Update ml_engine/strategies.py
Browse files- ml_engine/strategies.py +65 -166
ml_engine/strategies.py
CHANGED
|
@@ -1,17 +1,19 @@
|
|
| 1 |
-
# ml_engine/strategies.py (
|
| 2 |
import asyncio
|
| 3 |
|
| 4 |
-
#
|
| 5 |
from .patterns import ChartPatternAnalyzer
|
| 6 |
|
| 7 |
class PatternEnhancedStrategyEngine:
|
| 8 |
-
|
|
|
|
| 9 |
self.data_manager = data_manager
|
| 10 |
-
self.
|
| 11 |
self.pattern_analyzer = ChartPatternAnalyzer()
|
|
|
|
| 12 |
|
| 13 |
async def enhance_strategy_with_patterns(self, strategy_scores, pattern_analysis, symbol):
|
| 14 |
-
"""
|
| 15 |
if not pattern_analysis or pattern_analysis.get('pattern_detected') in ['no_clear_pattern', 'insufficient_data']:
|
| 16 |
return strategy_scores
|
| 17 |
|
|
@@ -23,18 +25,17 @@ class PatternEnhancedStrategyEngine:
|
|
| 23 |
enhancement_factor = self._calculate_pattern_enhancement(pattern_confidence, pattern_name)
|
| 24 |
enhanced_strategies = self._get_pattern_appropriate_strategies(pattern_name, predicted_direction)
|
| 25 |
|
| 26 |
-
|
| 27 |
|
| 28 |
for strategy in enhanced_strategies:
|
| 29 |
if strategy in strategy_scores:
|
| 30 |
original_score = strategy_scores[strategy]
|
| 31 |
strategy_scores[strategy] = min(original_score * enhancement_factor, 1.0)
|
| 32 |
-
print(f" 📈 {strategy}: {original_score:.3f} → {strategy_scores[strategy]:.3f}")
|
| 33 |
|
| 34 |
return strategy_scores
|
| 35 |
|
| 36 |
def _calculate_pattern_enhancement(self, pattern_confidence, pattern_name):
|
| 37 |
-
"""
|
| 38 |
base_enhancement = 1.0 + (pattern_confidence * 0.3)
|
| 39 |
high_reliability_patterns = ['Double Top', 'Double Bottom', 'Head & Shoulders', 'Cup and Handle']
|
| 40 |
if pattern_name in high_reliability_patterns:
|
|
@@ -42,26 +43,29 @@ class PatternEnhancedStrategyEngine:
|
|
| 42 |
return min(base_enhancement, 1.5)
|
| 43 |
|
| 44 |
def _get_pattern_appropriate_strategies(self, pattern_name, direction):
|
| 45 |
-
"""
|
| 46 |
reversal_patterns = ['Double Top', 'Double Bottom', 'Head & Shoulders', 'Triple Top', 'Triple Bottom']
|
| 47 |
continuation_patterns = ['Flags', 'Pennants', 'Triangles', 'Rectangles']
|
| 48 |
|
| 49 |
if pattern_name in reversal_patterns:
|
| 50 |
if direction == 'down':
|
| 51 |
-
return ['breakout_momentum', 'trend_following']
|
| 52 |
else:
|
| 53 |
-
return ['mean_reversion', 'breakout_momentum']
|
| 54 |
elif pattern_name in continuation_patterns:
|
| 55 |
return ['trend_following', 'breakout_momentum']
|
| 56 |
else:
|
| 57 |
return ['breakout_momentum', 'hybrid_ai']
|
| 58 |
|
| 59 |
class MultiStrategyEngine:
|
| 60 |
-
|
|
|
|
| 61 |
self.data_manager = data_manager
|
| 62 |
-
self.
