Datasets:
Tasks:
Audio Classification
Formats:
parquet
Size:
1K - 10K
ArXiv:
Tags:
arxiv:2606.01686
music
ai-generated-music
ai-generated-music-detection
plagiarism-detection
ace-step
License:
File size: 15,610 Bytes
b347b70 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 | #!/usr/bin/env python3
"""
Common utilities for AI Music Dataset collection pipeline.
- Metadata management (JSON per-track + summary CSV)
- Disk space monitoring
- Download helpers with retry/rate-limit
- Audio validation
"""
import json
import csv
import os
import shutil
import time
import hashlib
import logging
import subprocess
from pathlib import Path
from datetime import datetime, timezone
from dataclasses import dataclass, field, asdict
from typing import Optional, List, Dict, Any
from functools import wraps
import requests
from tqdm import tqdm
# ─── Configuration ───────────────────────────────────────────
BASE_DIR = Path("/ssd_data/dataset/ai_music_dataset")
FAKE_DIR = BASE_DIR / "fake"
METADATA_DIR = BASE_DIR / "metadata"
SCRIPTS_DIR = BASE_DIR / "scripts"
TARGET_PER_FOLDER = 2000
MIN_FREE_SPACE_GB = 50 # Stop collection if disk < 50GB free
MAX_RETRIES = 3
RETRY_DELAY = 2 # seconds
logging.basicConfig(
level=logging.INFO,
format="%(asctime)s [%(levelname)s] %(name)s: %(message)s",
handlers=[
logging.StreamHandler(),
logging.FileHandler(BASE_DIR / "collection.log", mode="a"),
],
)
logger = logging.getLogger("ai_music_dataset")
# ─── Metadata Schema ────────────────────────────────────────
@dataclass
class TrackMetadata:
"""Per-track metadata schema."""
# Core identification
track_id: str # Unique ID (UUID or platform ID)
filename: str # Local filename
category: str # A_commercial, A_opensource, B_hybrid, C_mixing
subcategory: str # e.g., suno_v4, musicgen_large, B1_ai_cover_of_human
# Source information
source_platform: str # suno, udio, mureka, musicgen, acestep, etc.
source_type: str # "commercial" or "open-source"
model_name: str # Full model name
model_version: str # v3, v3.5, v4, Small, Large, etc.
# Audio properties
duration_sec: Optional[float] = None
sample_rate: Optional[int] = None
channels: Optional[int] = None
bitrate_kbps: Optional[int] = None
file_size_bytes: Optional[int] = None
audio_format: str = "mp3"
# Content metadata
title: Optional[str] = None
prompt: Optional[str] = None # Generation prompt (if available)
genre: Optional[str] = None
tags: List[str] = field(default_factory=list)
lyrics: Optional[str] = None
# Source URLs
source_url: Optional[str] = None
download_url: Optional[str] = None
page_url: Optional[str] = None
# For hybrid (B-category) tracks
original_source: Optional[str] = None # Original human/AI track used
processing_method: Optional[str] = None # How it was processed
ai_component: Optional[str] = None # Which part is AI
human_component: Optional[str] = None # Which part is human
# Collection metadata
collected_at: str = ""
collection_method: str = "" # "crawl", "api", "generate", "process"
md5_hash: Optional[str] = None
# Quality flags
is_valid: bool = True
validation_notes: Optional[str] = None
def __post_init__(self):
if not self.collected_at:
self.collected_at = datetime.now(timezone.utc).isoformat()
class MetadataManager:
"""Manages per-folder metadata as JSONL + summary CSV."""
def __init__(self, folder_path: Path):
self.folder_path = Path(folder_path)
self.jsonl_path = self.folder_path / "metadata.jsonl"
self.summary_path = self.folder_path / "summary.json"
self._existing_ids = set()
self._load_existing()
def _load_existing(self):
"""Load existing track IDs to avoid duplicates."""
if self.jsonl_path.exists():
with open(self.jsonl_path, "r", encoding="utf-8") as f:
for line in f:
try:
data = json.loads(line.strip())
self._existing_ids.add(data.get("track_id", ""))
except json.JSONDecodeError:
continue
def has_track(self, track_id: str) -> bool:
return track_id in self._existing_ids
def add_track(self, meta: TrackMetadata):
"""Append track metadata to JSONL file."""
if meta.track_id in self._existing_ids:
return # Skip duplicate
with open(self.jsonl_path, "a", encoding="utf-8") as f:
f.write(json.dumps(asdict(meta), ensure_ascii=False) + "\n")
self._existing_ids.add(meta.track_id)
def get_count(self) -> int:
return len(self._existing_ids)
def update_summary(self):
"""Write summary.json with aggregate stats."""
