| #!/usr/bin/env python3 | |
| """summarize.py - one line per clip from the per-second JSONL.""" | |
| import json, sys | |
| from collections import defaultdict | |
| clips = defaultdict(list) | |
| for line in open(sys.argv[1]): | |
| r = json.loads(line); clips[r["clip_id"]].append(r) | |
| print(f"{'clip_id':<13}{'secs':>5}{'fps':>5}{'luma':>6}{'sat':>6}{'motion':>8}{'busiest_s':>10}{'rms_db':>7}{'offset':>7}{'match':>6} name==sha256") | |
| for cid, rows in clips.items(): | |
| r0 = rows[0] | |
| busiest = max(rows, key=lambda r: r["motion_max"])["second"] | |
| avg = lambda k: sum(r[k] for r in rows) / len(rows) | |
| ok = r0["source_url"].rsplit("/", 1)[1].split(".")[0] == r0["sha256"] | |
| print(f"{cid:<13}{len(rows):>5}{r0['fps']:>5.0f}{avg('luma_mean'):>6.2f}{avg('saturation_mean'):>6.2f}" | |
| f"{avg('motion_mean'):>8.4f}{busiest:>10}{avg('audio_rms_dbfs'):>7.1f}{r0['soundtrack_offset_s']:>+7.2f}" | |
| f"{r0['soundtrack_match']:>6.2f} {ok}") | |