cleanup: remove dead code and duplicate docs

- Remove session-ses_2f27.md (161KB raw session log)
- Remove 49 ROOT_* duplicate files across REFERENCE/
- Remove 14 duplicate files between REFERENCE/ root and history/
- Remove asr_legacy.rs (dead code, replaced by asr.rs)
- Remove src/core/worker/ (duplicate JobWorker)
- Remove src/core/layers/ (empty directory)
- Remove 4 .bak files in src/
- Remove 7 dead private methods in worker/processor.rs
- Remove backup directory from git tracking

scripts/music_segmentation_processor.py

@@ -8,7 +8,6 @@ import librosa
 import numpy as np
 import os
 import json
-import matplotlib.pyplot as plt  # Only for debug if needed, but we stick to console for now
 
 # 設定
 UUID = os.getenv("UUID", "384b0ff44aaaa1f1")
@@ -29,7 +28,7 @@ def analyze_music_segmentation(audio_path):
     hop_length = int(1.0 * sr)
     chroma = librosa.feature.chroma_stft(y=y, sr=sr, hop_length=hop_length)
 
-    print(f"📊 Analyzing transitions...")
+    print("📊 Analyzing transitions...")
 
     # 2. 計算自我相似度矩陣 (Self-Similarity Matrix) - 優化版
     # 這裡我們簡化為計算相鄰片段的餘弦距離 (Cosine Distance)
@@ -45,7 +44,6 @@ def analyze_music_segmentation(audio_path):
 
     # 使用 librosa 的 onset_strength 的變體，但針對 Chroma
     # 這裡手動計算 Cosine Distance 以確保準確度
-    from sklearn.metrics.pairwise import cosine_similarity
 
     # 為了效能，我們不逐一計算，而是使用向量化的方法
     # 計算 frame[t] 和 frame[t+lag] 的差異
@@ -127,12 +125,12 @@ if __name__ == "__main__":
     with open(OUTPUT_JSON, "w", encoding="utf-8") as f:
         json.dump({"music_segments": segments}, f, indent=2, ensure_ascii=False)
 
-    print(f"\n🎉 Analysis Complete!")
+    print("\n🎉 Analysis Complete!")
     print(f"✅ Identified {len(segments)} music-based scenes.")
     print(f"💾 Saved to {OUTPUT_JSON}")
 
     # 顯示結果
-    print(f"\n🎶 Top Music Segments:")
+    print("\n🎶 Top Music Segments:")
     for i, seg in enumerate(segments[:20]):
         m_s, s_s = divmod(seg["start_time"], 60)
         print(f"  {i + 1:02d}. [{int(m_s):02d}:{s_s:05.2f}] - {seg['duration']}s")