|
| 63 |
|
| 64 |
-
|
|
|
|
|
|
|
| 65 |
|
| 66 |
self.strategies = {
|
| 67 |
'trend_following': self._trend_following_strategy,
|
|
@@ -70,25 +74,29 @@ class MultiStrategyEngine:
|
|
| 70 |
'volume_spike': self._volume_spike_strategy,
|
| 71 |
'whale_tracking': self._whale_tracking_strategy,
|
| 72 |
'pattern_recognition': self._pattern_recognition_strategy,
|
| 73 |
-
'hybrid_ai': self._hybrid_ai_strategy
|
| 74 |
}
|
| 75 |
|
| 76 |
async def evaluate_all_strategies(self, symbol_data, market_context):
|
| 77 |
-
"""
|
| 78 |
try:
|
| 79 |
-
|
|
|
|
|
|
|
| 80 |
try:
|
| 81 |
market_condition = market_context.get('market_trend', 'sideways_market')
|
| 82 |
-
|
|
|
|
| 83 |
except Exception as e:
|
|
|
|
| 84 |
optimized_weights = await self.get_default_weights()
|
| 85 |
else:
|
| 86 |
optimized_weights = await self.get_default_weights()
|
|
|
|
| 87 |
|
| 88 |
strategy_scores = {}
|
| 89 |
base_scores = {}
|
| 90 |
|
| 91 |
-
# 🔴 تقييم الاستراتيجيات الأساسية أولاً
|
| 92 |
primary_strategies = [s for s in self.strategies.keys() if s != 'hybrid_ai']
|
| 93 |
|
| 94 |
for strategy_name in primary_strategies:
|
|
@@ -102,10 +110,9 @@ class MultiStrategyEngine:
|
|
| 102 |
weighted_score = base_score * weight
|
| 103 |
strategy_scores[strategy_name] = min(weighted_score, 1.0)
|
| 104 |
except Exception as error:
|
| 105 |
-
print(f"❌
|
| 106 |
continue
|
| 107 |
|
| 108 |
-
# 🔴 تقييم استراتيجية hybrid_ai (الميتا) بعد حساب الدرجات الأساسية
|
| 109 |
try:
|
| 110 |
hybrid_score = await self._hybrid_ai_strategy(symbol_data, market_context, base_scores)
|
| 111 |
if hybrid_score is not None:
|
|
@@ -113,9 +120,9 @@ class MultiStrategyEngine:
|
|
| 113 |
weight = optimized_weights.get('hybrid_ai', 0.1)
|
| 114 |
strategy_scores['hybrid_ai'] = min(hybrid_score * weight, 1.0)
|
| 115 |
except Exception as e:
|
| 116 |
-
print(f"❌
|
| 117 |
|
| 118 |
-
#
|
| 119 |
pattern_analysis = symbol_data.get('pattern_analysis')
|
| 120 |
if pattern_analysis:
|
| 121 |
strategy_scores = await self.pattern_enhancer.enhance_strategy_with_patterns(
|
|
@@ -132,270 +139,162 @@ class MultiStrategyEngine:
|
|
| 132 |
return strategy_scores, base_scores
|
| 133 |
|
| 134 |
except Exception as error:
|
| 135 |
-
print(f"❌
|
| 136 |
return {}, {}
|
| 137 |
|
| 138 |
async def get_default_weights(self):
|
| 139 |
-
"""
|
| 140 |
return {
|
| 141 |
'trend_following': 0.15,
|
| 142 |
'mean_reversion': 0.12,
|
| 143 |
-
'breakout_momentum': 0.20,
|
| 144 |
-
'volume_spike': 0.13,
|
| 145 |
'whale_tracking': 0.20,
|
| 146 |
-
'pattern_recognition': 0.10,
|
| 147 |
-
'hybrid_ai': 0.10
|
| 148 |
}
|
| 149 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 150 |
async def _trend_following_strategy(self, symbol_data, market_context):
|
| 151 |
-
"""
|
| 152 |
-
🔴 (محسنة) استراتيجية تتبع الاتجاه - تركز على الأطر السريعة وقواعد ADX
|
| 153 |
-
القاعدة: EMA(21) > EMA(50) و ADX > 20
|
| 154 |
-
"""
|
| 155 |
try:
|
| 156 |
score = 0.0
|
| 157 |
indicators = symbol_data.get('advanced_indicators', {})
|
| 158 |
-
|
| 159 |
-
# التركيز على الأطر الزمنية السريعة (1h, 15m, 5m)
|
| 160 |
for timeframe in ['1h', '15m', '5m']:
|
| 161 |
if timeframe in indicators:
|
| 162 |
tf_indicators = indicators[timeframe]
|
| 163 |
-
|
| 164 |
ema_21 = tf_indicators.get('ema_21')
|
| 165 |