tracks = []
total_size = 0
total_duration = 0.0
if self.jsonl_path.exists():
with open(self.jsonl_path, "r", encoding="utf-8") as f:
for line in f:
try:
data = json.loads(line.strip())
tracks.append(data)
total_size += data.get("file_size_bytes", 0) or 0
total_duration += data.get("duration_sec", 0) or 0
except json.JSONDecodeError:
continue
summary = {
"folder": str(self.folder_path),
"total_tracks": len(tracks),
"target_tracks": TARGET_PER_FOLDER,
"total_size_mb": round(total_size / (1024 * 1024), 2),
"total_duration_hours": round(total_duration / 3600, 2),
"avg_duration_sec": round(total_duration / max(len(tracks), 1), 1),
"collection_progress": f"{len(tracks)}/{TARGET_PER_FOLDER}",
"last_updated": datetime.now(timezone.utc).isoformat(),
"versions": {},
}
# Count by version
for t in tracks:
ver = t.get("model_version", "unknown")
summary["versions"][ver] = summary["versions"].get(ver, 0) + 1
with open(self.summary_path, "w", encoding="utf-8") as f:
json.dump(summary, f, ensure_ascii=False, indent=2)
return summary
# ─── Disk Space Monitor ─────────────────────────────────────
def check_disk_space(path: str = "/ssd_data") -> Dict[str, float]:
"""Check available disk space in GB."""
usage = shutil.disk_usage(path)
return {
"total_gb": round(usage.total / (1024**3), 1),
"used_gb": round(usage.used / (1024**3), 1),
"free_gb": round(usage.free / (1024**3), 1),
"usage_percent": round(usage.used / usage.total * 100, 1),
}
def ensure_disk_space(min_free_gb: float = MIN_FREE_SPACE_GB) -> bool:
"""Return True if enough disk space, False otherwise."""
space = check_disk_space()
if space["free_gb"] < min_free_gb:
logger.warning(
f"Low disk space! {space['free_gb']:.1f}GB free "
f"(minimum: {min_free_gb}GB). Stopping collection."
)
return False
return True
# ─── Download Helpers ────────────────────────────────────────
def download_file(
url: str,
output_path: Path,
headers: Optional[Dict] = None,
timeout: int = 60,
max_retries: int = MAX_RETRIES,
) -> bool:
"""Download a file with retry logic."""
output_path = Path(output_path)
if output_path.exists() and output_path.stat().st_size > 0:
return True # Already downloaded
for attempt in range(max_retries):
try:
resp = requests.get(url, headers=headers, timeout=timeout, stream=True)
resp.raise_for_status()
output_path.parent.mkdir(parents=True, exist_ok=True)
with open(output_path, "wb") as f:
for chunk in resp.iter_content(chunk_size=8192):
f.write(chunk)
# Verify non-empty
if output_path.stat().st_size > 1000: # At least 1KB
return True
else:
output_path.unlink(missing_ok=True)
logger.warning(f"Downloaded file too small: {output_path}")
except requests.RequestException as e:
logger.warning(f"Download attempt {attempt + 1}/{max_retries} failed: {e}")
if attempt < max_retries - 1:
time.sleep(RETRY_DELAY * (attempt + 1))
return False
def compute_md5(filepath: Path) -> str:
"""Compute MD5 hash of a file."""
md5 = hashlib.md5()
with open(filepath, "rb") as f:
for chunk in iter(lambda: f.read(8192), b""):
md5.update(chunk)
return md5.hexdigest()
# ─── Audio Validation ───────────────────────────────────────
def get_audio_info(filepath: Path) -> Dict[str, Any]:
"""Get audio metadata using ffprobe."""
try:
cmd = [
"ffprobe",
"-v",
"quiet",
"-print_format",
"json",
"-show_format",
"-show_streams",
str(filepath),
]
result = subprocess.run(cmd, capture_output=True, text=True, timeout=30)
if result.returncode != 0:
return {}
info = json.loads(result.stdout)
fmt = info.get("format", {})
streams = info.get("streams", [{}])
audio_stream = next((s for s in streams if s.get("codec_type") == "audio"), {})
return {
"duration_sec": float(fmt.get("duration", 0)),
"sample_rate": int(audio_stream.get("sample_rate", 0)),
"channels": int(audio_stream.get("channels", 0)),
"bitrate_kbps": int(fmt.get("bit_rate", 0)) // 1000,
"file_size_bytes": int(fmt.get("size", 0)),
"audio_format": fmt.get("format_name", "unknown"),
}
except Exception as e:
logger.error(f"ffprobe failed for {filepath}: {e}")
return {}
def validate_audio(filepath: Path, min_duration: float = 5.0) -> bool:
"""Validate that an audio file is playable and meets minimum duration."""
