results: just results = eval2 deploy table (time/headline/deploy/arm/pair/seed/train/argv); hard eval2 cutoff; archive eval1 (Q1-Q13 + 352 old logs)

scripts/results.py

@@ -20,6 +20,11 @@ from tabulate import tabulate
 
 LOG_DIR = Path("logs")
 TS_RE = re.compile(r"(\d{8}T\d{6})")
+# Hard cutoff: only show eval2-era runs (recency-clean test set, dir6+ onward). Runs before
+# this are the OLD eval (contaminated holdout); their curated findings live in
+# docs/results_eval1_archive.md. Robust to old logs being present -- filters by the log's
+# own timestamp, so we don't rely on moving files out of logs/.
+EVAL2_CUTOFF = "20260607T000000"
 # Column positions are read from the header row by NAME, not hardcoded -- the
 # per-step table layout has changed over time (sprd/N dropped, cin/cout/hk_dep
 # added) so fixed indices silently mis-read newer logs and crash on smoke logs.
@@ -65,6 +70,9 @@ def _cfg(argv: str, preset_line: str) -> dict:
 
 
 def parse_log(path: Path) -> dict | None:
+    ts_m = TS_RE.search(path.name)
+    if ts_m and ts_m.group(1) < EVAL2_CUTOFF:
+        return None  # pre-eval2 (OLD eval) -> docs/results_eval1_archive.md
     txt = path.read_text(errors="replace")
     argv = next((l for l in txt.splitlines() if "argv:" in l), None)
     preset_line = next((l for l in txt.splitlines() if "preset=" in l and "arm=" in l), "")