File size: 6,386 Bytes
5fb1693 6b333a3 5fb1693 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 | #!/usr/bin/env python3
"""Download source videos and reconstruct annotation-aligned Sekai2 clips."""
from __future__ import annotations
import argparse
import csv
import json
import shutil
import subprocess
from collections import defaultdict
from pathlib import Path
DEFAULT_FORMAT = (
"bestvideo[height<=720][fps<=30]+bestaudio/"
"best[height<=720][fps<=30]"
)
def run(command: list[str]) -> None:
subprocess.run(command, check=True)
def read_rows(path: Path, clip_ids: set[str], dataset: str | None, limit: int | None):
rows = []
with path.open(encoding="utf-8-sig", newline="") as handle:
for row in csv.DictReader(handle):
if dataset and row["dataset"] != dataset:
continue
if clip_ids and row["clip_id"] not in clip_ids:
continue
rows.append(row)
if limit is not None and len(rows) >= limit:
break
if clip_ids:
found = {row["clip_id"] for row in rows}
missing = sorted(clip_ids - found)
if missing:
raise ValueError(f"clip_id not found: {missing}")
return rows
def find_cached_source(cache_dir: Path, video_id: str) -> Path | None:
candidates = [
path
for path in cache_dir.glob(video_id + ".*")
if path.suffix not in {".json", ".part", ".ytdl"}
]
return sorted(candidates)[0] if candidates else None
def download_source(
row: dict[str, str], cache_dir: Path, yt_dlp: str, format_selector: str
) -> Path:
existing = find_cached_source(cache_dir, row["video_id"])
if existing:
return existing
template = str(cache_dir / (row["video_id"] + ".%(ext)s"))
command = [
yt_dlp,
"--no-playlist",
"--no-part",
"--write-info-json",
"--merge-output-format",
"mkv",
"-f",
format_selector,
"-o",
template,
row["url"],
]
run(command)
downloaded = find_cached_source(cache_dir, row["video_id"])
if downloaded is None:
raise RuntimeError(f"yt-dlp produced no media file for {row['video_id']}")
return downloaded
def reconstruct_one(row: dict[str, str], source: Path, output: Path, ffmpeg: str) -> None:
output.parent.mkdir(parents=True, exist_ok=True)
command = [
ffmpeg,
"-y",
"-nostdin",
"-loglevel",
"error",
"-i",
str(source),
# Output-side -ss is slower but frame accurate. Do not move it before -i.
"-ss",
row["start_time"],
"-t",
row["duration"],
"-map",
"0:v:0",
"-map",
"0:a:0?",
"-vf",
(
f"fps={row['fps']},"
f"scale={row['frame_width']}:{row['frame_height']}:flags=lanczos"
),
"-frames:v",
row["num_frames"],
"-fps_mode",
"cfr",
"-c:v",
"libx264",
"-preset",
"medium",
"-crf",
"18",
"-pix_fmt",
"yuv420p",
"-c:a",
"aac",
"-ar",
"48000",
"-ac",
"2",
"-movflags",
"+faststart",
str(output),
]
run(command)
def probe_video(path: Path, ffprobe: str) -> dict[str, object]:
command = [
ffprobe,
"-v",
"error",
"-select_streams",
"v:0",
"-count_frames",
"-show_entries",
"stream=width,height,avg_frame_rate,nb_read_frames",
"-of",
"json",
str(path),
]
output = subprocess.run(command, check=True, capture_output=True, text=True).stdout
return json.loads(output)["streams"][0]
def validate_encoded(row: dict[str, str], output: Path, ffprobe: str) -> None:
stream = probe_video(output, ffprobe)
numerator, denominator = map(int, stream["avg_frame_rate"].split("/"))
fps = numerator / denominator
expected = {
"width": int(row["frame_width"]),
"height": int(row["frame_height"]),
"frames": int(row["num_frames"]),
}
actual = {
"width": int(stream["width"]),
"height": int(stream["height"]),
"frames": int(stream["nb_read_frames"]),
}
if actual != expected or abs(fps - float(row["fps"])) > 1e-6:
raise RuntimeError(
f"encoded verification failed for {output}: expected={expected}, "
f"actual={actual}, fps={fps}"
)
def main() -> int:
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("manifest", type=Path)
parser.add_argument("--output-dir", type=Path, required=True)
parser.add_argument("--cache-dir", type=Path, required=True)
parser.add_argument("--clip-id", action="append", default=[])
parser.add_argument("--dataset", choices=("sekai", "sekai2"))
parser.add_argument("--limit", type=int)
parser.add_argument("--format", default=DEFAULT_FORMAT)
parser.add_argument("--yt-dlp", default="yt-dlp")
parser.add_argument("--ffmpeg", default="ffmpeg")
parser.add_argument("--ffprobe", default="ffprobe")
parser.add_argument(
"--remove-source",
action="store_true",
help="Delete each cached source after all selected clips are reconstructed.",
)
args = parser.parse_args()
for executable in (args.yt_dlp, args.ffmpeg, args.ffprobe):
if shutil.which(executable) is None:
raise RuntimeError(f"required executable not found: {executable}")
rows = read_rows(args.manifest, set(args.clip_id), args.dataset, args.limit)
if not rows:
raise RuntimeError("no perspective rows selected")
args.cache_dir.mkdir(parents=True, exist_ok=True)
grouped: dict[str, list[dict[str, str]]] = defaultdict(list)
for row in rows:
grouped[row["video_id"]].append(row)
for video_rows in grouped.values():
source = download_source(video_rows[0], args.cache_dir, args.yt_dlp, args.format)
for row in video_rows:
output = args.output_dir / row["dataset"] / row["clip_id"]
reconstruct_one(row, source, output, args.ffmpeg)
validate_encoded(row, output, args.ffprobe)
print(f"OK {row['clip_id']} -> {output}")
if args.remove_source:
source.unlink()
return 0
if __name__ == "__main__":
raise SystemExit(main())
|