Keep the ablation arms inside the line budget the repo enforces

copy_store() moved into bench/scratch.py: systems.py had reached 512 lines
and bench/filesize_check fails anything over 500. The move is behaviour
identical, so the scale reports keep the commit stamp they were generated
against — the code they describe is complete at that commit.

bench/dense_scale.py

@@ -21,7 +21,8 @@ import numpy as np
 from bench.query_set import load_queries, scale_probes
 from bench.report import git_commit
 from bench.run import latency
-from bench.systems import DENSE_MODEL, MODELS, copy_store, machine_facts
+from bench.scratch import copy_store
+from bench.systems import DENSE_MODEL, MODELS, machine_facts
 
 if TYPE_CHECKING:
     from ely_eye.db import Database

bench/leg_check.py

@@ -25,7 +25,8 @@ from bench.metrics import aggregate, score_query
 from bench.query_set import load_queries, scale_probes
 from bench.report import git_commit
 from bench.run import latency
-from bench.systems import copy_store, machine_facts
+from bench.scratch import copy_store
+from bench.systems import machine_facts
 
 DEFAULT_HOME = Path("bench/runs/dense_scale")
 DEFAULT_SCRATCH = Path("bench/runs/leg_check")

bench/scale_check.py

@@ -17,7 +17,8 @@ from typing import Any
 from bench.query_set import load_queries, scale_probes
 from bench.report import git_commit
 from bench.run import latency
-from bench.systems import copy_store, machine_facts
+from bench.scratch import copy_store
+from bench.systems import machine_facts
 
 DEFAULT_HOME = Path("bench/runs/baseline")
 DEFAULT_SCRATCH = Path("bench/runs/scale")

bench/scratch.py

@@ -0,0 +1,24 @@
+"""Every harness measures a copy. The frozen baseline store is cited evidence, and the product
+migrates whatever database it opens."""
+
+from __future__ import annotations
+
+import shutil
+import time
+from pathlib import Path
+
+
+def copy_store(home: Path, scratch: Path) -> tuple[Path, float]:
+    """Work on a copy: the baseline store is frozen evidence and must not be migrated in place."""
+    source = home / "data" / "ely_eye.sqlite"
+    if not source.is_file():
+        raise SystemExit(
+            f"{source}: no product database, build it with python -m bench.ingest_baseline"
+        )
+    data = scratch / "data"
+    data.mkdir(parents=True, exist_ok=True)
+    began = time.perf_counter()
+    for suffix in ("", "-wal", "-shm"):
+        if (path := Path(str(source) + suffix)).is_file():
+            shutil.copyfile(path, data / (source.name + suffix))
+    return scratch, time.perf_counter() - began

bench/systems.py

@@ -13,7 +13,6 @@ import json
 import os
 import platform
 import re
-import shutil
 import sqlite3
 import time
 from collections.abc import Sequence
@@ -24,6 +23,7 @@ from typing import TYPE_CHECKING, Any
 import numpy as np
 
 from bench.ingest_baseline import source_map
+from bench.scratch import copy_store
 
 if TYPE_CHECKING:
     from ely_eye.db import Database
@@ -446,20 +446,6 @@ def _corpus_ids(database: Database) -> dict[str, str]:
     return ids
 
 
-def copy_store(home: Path, scratch: Path) -> tuple[Path, float]:
-    """Work on a copy: the baseline store is frozen evidence and must not be migrated in place."""
-    source = home / "data" / "ely_eye.sqlite"
-    if not source.is_file():
-        raise SystemExit(f"{source}: no product database, build it with python -m bench.ingest_baseline")
-    data = scratch / "data"
-    data.mkdir(parents=True, exist_ok=True)
-    began = time.perf_counter()
-    for suffix in ("", "-wal", "-shm"):
-        if (path := Path(str(source) + suffix)).is_file():
-            shutil.copyfile(path, data / (source.name + suffix))
-    return scratch, time.perf_counter() - began
-
-
 def live_product_rankings(
     db_home: Path,
     queries: Sequence[str],

bench/term_sweep.py

@@ -20,7 +20,8 @@ from bench.metrics import aggregate, score_query
 from bench.query_set import load_queries, scale_probes
 from bench.report import git_commit
 from bench.run import latency
-from bench.systems import copy_store, live_product_rankings, machine_facts
+from bench.scratch import copy_store
+from bench.systems import live_product_rankings, machine_facts
 
 FRACTIONS = [0.20, 0.10, 0.05, 0.02, 0.01, 0.005]
 TOP_K = 10