优化6个类，现在ks降低到0.28，史称3.0版本

example/program_stats.py

@@ -0,0 +1,142 @@
+#!/usr/bin/env python3
+"""Compute program-style stats (dwell, change count, step size) for selected features."""
+
+import argparse
+import csv
+import gzip
+import json
+from pathlib import Path
+from typing import Dict, List
+
+
+def parse_args():
+    base_dir = Path(__file__).resolve().parent
+    parser = argparse.ArgumentParser(description="Program stats for setpoints/demands.")
+    parser.add_argument("--generated", default=str(base_dir / "results" / "generated.csv"))
+    parser.add_argument("--reference", default=str(base_dir / "config.json"))
+    parser.add_argument("--features", default="", help="comma-separated list; empty = auto from eval")
+    parser.add_argument("--config", default=str(base_dir / "config.json"))
+    parser.add_argument("--out", default=str(base_dir / "results" / "program_stats.json"))
+    parser.add_argument("--max-rows", type=int, default=200000)
+    return parser.parse_args()
+
+
+def resolve_reference_glob(ref_arg: str) -> str:
+    ref_path = Path(ref_arg)
+    if ref_path.suffix == ".json":
+        cfg = json.loads(ref_path.read_text(encoding="utf-8"))
+        data_glob = cfg.get("data_glob") or cfg.get("data_path") or ""
+        if not data_glob:
+            raise SystemExit("reference config has no data_glob/data_path")
+        combined = ref_path.parent / data_glob
+        # avoid resolve on glob patterns
+        if "*" in str(combined) or "?" in str(combined):
+            return str(combined)
+        return str(combined.resolve())
+    return str(ref_path)
+
+
+def read_series(path: Path, cols: List[str], max_rows: int) -> Dict[str, List[float]]:
+    vals = {c: [] for c in cols}
+    opener = gzip.open if str(path).endswith(".gz") else open
+    with opener(path, "rt", newline="") as fh:
+        reader = csv.DictReader(fh)
+        for i, row in enumerate(reader):
+            for c in cols:
+                try:
+                    vals[c].append(float(row[c]))
+                except Exception:
+                    pass
+            if max_rows > 0 and i + 1 >= max_rows:
+                break
+    return vals
+
+
+def dwell_and_steps(series: List[float]):
+    if not series:
+        return {
+            "num_changes": 0,
+            "mean_dwell": None,
+            "median_dwell": None,
+            "mean_step": None,
+            "median_step": None,
+        }
+    changes = 0
+    dwells = []
+    steps = []
+    current = series[0]
+    dwell = 1
+    for v in series[1:]:
+        if v == current:
+            dwell += 1
+            continue
+        changes += 1
+        dwells.append(dwell)
+        steps.append(abs(v - current))
+        current = v
+        dwell = 1
+    dwells.append(dwell)
+
+    def mean(x):
+        return sum(x) / len(x) if x else None
+
+    def median(x):
+        if not x:
+            return None
+        xs = sorted(x)
+        mid = len(xs) // 2
+        return xs[mid] if len(xs) % 2 == 1 else 0.5 * (xs[mid - 1] + xs[mid])
+
+    return {
+        "num_changes": changes,
+        "mean_dwell": mean(dwells),
+        "median_dwell": median(dwells),
+        "mean_step": mean(steps),
+        "median_step": median(steps),
+    }
+
+
+def main():
+    args = parse_args()
+    out_path = Path(args.out)
+
+    eval_path = Path("results") / "eval.json"
+    auto_feats = []
+    if eval_path.exists():
+        data = json.loads(eval_path.read_text(encoding="utf-8"))
+        ks = data.get("continuous_ks", {})
+        auto_feats = [k for k, v in ks.items() if v >= 0.6]
+
+    features = [f.strip() for f in args.features.split(",") if f.strip()] or auto_feats
+    if not features and Path(args.config).exists():
+        cfg = json.loads(Path(args.config).read_text(encoding="utf-8"))
+        features = cfg.get("type1_features", []) or []
+    if not features:
+        raise SystemExit("no features specified and no eval.json with ks>=0.6")
+
+    # generated series
+    gen_vals = read_series(Path(args.generated), features, args.max_rows)
+
+    # reference series (aggregate across files)
+    ref_glob = resolve_reference_glob(args.reference)
+    ref_paths = sorted(Path(ref_glob).parent.glob(Path(ref_glob).name))
+    if not ref_paths:
+        raise SystemExit(f"no reference files matched: {ref_glob}")
+    real_vals = {c: [] for c in features}
+    for p in ref_paths:
+        vals = read_series(p, features, args.max_rows)
+        for c in features:
+            real_vals[c].extend(vals[c])
+
+    out = {"features": features, "generated": {}, "reference": {}}
+    for c in features:
+        out["generated"][c] = dwell_and_steps(gen_vals[c])
+        out["reference"][c] = dwell_and_steps(real_vals[c])
+
+    out_path.parent.mkdir(parents=True, exist_ok=True)
+    out_path.write_text(json.dumps(out, indent=2), encoding="utf-8")
+    print("wrote", out_path)
+
+
+if __name__ == "__main__":
+    main()