添加 StructureZone 结构价值区系统，支持多周期支撑/阻力分析

- 新增 ChanZone.py: 从笔中枢/线段中枢/EMA52 提取价格区，聚类评分
- ChanLun.py 新增 get_structure_zones() 方法
- web/app.py: 独立拉取多周期数据 + 缓存 + limit 传参避免全量传输
- web/index.html: 结构区勾选框 + K线数量输入 + 半透明填充区绘制
- tests/test_chan_zone.py: 24 个单元测试

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

CLAUDE.md

@@ -12,36 +12,38 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co
 
 Data processing pipeline (each step feeds the next):
 
-1. **`ChanKLU.py`** — Raw K-line unit with TA indicators (EMA, MACD, RSI, Bollinger Bands)
-2. **`ChanKLC.py`** — Combined K-line: inclusion processing (包含处理), fractal (分型) detection
+1. **`ChanKLU.py`** — Raw K-line unit with TA indicators (EMA, MACD, RSI, Bollinger Bands) and candlestick pattern recognition (`Chan_KLU_PATTERN`)
+2. **`ChanKLC.py`** — Combined K-line: inclusion processing (包含处理), fractal (分型) detection. Linked-list structure with `.next`/`.pre` pointers
 3. **`ChanBI.py`** — Stroke (笔): basic trend unit connecting alternating fractals
-4. **`ChanSEG.py`** — Segment (线段): built from strokes
-5. **`ChanZS.py`** / **`ChanBIZS.py`** — Center/pivot (中枢): consolidation zones (stroke-level and segment-level)
-6. **`ChanBSP.py`** — Buy/Sell points (买卖点): Type 1/2/3 signals
-7. **`ChanLun.py`** — Main orchestrator: ties all steps together, entry point
-8. **`TF_DF.py`** — Timeframe-aware DataFrame processor: resamples data, runs the full pipeline per timeframe, handles multi-timeframe analysis
+4. **`ChanSBI.py`** — Special Stroke: aggregates multiple BI into higher-level units with fractal detection, feeds into SEG
+5. **`ChanSEG.py`** — Segment (线段): built from SBI strokes
+6. **`ChanZS.py`** / **`ChanBIZS.py`** — Center/pivot (中枢): consolidation zones (segment-level and stroke-level)
+7. **`ChanBSP.py`** — Buy/Sell points (买卖点): Type 1/2/3 signals
+8. **`ChanLun.py`** — Main orchestrator: ties all steps together, entry point
+9. **`TF_DF.py`** — Timeframe-aware DataFrame processor: resamples data, runs the full pipeline per timeframe, handles multi-timeframe analysis
 
 ### Support modules
 
 - **`ChanEnum.py`** — All enumerations: K-line types, fractal types, MACD states, buy/sell point types, EMA position/semantic states, K-line patterns
+- **`ChanCTime.py`** — Chan theory time utility: auto-adaptive day understanding (e.g. crypto 24h vs stock market hours)
 - **`ChanMACD.py`** / **`ChanMACDHistSet.py`** / **`ChanMACDSeg.py`** / **`ChanMACDUnitTF.py`** — MACD state analysis and divergence detection
-- **`ChanKLU.py`** — K-line unit with candlestick pattern recognition (`Chan_KLU_PATTERN`)
 - **`ChanPY.py`** — Consolidation (盘整) analysis
 - **`ChanHeng.py`** — Sideways market analysis
 - **`Chan_FX_Box.py`** — Fractal box (分型箱体) detection
+- **`ChanLun_Classifier.py`** — Standalone classifier script: runs full pipeline and classifies market states
 
 ### Services
 
 - **`data_provider/`** — FastAPI data service: fetches crypto data from Binance via CCXT, caches to CSV, serves REST API + WebSocket. Synthesizes derived timeframes (e.g. 5m/15m/4h from 1m/1h base). Port 9009.
 - **`web/`** — Flask web UI for interactive chart visualization with Chan theory overlays. Port 8123.
-- **`strategies/`** — Freqtrade trading strategies using the Chan theory engine (40+ strategies)
+- **`strategies/`** — Freqtrade trading strategies using the Chan theory engine (53 strategies)
 - **`config/`** — Freqtrade JSON config files per pair/timeframe
 
 ### Data Flow
 
 ```
 Exchange (CCXT) → data_provider (CSV cache) → Freqtrade → Strategy → ChanLun → TF_DF
-                                                                              → KLU → KLC → BI → SEG → ZS → BSP
+                                                                              → KLU → KLC → BI → SBI → SEG → ZS → BSP
 ```
 
 ## Common Commands