ema_50 = tf_indicators.get('ema_50')
|
| 166 |
adx = tf_indicators.get('adx', 0)
|
| 167 |
-
|
| 168 |
if ema_21 is not None and ema_50 is not None:
|
| 169 |
-
# 1. EMA(21) > EMA(50) (اتجاه صاعد)
|
| 170 |
if ema_21 > ema_50:
|
| 171 |
score += 0.2
|
| 172 |
-
|
| 173 |
-
# 2. ADX > 20 (الاتجاه قوي)
|
| 174 |
if adx > 20:
|
| 175 |
score += 0.1
|
| 176 |
-
|
| 177 |
-
# 3. السعر الحالي فوق المتوسطات (تأكيد إضافي)
|
| 178 |
if symbol_data['current_price'] > ema_21:
|
| 179 |
score += 0.05
|
| 180 |
-
|
| 181 |
-
# تعديل الدرجة النهائية (0.35 * 3 = 1.05 كحد أقصى)
|
| 182 |
return min(score, 1.0)
|
| 183 |
-
except Exception
|
| 184 |
-
print(f"❌ خطأ في استراتيجية تتبع الاتجاه: {error}")
|
| 185 |
-
return None
|
| 186 |
|
| 187 |
def _check_ema_alignment(self, indicators):
|
| 188 |
-
# (دالة مساعدة - قد لا تكون مستخدمة في المنطق الجديد)
|
| 189 |
required_emas = ['ema_9', 'ema_21', 'ema_50']
|
| 190 |
if all(ema in indicators for ema in required_emas):
|
| 191 |
return (indicators['ema_9'] > indicators['ema_21'] > indicators['ema_50'])
|
| 192 |
return False
|
| 193 |
|
| 194 |
async def _mean_reversion_strategy(self, symbol_data, market_context):
|
| 195 |
-
"""
|
| 196 |
-
🔴 (محسنة) استراتيجية العودة للمتوسط - أطر متعددة وقاعدة "OR"
|
| 197 |
-
القاعدة: RSI < 25 أو السعر تحت -2σ Bollinger
|
| 198 |
-
"""
|
| 199 |
try:
|
| 200 |
score = 0.0
|
| 201 |
current_price = symbol_data['current_price']
|
| 202 |
indicators = symbol_data.get('advanced_indicators', {})
|
| 203 |
-
|
| 204 |
-
# التركيز على الأطر (1h, 15m)
|
| 205 |
for timeframe in ['1h', '15m']:
|
| 206 |
if timeframe in indicators:
|
| 207 |
tf_indicators = indicators[timeframe]
|
| 208 |
-
|
| 209 |
rsi_value = tf_indicators.get('rsi', 50)
|
| 210 |
bb_lower = tf_indicators.get('bb_lower')
|
| 211 |
-
bb_upper = tf_indicators.get('bb_upper')
|
| 212 |
-
|
| 213 |
-
|
| 214 |
-
continue # نحتاج بيانات BB
|
| 215 |
-
|
| 216 |
-
# حساب موقع النطاق (0.0 إلى 1.0)
|
| 217 |
-
position_in_band = 0.5 # افتراضي
|
| 218 |
if (bb_upper - bb_lower) > 0:
|
| 219 |
position_in_band = (current_price - bb_lower) / (bb_upper - bb_lower)
|
| 220 |
-
|
| 221 |
-
# 1. RSI < 25 (تشبع بيع قوي)
|
| 222 |
is_rsi_oversold = rsi_value < 25
|
| 223 |
-
|
| 224 |
-
# 2. السعر قرب النطاق السفلي لـ BB (مثال: < 0.1 يعادل تقريباً -2σ)
|
| 225 |
is_bb_oversold = position_in_band < 0.1
|
| 226 |
-
|
| 227 |
-
# 3. تطبيق قاعدة "OR"
|
| 228 |
if is_rsi_oversold or is_bb_oversold:
|
| 229 |
-
score += 0.4
|
| 230 |
-
|
| 231 |
-
# (مكافأة إذا تحققا معاً)
|
| 232 |
if is_rsi_oversold and is_bb_oversold:
|
| 233 |
score += 0.2
|
| 234 |
-
|
| 235 |
-
# تعديل الدرجة النهائية (0.6 * 2 = 1.2 كحد أقصى)
|
| 236 |
return min(score, 1.0)
|
| 237 |
-
except Exception
|
| 238 |
-
print(f"❌ خطأ في استراتيجية العودة للمتوسط: {error}")
|
| 239 |
-
return None
|
| 240 |
|
| 241 |
async def _breakout_momentum_strategy(self, symbol_data, market_context):
|
| 242 |
-
"""
|
| 243 |
-
🔴 (محسنة) استراتيجية زخم الاختراق - قواعد صارمة (حجم + زخم + تقلب)
|
| 244 |
-
القاعدة: اختراق مع حجم > 1.5 * avg_volume(20) + MACD Hist إيجابي + ATR% كافٍ
|
| 245 |
-
"""
|
| 246 |
try:
|
| 247 |
score = 0.0
|
| 248 |
current_price = symbol_data['current_price']
|
| 249 |
indicators = symbol_data.get('advanced_indicators', {})
|
| 250 |
-
|
| 251 |
-
