ChanPivotMonitor: 实时中枢特征跟踪 + Telegram推送

- ChanPivotClassifier: 提取 calc_duration/contraction/shift 为 @staticmethod，新增 compute_features()
- ChanPivotMonitor: 实时追踪当前中枢，bi_count 增长时重新计算 shift/contraction/duration
- bsp_monitor/fetcher: 改用 data_provider HTTP API 替代直连 CCXT
- bsp_monitor/notify: 新增 send_telegram_message() 通用推送
- bsp_monitor/main: 集成 ChanPivotMonitor，有新笔或 BSP 时推送到 Telegram

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

ChanPivotClassifier.py

@@ -31,14 +31,16 @@ class ChanPivotClassifier:
     # Feature extraction
     # ------------------------------------------------------------------
 
-    def _calc_duration(self, zs) -> int:
+    @staticmethod
+    def calc_duration(zs) -> int:
         """持续时间: 第一笔首K → 最后一笔末K 的 index 差"""
         bi_list = zs.bi_list
         start_idx = bi_list[0].start_klc.index
         end_idx = bi_list[-1].end_klc.index
         return end_idx - start_idx
 
-    def _calc_contraction(self, zs) -> float:
+    @staticmethod
+    def calc_contraction(zs) -> float:
         """收敛率: 后窗口振幅均值 / 前窗口振幅均值"""
         bi_list = zs.bi_list
         if len(bi_list) < 4:
@@ -55,7 +57,8 @@ class ChanPivotClassifier:
             return 1.0
         return last_mean / first_mean
 
-    def _calc_shift(self, zs) -> tuple[float, float]:
+    @staticmethod
+    def calc_shift(zs) -> tuple[float, float]:
         """重心漂移: 前后半段重心均值差 (原始值, 归一化值)"""
         bi_list = zs.bi_list
         mid = len(bi_list) // 2
@@ -76,6 +79,39 @@ class ChanPivotClassifier:
 
         return shift_raw, shift_norm
 
+    @staticmethod
+    def compute_duration_norm(duration_raw: int, historical_durations: list) -> float:
+        """用历史窗口均值归一化 duration"""
+        if not historical_durations:
+            return 1.0
+        avg = sum(historical_durations) / len(historical_durations)
+        if avg == 0:
+            return 1.0
+        return duration_raw / avg
+
+    @staticmethod
+    def compute_features(zs, historical_durations: list | None = None):
+        """计算单个中枢的全部结构特征（实时友好）"""
+        duration_raw = ChanPivotClassifier.calc_duration(zs)
+        contraction = ChanPivotClassifier.calc_contraction(zs)
+        shift_raw, shift_norm = ChanPivotClassifier.calc_shift(zs)
+
+        if historical_durations is not None and len(historical_durations) > 0:
+            duration_norm = ChanPivotClassifier.compute_duration_norm(
+                duration_raw, historical_durations
+            )
+        else:
+            duration_norm = 1.0
+
+        return {
+            "duration_raw": duration_raw,
+            "duration_norm": round(duration_norm, 4),
+            "contraction": round(contraction, 4),
+            "shift_raw": round(shift_raw, 6),
+            "shift_norm": round(shift_norm, 4),
+            "zs_height": round(zs.zg - zs.zd, 6),
+        }
+
     # ------------------------------------------------------------------
     # Label computation
     # ------------------------------------------------------------------
@@ -189,9 +225,9 @@ class ChanPivotClassifier:
             if not zs.is_sure or len(zs.bi_list) < 3:
                 continue
 
-            duration_raw = self._calc_duration(zs)
-            contraction = self._calc_contraction(zs)
-            shift_raw, shift_norm = self._calc_shift(zs)
+            duration_raw = ChanPivotClassifier.calc_duration(zs)
+            contraction = ChanPivotClassifier.calc_contraction(zs)
+            shift_raw, shift_norm = ChanPivotClassifier.calc_shift(zs)
 
             raw.append({
                 "zs": zs,
@@ -203,17 +239,14 @@ class ChanPivotClassifier:
                 "zs_height": zs.zg - zs.zd,
             })
 
-        # 归一化 duration: 除以均值
-        if raw:
-            avg_duration = sum(r["duration_raw"] for r in raw) / len(raw)
-        else:
-            avg_duration = 1
-
         # 第二遍：组装输出 + 计算 label
         result = []
         for r in raw:
             zs = r["zs"]
-            duration_norm = r["duration_raw"] / avg_duration if avg_duration > 0 else 1.0
+            historical = [x["duration_raw"] for x in raw]
+            duration_norm = ChanPivotClassifier.compute_duration_norm(
+                r["duration_raw"], historical
+            )
             label_info = self._compute_label(zs, r["contraction"], r["shift_norm"])
 
             # 时间处理