs and their ranking order
-match `hard.json`.
+match `hardset.json`.

metadata/{splits → hardset}/reproduce_hard.py

@@ -98,7 +98,7 @@ def rank_tasks(inputs: pd.DataFrame) -> pd.DataFrame:
 def main() -> int:
     parser = argparse.ArgumentParser(description=__doc__)
     parser.add_argument("--inputs", type=Path, default=HERE / "hard_selection_inputs_180.csv")
-    parser.add_argument("--hard-json", type=Path, default=HERE / "hard.json")
+    parser.add_argument("--hard-json", type=Path, default=HERE / "hardset.json")
     parser.add_argument("--ranking-out", type=Path, help="Optional path for the recomputed ranking CSV")
     args = parser.parse_args()
 
@@ -122,7 +122,7 @@ def main() -> int:
               f"published-only={sorted(selected - set(computed))}")
         return 1
     print(f"Reproduced {len(selected)} hard tasks from {len(ranking)} frozen input rows.")
-    print("Task IDs and ranking order match hard.json.")
+    print("Task IDs and ranking order match hardset.json.")
     return 0