# التركيز على الأطر (1h, 15m, 5m)
|
| 252 |
for timeframe in ['1h', '15m', '5m']:
|
| 253 |
if timeframe in indicators:
|
| 254 |
tf_indicators = indicators[timeframe]
|
| 255 |
-
|
| 256 |
-
# 1. قاعدة الحجم (volume_ratio هو current_vol / avg_vol)
|
| 257 |
volume_ratio = tf_indicators.get('volume_ratio', 0)
|
| 258 |
-
if volume_ratio < 1.5:
|
| 259 |
-
|
| 260 |
-
|
| 261 |
-
score += 0.2 # (مكافأة أولية للحجم)
|
| 262 |
-
|
| 263 |
-
# 2. قاعدة الزخم (MACD Histogram إيجابي)
|
| 264 |
macd_hist = tf_indicators.get('macd_hist', 0)
|
| 265 |
if macd_hist > 0:
|
| 266 |
score += 0.1
|
| 267 |
-
|
| 268 |
-
# 3. قاعدة التقلب (يجب أن يكون هناك تقلب، مثال: ATR% > 1.5%)
|
| 269 |
-
# (نفترض أن 'atr_percent' محسوب في indicators.py)
|
| 270 |
atr_percent = tf_indicators.get('atr_percent', 0)
|
| 271 |
-
if atr_percent > 1.5:
|
| 272 |
score += 0.1
|
| 273 |
-
|
| 274 |
-
# 4. السعر فوق VWAP (تأكيد إضافي)
|
| 275 |
vwap = tf_indicators.get('vwap')
|
| 276 |
if vwap and current_price > vwap:
|
| 277 |
score += 0.05
|
| 278 |
-
|
| 279 |
-
# تعديل الدرجة النهائية (0.45 * 3 = 1.35 كحد أقصى)
|
| 280 |
return min(score, 1.0)
|
| 281 |
-
except Exception
|
| 282 |
-
print(f"❌ خطأ في استراتيجية زخم الاختراق: {error}")
|
| 283 |
-
return None
|
| 284 |
|
| 285 |
async def _volume_spike_strategy(self, symbol_data, market_context):
|
| 286 |
-
"""
|
| 287 |
-
(بدون تغيير جوهري) استراتيجية ارتفاع الحجم - (جيدة للسكالبينغ)
|
| 288 |
-
"""
|
| 289 |
try:
|
| 290 |
score = 0.0
|
| 291 |
indicators = symbol_data.get('advanced_indicators', {})
|
| 292 |
-
|
| 293 |
for timeframe in ['1h', '15m', '5m']:
|
| 294 |
if timeframe in indicators:
|
| 295 |
volume_ratio = indicators[timeframe].get('volume_ratio', 0)
|
| 296 |
-
if volume_ratio > 3.0:
|
| 297 |
-
|
| 298 |
-
elif volume_ratio >
|
| 299 |
-
score += 0.25
|
| 300 |
-
elif volume_ratio > 1.5:
|
| 301 |
-
score += 0.15
|
| 302 |
-
|
| 303 |
return min(score, 1.0)
|
| 304 |
-
except Exception
|
| 305 |
-
print(f"❌ خطأ في استراتيجية ارتفاع الحجم: {error}")
|
| 306 |
-
return None
|
| 307 |
|
| 308 |
async def _whale_tracking_strategy(self, symbol_data, market_context):
|
| 309 |
-
"""
|
| 310 |
-
(بدون تغيير) استراتيجية تتبع الحيتان - (تعتمد على الخدمة الخارجية)
|
| 311 |
-
"""
|
| 312 |
try:
|
| 313 |
whale_data = symbol_data.get('whale_data', {})
|
| 314 |
if not whale_data.get('data_available', False):
|
| 315 |
-
return None
|
| 316 |
-
|
| 317 |
-
# المنطق مجرد في data_manager/whale_monitor
|
| 318 |
whale_signal = await self.data_manager.get_whale_trading_signal(
|
| 319 |
symbol_data['symbol'], whale_data, market_context
|
| 320 |
)
|
| 321 |
-
|
| 322 |
if whale_signal and whale_signal.get('action') != 'HOLD':
|
| 323 |
confidence = whale_signal.get('confidence', 0)
|
| 324 |
if whale_signal.get('action') in ['STRONG_BUY', 'BUY']:
|
| 325 |
return min(confidence * 1.2, 1.0)
|
| 326 |
-
# (لا نهتم بالبيع في نظام SPOT)
|
| 327 |
-
|
| 328 |
-
return None
|
| 329 |
-
except Exception as error:
|
| 330 |
-
print(f"❌ خطأ في استراتيجية تتبع الحيتان: {error}")
|
| 331 |
return None
|
|
|
|
| 332 |
|
| 333 |
async def _pattern_recognition_strategy(self, symbol_data, market_context):
|
| 334 |
-
"""
|
| 335 |
-
(بدون تغيير) استراتيجية التعرف على الأنماط - (تعتمد على MLProcessor)
|
| 336 |
-
"""
|
| 337 |
try:
|
| 338 |
