Upload scripts/verify_full_dataset.py

scripts/verify_full_dataset.py

@@ -0,0 +1,78 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""Quick integrity checks for the full SeismicX-Cont release folder."""
+
+from __future__ import annotations
+
+import argparse
+import json
+import sqlite3
+from pathlib import Path
+
+import h5py
+
+
+def count_h5_segments(path: Path) -> tuple[int, int, int]:
+    segments = 0
+    stations = set()
+    channels = set()
+    with h5py.File(path, "r") as h5:
+        for year_id in h5:
+            year = h5[year_id]
+            if not isinstance(year, h5py.Group):
+                continue
+            for day_id in year:
+                day = year[day_id]
+                if not isinstance(day, h5py.Group) or "stations" not in day:
+                    continue
+                for station_id, station_grp in day["stations"].items():
+                    stations.add(station_id)
+                    waveform = station_grp.get("waveform")
+                    if waveform is None:
+                        continue
+                    for channel, channel_grp in waveform.items():
+                        if not isinstance(channel_grp, h5py.Group):
+                            continue
+                        channels.add(channel)
+                        segments += sum(1 for key in channel_grp if key.isdigit())
+    return segments, len(stations), len(channels)
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("--root", type=Path, default=Path(__file__).resolve().parents[1])
+    parser.add_argument("--scan-hdf5", action="store_true", help="Open each HDF5 file and count waveform segments.")
+    args = parser.parse_args()
+
+    root = args.root
+    h5_files = sorted((root / "data" / "hdf5").glob("continuous_waveform_usa_*.h5"))
+    label_json = root / "data" / "label" / "annotations_for_continuous_hdf5.json"
+    db_path = root / "data" / "index" / "waveform_index.sqlite"
+
+    print(f"HDF5 files: {len(h5_files)}")
+    print(f"HDF5 total size: {sum(p.stat().st_size for p in h5_files) / 1024**3:.2f} GiB")
+    if args.scan_hdf5:
+        total_segments = 0
+        for path in h5_files:
+            segments, stations, channels = count_h5_segments(path)
+            total_segments += segments
+            print(f"  {path.name}: segments={segments:,}, stations={stations:,}, channels={channels:,}")
+        print(f"HDF5 scanned segments: {total_segments:,}")
+
+    with label_json.open("r", encoding="utf-8") as f:
+        labels = json.load(f)
+    print("Annotation summary:")
+    print(json.dumps(labels.get("summary", {}), ensure_ascii=False, indent=2))
+
+    conn = sqlite3.connect(db_path)
+    cur = conn.cursor()
+    n_files = cur.execute("SELECT COUNT(*) FROM hdf5_files").fetchone()[0]
+    n_segments = cur.execute("SELECT COUNT(*) FROM waveform_segments").fetchone()[0]
+    n_stations = cur.execute("SELECT COUNT(*) FROM stations").fetchone()[0]
+    conn.close()
+    print("SQLite index:")
+    print(f"  files={n_files}, waveform_segments={n_segments:,}, stations={n_stations:,}")
+
+
+if __name__ == "__main__":
+    main()