info = get_audio_info(filepath)
if not info:
return False
duration = info.get("duration_sec", 0)
return duration >= min_duration
# ─── Rate Limiter ────────────────────────────────────────────
class RateLimiter:
"""Simple rate limiter for API/crawl requests."""
def __init__(self, requests_per_second: float = 2.0):
self.min_interval = 1.0 / requests_per_second
self.last_request = 0.0
def wait(self):
elapsed = time.time() - self.last_request
if elapsed < self.min_interval:
time.sleep(self.min_interval - elapsed)
self.last_request = time.time()
# ─── Progress Reporter ──────────────────────────────────────
def print_collection_status():
"""Print overall collection status across all folders."""
print("\n" + "=" * 80)
print("AI Music Dataset Collection Status")
print("=" * 80)
space = check_disk_space()
print(
f"Disk: {space['used_gb']:.1f}GB used / {space['total_gb']:.1f}GB total "
f"({space['free_gb']:.1f}GB free, {space['usage_percent']:.1f}%)"
)
print()
categories = ["A_commercial", "A_opensource", "B_hybrid", "C_mixing"]
total_tracks = 0
total_size = 0
for cat in categories:
cat_dir = FAKE_DIR / cat
if not cat_dir.exists():
continue
print(f" [{cat}]")
for sub in sorted(cat_dir.iterdir()):
if sub.is_dir():
meta_mgr = MetadataManager(sub)
count = meta_mgr.get_count()
# Also count actual audio files as fallback
audio_files = (
list(sub.glob("*.mp3"))
+ list(sub.glob("*.wav"))
+ list(sub.glob("*.flac"))
)
file_count = len(audio_files)
actual_count = max(count, file_count)
total_tracks += actual_count
folder_size = sum(
f.stat().st_size for f in sub.rglob("*") if f.is_file()
)
total_size += folder_size
bar = "█" * (actual_count * 30 // TARGET_PER_FOLDER) + "░" * (
30 - actual_count * 30 // TARGET_PER_FOLDER
)
print(
f" {sub.name:<25} [{bar}] {actual_count:>5}/{TARGET_PER_FOLDER} "
f"({folder_size / 1024 / 1024:.0f}MB)"
)
print()
print(f" Total: {total_tracks} tracks, {total_size / 1024 / 1024 / 1024:.2f}GB")
print("=" * 80)
# ─── Prompt Generators ──────────────────────────────────────
MUSIC_PROMPTS = [
# Genre variety for open-source generation
"upbeat pop song with catchy melody and electronic beats",
"ambient electronic music with atmospheric pads and gentle rhythms",
"acoustic folk song with fingerpicked guitar and warm vocals",
"hip hop beat with heavy bass and trap hi-hats",
"classical piano sonata in the style of Chopin",
"jazz fusion with complex chord progressions and saxophone",
"heavy metal with distorted guitars and double bass drums",
"lo-fi hip hop chill beats for studying",
"reggae song with offbeat guitar and bass groove",
"country music with acoustic guitar and slide steel",
"EDM festival anthem with big drops and synth leads",
"R&B smooth vocals over neo-soul chords",
"indie rock with jangly guitars and reverb",
"orchestral film score with dramatic strings and brass",
"bossa nova with nylon guitar and soft percussion",
"synthwave retro 80s electronic music",
"punk rock fast tempo with power chords",
"blues guitar solo with emotional bends",
"K-pop song with catchy chorus and dance beat",
"Latin reggaeton with dembow rhythm",
"psychedelic rock with wah guitar and space effects",
"minimal techno with hypnotic loops",
"gospel choir with powerful harmonies",
"Celtic folk music with fiddle and tin whistle",
"Afrobeat with polyrhythmic percussion and horn section",
"drum and bass with fast breakbeats",
"soul music with Motown-style arrangement",
"grunge alternative rock with distortion",
"world music fusion with tabla and sitar",
"chiptune 8-bit video game music",
"tropical house with marimba and steel drums",
"progressive rock epic with time signature changes",
"acoustic singer-songwriter ballad",
"dubstep with heavy wobble bass",
"flamenco guitar with passionate rhythm",
"new age meditation music with nature sounds",
"big band swing jazz",
"post-rock with atmospheric guitar layers",
"dancehall with Caribbean rhythm",
"shoegaze with wall of sound guitars",
]
def get_diverse_prompts(n: int) -> List[str]:
"""Get n diverse music prompts, cycling through the list."""
prompts = []
for i in range(n):
prompts.append(MUSIC_PROMPTS[i % len(MUSIC_PROMPTS)])
return prompts
if __name__ == "__main__":
print_collection_status()
|