score = 0.0
|
| 339 |
pattern_analysis = symbol_data.get('pattern_analysis')
|
| 340 |
-
|
| 341 |
-
# 1. الاعتماد على نتيجة ChartPatternAnalyzer
|
| 342 |
if pattern_analysis and pattern_analysis.get('pattern_confidence', 0) > 0.6:
|
| 343 |
-
# يجب أن يكون النمط صاعداً
|
| 344 |
if pattern_analysis.get('predicted_direction') == 'up':
|
| 345 |
score += pattern_analysis.get('pattern_confidence', 0) * 0.8
|
| 346 |
else:
|
| 347 |
-
# (منطق احتياطي إذا فشل تحليل النمط)
|
| 348 |
indicators = symbol_data.get('advanced_indicators', {})
|
| 349 |
if '1h' in indicators:
|
| 350 |
tf_indicators = indicators['1h']
|
| 351 |
if (tf_indicators.get('rsi', 50) > 60 and
|
| 352 |
tf_indicators.get('macd_hist', 0) > 0):
|
| 353 |
score += 0.3
|
| 354 |
-
|
| 355 |
return min(score, 1.0)
|
| 356 |
-
except Exception
|
| 357 |
-
print(f"❌ خطأ في استراتيجية التعرف على الأنماط: {error}")
|
| 358 |
-
return None
|
| 359 |
|
| 360 |
async def _hybrid_ai_strategy(self, symbol_data, market_context, base_scores):
|
| 361 |
-
"""
|
| 362 |
-
🔴 (محسنة) استراتيجية "الميتا" الهجينة - تقيس "تطابق" الإشارات
|
| 363 |
-
"""
|
| 364 |
try:
|
| 365 |
score = 0.0
|
| 366 |
-
|
| 367 |
-
# 1. درجة مونت كارلو (أساسية)
|
| 368 |
monte_carlo_prob = symbol_data.get('monte_carlo_probability')
|
| 369 |
if monte_carlo_prob is not None:
|
| 370 |
-
score += monte_carlo_prob * 0.4
|
| 371 |
-
|
| 372 |
-
# 2. درجة توافق الاستراتيجيات (Meta-Score)
|
| 373 |
-
# (نستخدم base_scores التي تم حسابها في evaluate_all_strategies)
|
| 374 |
|
| 375 |
breakout_score = base_scores.get('breakout_momentum', 0)
|
| 376 |
volume_score = base_scores.get('volume_spike', 0)
|
| 377 |
whale_score = base_scores.get('whale_tracking', 0)
|
| 378 |
pattern_score = base_scores.get('pattern_recognition', 0)
|
| 379 |
|
| 380 |
-
|
| 381 |
-
if breakout_score > 0.
|
| 382 |
-
|
| 383 |
-
|
| 384 |
-
# سيناريو 2: اختراق مدعوم بالحيتان
|
| 385 |
-
if breakout_score > 0.6 and whale_score > 0.7:
|
| 386 |
-
score += 0.4
|
| 387 |
-
|
| 388 |
-
# سيناريو 3: نمط مدعوم بالحجم
|
| 389 |
-
if pattern_score > 0.7 and volume_score > 0.5:
|
| 390 |
-
score += 0.2
|
| 391 |
-
|
| 392 |
-
# سيناريو 4: كل شيء متوافق (إشارة قوية جداً)
|
| 393 |
if breakout_score > 0.7 and whale_score > 0.7 and volume_score > 0.7:
|
| 394 |
-
score = 1.0
|
| 395 |
-
|
| 396 |
return max(0.0, min(score, 1.0))
|
| 397 |
-
except Exception
|
| 398 |
-
print(f"❌ خطأ في استراتيجية الهجين الذكية: {error}")
|
| 399 |
-
return None
|
| 400 |
|
| 401 |
-
print("✅ ML Module: Strategy Engine loaded (
|
|
|
|
| 1 |
+
# ml_engine/strategies.py (Updated to use LearningHub for weights)
|
| 2 |
import asyncio
|
| 3 |
|
| 4 |
+
# (Import from internal modules)
|
| 5 |
from .patterns import ChartPatternAnalyzer
|
| 6 |
|
| 7 |
class PatternEnhancedStrategyEngine:
|
| 8 |
+
# 🔴 --- START OF CHANGE --- 🔴
|
| 9 |
+
def __init__(self, data_manager, learning_hub): # (Changed from learning_engine)
|
| 10 |
self.data_manager = data_manager
|
| 11 |
+
self.learning_hub = learning_hub # (Changed from learning_engine)
|
| 12 |
self.pattern_analyzer = ChartPatternAnalyzer()
|
| 13 |
+
# 🔴 --- END OF CHANGE --- 🔴
|
| 14 |
|
| 15 |
async def enhance_strategy_with_patterns(self, strategy_scores, pattern_analysis, symbol):
|
| 16 |
+
"""(Unchanged logic)"""
|
| 17 |
