File size: 4,850 Bytes
155a87f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
#!/usr/bin/env python3
"""Scrape a diverse multi-channel cover-video corpus for StegoReels.

Reads a channel list (--channels-file, see dataset/config/channels.csv),
scrapes each channel's Shorts tab up to a per-channel cap, and tags every
clip's .meta.json with that channel's category/split. Wraps yt_scrapper.py
for the actual download/filter logic. Verify channel handles before a large
run -- creator @handles change and are not guaranteed current.

Usage:
    python scrape_dataset.py --target-total 1000
"""

from __future__ import annotations

import argparse
import csv
import json
import logging
import math
from pathlib import Path

from yt_scrapper import DEFAULT_MAX_DURATION_S, DEFAULT_MIN_DURATION_S, ScrapeConfig, run

LOGGER = logging.getLogger("scrape_dataset")

SCRIPT_DIR = Path(__file__).resolve().parent
DATASET_DIR = SCRIPT_DIR.parent


def load_channels(path: Path) -> list[dict]:
    with path.open(encoding="utf-8") as f:
        rows = [row for row in csv.DictReader(f) if row.get("channel_url", "").strip()]
    if not rows:
        raise SystemExit(f"No channels found in {path}")
    return rows


def build_manifest(output_dir: Path, manifest_path: Path) -> int:
    """Aggregate every *.meta.json under output_dir into one manifest.jsonl."""
    count = 0
    with manifest_path.open("w", encoding="utf-8") as out:
        for meta_file in sorted(output_dir.glob("*.meta.json")):
            meta = json.loads(meta_file.read_text(encoding="utf-8"))
            out.write(json.dumps(meta, ensure_ascii=False) + "\n")
            count += 1
    return count


def parse_args(argv: list[str] | None = None) -> argparse.Namespace:
    parser = argparse.ArgumentParser(description=__doc__)
    parser.add_argument(
        "--channels-file", type=Path, default=DATASET_DIR / "config" / "channels.csv"
    )
    parser.add_argument("--output-dir", type=Path, default=DATASET_DIR / "raw")
    parser.add_argument("--archive-file", type=Path, default=None)
    parser.add_argument("--target-total", type=int, default=1000)
    parser.add_argument(
        "--per-channel-max", type=int, default=None, help="Overrides target-total / #channels."
    )
    parser.add_argument(
        "--val-test-fraction",
        type=float,
        default=0.35,
        help="Per-channel clip cap for val/test channels, as a fraction of the train cap "
        "(val/test only need enough clips to evaluate on, not train-level volume).",
    )
    parser.add_argument("--min-duration", type=int, default=DEFAULT_MIN_DURATION_S)
    parser.add_argument("--max-duration", type=int, default=DEFAULT_MAX_DURATION_S)
    parser.add_argument("--no-require-vertical", dest="require_vertical", action="store_false")
    parser.add_argument("--sleep-interval", type=float, default=1.0)
    parser.add_argument("--max-sleep-interval", type=float, default=3.0)
    parser.add_argument("--cookies-from-browser", default=None)
    parser.add_argument("--dry-run", action="store_true")
    parser.add_argument("-v", "--verbose", action="store_true")
    parser.set_defaults(require_vertical=True)
    return parser.parse_args(argv)


def main(argv: list[str] | None = None) -> None:
    args = parse_args(argv)
    logging.basicConfig(
        level=logging.DEBUG if args.verbose else logging.INFO,
        format="%(asctime)s [%(levelname)s] %(message)s",
    )

    channels = load_channels(args.channels_file)
    per_channel_max = args.per_channel_max or math.ceil(args.target_total / len(channels))
    archive_file = args.archive_file or (args.output_dir / ".yt_dlp_archive.txt")

    LOGGER.info(
        "%d channels, cap %d clips/channel (target ~%d total)",
        len(channels),
        per_channel_max,
        args.target_total,
    )

    for row in channels:
        cap = (
            per_channel_max
            if row["split"] == "train"
            else max(1, round(per_channel_max * args.val_test_fraction))
        )
        cfg = ScrapeConfig(
            output_dir=args.output_dir,
            archive_file=archive_file,
            max_videos=cap,
            min_duration=args.min_duration,
            max_duration=args.max_duration,
            require_vertical=args.require_vertical,
            sleep_interval=args.sleep_interval,
            max_sleep_interval=args.max_sleep_interval,
            cookies_from_browser=args.cookies_from_browser,
            verbose=args.verbose,
            dry_run=args.dry_run,
        )
        extra_meta = {"category": row["category"], "split": row["split"]}
        run(cfg, [row["channel_url"]], extra_meta=extra_meta)

    if not args.dry_run:
        manifest_path = args.output_dir / "manifest.jsonl"
        n = build_manifest(args.output_dir, manifest_path)
        LOGGER.info("Wrote manifest with %d clips -> %s", n, manifest_path)


if __name__ == "__main__":
    main()