Spaces:
Sleeping
Sleeping
| import os | |
| import glob | |
| import sys | |
| print("=" * 60) | |
| print("PHASE 2: DATASET SANITY CHECK") | |
| print("=" * 60) | |
| dataset_root = "dataset" | |
| if not os.path.exists(dataset_root): | |
| print(f"❌ FAIL: Dataset directory '{dataset_root}' not found") | |
| sys.exit(1) | |
| print(f"✅ Dataset root exists: {dataset_root}") | |
| # Expected structure | |
| classes = ["human", "ai_generated"] | |
| languages = ["tamil", "english", "hindi", "malayalam", "telugu"] | |
| total_files = 0 | |
| class_counts = {} | |
| issues = [] | |
| for cls in classes: | |
| cls_path = os.path.join(dataset_root, cls) | |
| if not os.path.exists(cls_path): | |
| issues.append(f"❌ Missing class directory: {cls}") | |
| continue | |
| print(f"\n📁 Class: {cls}") | |
| class_total = 0 | |
| for lang in languages: | |
| lang_path = os.path.join(cls_path, lang) | |
| if not os.path.exists(lang_path): | |
| issues.append(f"❌ Missing language directory: {cls}/{lang}") | |
| print(f" {lang}: ❌ MISSING") | |
| continue | |
| # Count audio files | |
| audio_files = [] | |
| for ext in ['*.mp3', '*.wav', '*.MP3', '*.WAV']: | |
| audio_files.extend(glob.glob(os.path.join(lang_path, ext))) | |
| file_count = len(audio_files) | |
| class_total += file_count | |
| if file_count == 0: | |
| issues.append(f"⚠️ Empty directory: {cls}/{lang}") | |
| print(f" {lang}: ⚠️ EMPTY (0 files)") | |
| else: | |
| print(f" {lang}: ✅ {file_count} files") | |
| class_counts[cls] = class_total | |
| total_files += class_total | |
| print(f" Total for {cls}: {class_total} files") | |
| print("\n" + "=" * 60) | |
| print("DATASET SUMMARY") | |
| print("=" * 60) | |
| print(f"Human samples: {class_counts.get('human', 0)}") | |
| print(f"AI-generated samples: {class_counts.get('ai_generated', 0)}") | |
| print(f"Total samples: {total_files}") | |
| # Calculate distribution | |
| if total_files > 0: | |
| human_pct = (class_counts.get('human', 0) / total_files) * 100 | |
| ai_pct = (class_counts.get('ai_generated', 0) / total_files) * 100 | |
| print(f"\nDistribution:") | |
| print(f" Human: {human_pct:.1f}%") | |
| print(f" AI-generated: {ai_pct:.1f}%") | |
| # Check for critical issues | |
| critical_fail = False | |
| if class_counts.get('human', 0) == 0: | |
| print(f"\n❌ CRITICAL: No human samples found!") | |
| critical_fail = True | |
| if class_counts.get('ai_generated', 0) == 0: | |
| print(f"\n❌ CRITICAL: No AI-generated samples found!") | |
| critical_fail = True | |
| if total_files < 10: | |
| print(f"\n❌ CRITICAL: Insufficient samples ({total_files} < 10)") | |
| critical_fail = True | |
| # Display issues | |
| if issues: | |
| print(f"\n⚠️ Issues found ({len(issues)}):") | |
| for issue in issues: | |
| print(f" {issue}") | |
| if critical_fail: | |
| print("\n❌ PHASE 2 FAILED: Critical dataset issues detected") | |
| sys.exit(1) | |
| else: | |
| print("\n✅ PHASE 2 PASSED: Dataset structure is valid") | |