if not pattern_analysis or pattern_analysis.get('pattern_detected') in ['no_clear_pattern', 'insufficient_data']:
|
| 18 |
return strategy_scores
|
| 19 |
|
|
|
|
| 25 |
enhancement_factor = self._calculate_pattern_enhancement(pattern_confidence, pattern_name)
|
| 26 |
enhanced_strategies = self._get_pattern_appropriate_strategies(pattern_name, predicted_direction)
|
| 27 |
|
| 28 |
+
# (Omitted print statements for brevity)
|
| 29 |
|
| 30 |
for strategy in enhanced_strategies:
|
| 31 |
if strategy in strategy_scores:
|
| 32 |
original_score = strategy_scores[strategy]
|
| 33 |
strategy_scores[strategy] = min(original_score * enhancement_factor, 1.0)
|
|
|
|
| 34 |
|
| 35 |
return strategy_scores
|
| 36 |
|
| 37 |
def _calculate_pattern_enhancement(self, pattern_confidence, pattern_name):
|
| 38 |
+
"""(Unchanged logic)"""
|
| 39 |
base_enhancement = 1.0 + (pattern_confidence * 0.3)
|
| 40 |
high_reliability_patterns = ['Double Top', 'Double Bottom', 'Head & Shoulders', 'Cup and Handle']
|
| 41 |
if pattern_name in high_reliability_patterns:
|
|
|
|
| 43 |
return min(base_enhancement, 1.5)
|
| 44 |
|
| 45 |
def _get_pattern_appropriate_strategies(self, pattern_name, direction):
|
| 46 |
+
"""(Unchanged logic)"""
|
| 47 |
reversal_patterns = ['Double Top', 'Double Bottom', 'Head & Shoulders', 'Triple Top', 'Triple Bottom']
|
| 48 |
continuation_patterns = ['Flags', 'Pennants', 'Triangles', 'Rectangles']
|
| 49 |
|
| 50 |
if pattern_name in reversal_patterns:
|
| 51 |
if direction == 'down':
|
| 52 |
+
return ['breakout_momentum', 'trend_following']
|
| 53 |
else:
|
| 54 |
+
return ['mean_reversion', 'breakout_momentum']
|
| 55 |
elif pattern_name in continuation_patterns:
|
| 56 |
return ['trend_following', 'breakout_momentum']
|
| 57 |
else:
|
| 58 |
return ['breakout_momentum', 'hybrid_ai']
|
| 59 |
|
| 60 |
class MultiStrategyEngine:
|
| 61 |
+
# 🔴 --- START OF CHANGE --- 🔴
|
| 62 |
+
def __init__(self, data_manager, learning_hub): # (Changed from learning_engine)
|
| 63 |
self.data_manager = data_manager
|
| 64 |
+
self.learning_hub = learning_hub # (Changed from learning_engine)
|
| 65 |
|
| 66 |
+
# (Pass the hub to the enhancer)
|
| 67 |
+
self.pattern_enhancer = PatternEnhancedStrategyEngine(data_manager, learning_hub)
|
| 68 |
+
# 🔴 --- END OF CHANGE --- 🔴
|
| 69 |
|
| 70 |
self.strategies = {
|
| 71 |
'trend_following': self._trend_following_strategy,
|
|
|
|
| 74 |
'volume_spike': self._volume_spike_strategy,
|
| 75 |
'whale_tracking': self._whale_tracking_strategy,
|
| 76 |
'pattern_recognition': self._pattern_recognition_strategy,
|
| 77 |
+
'hybrid_ai': self._hybrid_ai_strategy
|
| 78 |
}
|
| 79 |
|
| 80 |
async def evaluate_all_strategies(self, symbol_data, market_context):
|
| 81 |
+
"""Evaluate all trading strategies"""
|
| 82 |
try:
|
| 83 |
+
# 🔴 --- START OF CHANGE --- 🔴
|
| 84 |
+
# (Get weights from the new Learning Hub)
|
| 85 |
+
if self.learning_hub and self.learning_hub.initialized:
|
| 86 |
try:
|
| 87 |
market_condition = market_context.get('market_trend', 'sideways_market')
|
| 88 |
+
# (Call the new hub function)
|
| 89 |
+
optimized_weights = await self.learning_hub.get_optimized_weights(market_condition)
|
| 90 |
except Exception as e:
|
| 91 |
+
print(f"⚠️ Error getting optimized weights from hub: {e}. Using defaults.")
|
| 92 |
optimized_weights = await self.get_default_weights()
|
| 93 |
else:
|
| 94 |
optimized_weights = await self.get_default_weights()
|
| 95 |
+
# 🔴 --- END OF CHANGE --- 🔴
|
| 96 |
|
| 97 |
strategy_scores = {}
|
| 98 |
base_scores = {}
|
| 99 |
|
|
|
|
| 100 |
primary_strategies = [s for s in self.strategies.keys() if s != 'hybrid_ai']
|
| 101 |
|
| 102 |
for strategy_name in primary_strategies:
|
|
|
|
| 110 |
weighted_score = base_score * weight
|
| 111 |
strategy_scores[strategy_name] = min(weighted_score, 1.0)
|
| 112 |
except Exception as error:
|
| 113 |
+
print(f"❌ Error evaluating strategy {strategy_name}: {error}")
|
| 114 |
continue
|
| 115 |
|
|
|
|
| 116 |
try:
|
| 117 |
hybrid_score = await self._hybrid_ai_strategy(symbol_data, market_context, base_scores)
|
| 118 |
if hybrid_score is not None:
|
|
|
|
| 120 |
weight = optimized_weights.get('hybrid_ai', 0.1)
|
| 121 |
strategy_scores['hybrid_ai'] = min(hybrid_score * weight, 1.0)
|
| 122 |
except Exception as e:
|
| 123 |
+
print(f"❌ Error in hybrid_ai strategy: {e}")
|
| 124 |
|
| 125 |
+
# Pattern enhancement (Unchanged)
|
| 126 |
pattern_analysis = symbol_data.get('pattern_analysis')
|
| 127 |
if pattern_analysis:
|
| 128 |
strategy_scores = await self.pattern_enhancer.enhance_strategy_with_patterns(
|
|
|
|
| 139 |
return strategy_scores, base_scores
|
| 140 |
|
| 141 |
except Exception as error:
|
| 142 |
+
print(f"❌ Error in evaluate_all_strategies: {error}")
|
| 143 |
return {}, {}
|
| 144 |
|
| 145 |
async def get_default_weights(self):
|
| 146 |
+
"""(Unchanged) Default weights"""
|
| 147 |
return {
|
| 148 |
'trend_following': 0.15,
|
| 149 |
'mean_reversion': 0.12,
|
| 150 |
+
'breakout_momentum': 0.20,
|
| 151 |
+
'volume_spike': 0.13,
|
| 152 |
'whale_tracking': 0.20,
|
| 153 |
+
'pattern_recognition': 0.10,
|
| 154 |
+
'hybrid_ai': 0.10
|
| 155 |
}
|
| 156 |
|
| 157 |
+
#
|
| 158 |
+
# (All individual strategy functions remain unchanged)
|
| 159 |
+
# (_trend_following_strategy, _mean_reversion_strategy, etc.)
|
| 160 |
+
# (Omitted for brevity)
|
| 161 |
+
#
|
| 162 |
async def _trend_following_strategy(self, symbol_data, market_context):
|
|
|
|
|
|
|
|
|
|
|
|
|
| 163 |
try:
|
| 164 |
score = 0.0
|
| 165 |
indicators = symbol_data.get('advanced_indicators', {})
|
|
|
|
|
|
|
| 166 |
for timeframe in ['1h', '15m', '5m']:
|
| 167 |
if timeframe in indicators:
|
| 168 |
tf_indicators = indicators[timeframe]
|
|
|
|
| 169 |
ema_21 = tf_indicators.get('ema_21')
|
| 170 |
ema_50 = tf_indicators.get('ema_50')
|
| 171 |
adx = tf_indicators.get('adx', 0)
|
|
|
|
| 172 |
if ema_21 is not None and ema_50 is not None:
|
|
|
|
| 173 |
if ema_21 > ema_50:
|
| 174 |
score += 0.2
|
|
|
|
|
|
|
| 175 |
if adx > 20:
|
| 176 |
score += 0.1
|
|
|
|
|
|
|
| 177 |
if symbol_data['current_price'] > ema_21:
|
| 178 |
score += 0.05
|
|
|
|
|
|
|
| 179 |
return min(score, 1.0)
|
| 180 |
+
except Exception: return None
|
|
|
|
|
|
|
| 181 |
|
| 182 |
def _check_ema_alignment(self, indicators):
|
|
|
|
| 183 |
required_emas = ['ema_9', 'ema_21', 'ema_50']
|
| 184 |
if all(ema in indicators for ema in required_emas):
|
| 185 |
return (indicators['ema_9'] > indicators['ema_21'] > indicators['ema_50'])
|
| 186 |
return False
|
| 187 |
|
| 188 |
async def _mean_reversion_strategy(self, symbol_data, market_context):
|
|
|
|
|
|
|
|
|
|
|
|
|
| 189 |
try:
|
| 190 |
score = 0.0
|
| 191 |
current_price = symbol_data['current_price']
|
| 192 |
indicators = symbol_data.get('advanced_indicators', {})
|
|
|
|
|
|
|
| 193 |
for timeframe in ['1h', '15m']:
|
| 194 |
if timeframe in indicators:
|
| 195 |
tf_indicators = indicators[timeframe]
|
|
|
|
| 196 |
rsi_value = tf_indicators.get('rsi', 50)
|
| 197 |
bb_lower = tf_indicators.get('bb_lower')
|
| 198 |
+
bb_upper = tf_indicators.get('bb_upper')
|
| 199 |
+
if bb_lower is None or bb_upper is None: continue
|
| 200 |
+
position_in_band = 0.5
|
|
|
|
|
|
|
|
|
|
|
|
|
| 201 |
if (bb_upper - bb_lower) > 0:
|
| 202 |
position_in_band = (current_price - bb_lower) / (bb_upper - bb_lower)
|
|
|
|
|
|
|
| 203 |
is_rsi_oversold = rsi_value < 25
|
|
|
|
|
|
|
| 204 |
is_bb_oversold = position_in_band < 0.1
|
|
|
|
|
|
|
| 205 |
if is_rsi_oversold or is_bb_oversold:
|
| 206 |
+
score += 0.4
|
|
|
|
|
|
|
| 207 |
if is_rsi_oversold and is_bb_oversold:
|
| 208 |
score += 0.2
|
|
|
|
|
|
|
| 209 |
return min(score, 1.0)
|
| 210 |
+
except Exception: return None
|
|
|
|
|
|
|
| 211 |
|
| 212 |
async def _breakout_momentum_strategy(self, symbol_data, market_context):
|
|
|
|
|
|
|
|
|
|
|
|
|
| 213 |
try:
|
| 214 |
score = 0.0
|
| 215 |
current_price = symbol_data['current_price']
|
| 216 |
indicators = symbol_data.get('advanced_indicators', {})
|
|
|
|
|
|
|
| 217 |
for timeframe in ['1h', '15m', '5m']:
|
| 218 |
if timeframe in indicators:
|
| 219 |
tf_indicators = indicators[timeframe]
|
|
|
|
|
|
|
| 220 |
volume_ratio = tf_indicators.get('volume_ratio', 0)
|
| 221 |
+
if volume_ratio < 1.5: continue
|
| 222 |
+
score += 0.2
|
|
|
|
|
|
|
|
|
|
|
|
|
| 223 |
macd_hist = tf_indicators.get('macd_hist', 0)
|
| 224 |
if macd_hist > 0:
|
| 225 |
score += 0.1
|
|
|
|
|
|
|
|
|
|
| 226 |
atr_percent = tf_indicators.get('atr_percent', 0)
|
| 227 |
+
if atr_percent > 1.5:
|
| 228 |
score += 0.1
|
|
|
|
|
|
|
| 229 |
vwap = tf_indicators.get('vwap')
|
| 230 |
if vwap and current_price > vwap:
|
| 231 |
score += 0.05
|
|
|
|
|
|
|
| 232 |
return min(score, 1.0)
|
| 233 |
+
except Exception: return None
|
|
|
|
|
|
|
| 234 |
|
| 235 |
async def _volume_spike_strategy(self, symbol_data, market_context):
|
|
|
|
|
|
|
|
|
|
| 236 |
try:
|
| 237 |
score = 0.0
|
| 238 |
indicators = symbol_data.get('advanced_indicators', {})
|
|
|
|
| 239 |
for timeframe in ['1h', '15m', '5m']:
|
| 240 |
if timeframe in indicators:
|
| 241 |
volume_ratio = indicators[timeframe].get('volume_ratio', 0)
|
| 242 |
+
if volume_ratio > 3.0: score += 0.45
|
| 243 |
+
elif volume_ratio > 2.0: score += 0.25
|
| 244 |
+
elif volume_ratio > 1.5: score += 0.15
|
|
|
|
|
|
|
|
|
|
|
|
|
| 245 |
return min(score, 1.0)
|
| 246 |
+
except Exception: return None
|
|
|
|
|
|
|
| 247 |
|
| 248 |
async def _whale_tracking_strategy(self, symbol_data, market_context):
|
|
|
|
|
|
|
|
|
|
| 249 |
try:
|
| 250 |
whale_data = symbol_data.get('whale_data', {})
|
| 251 |
if not whale_data.get('data_available', False):
|
| 252 |
+
return None
|
|
|
|
|
|
|
| 253 |
whale_signal = await self.data_manager.get_whale_trading_signal(
|
| 254 |
symbol_data['symbol'], whale_data, market_context
|
| 255 |
)
|
|
|
|
| 256 |
if whale_signal and whale_signal.get('action') != 'HOLD':
|
| 257 |
confidence = whale_signal.get('confidence', 0)
|
| 258 |
if whale_signal.get('action') in ['STRONG_BUY', 'BUY']:
|
| 259 |
return min(confidence * 1.2, 1.0)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 260 |
return None
|
| 261 |
+
except Exception: return None
|
| 262 |
|
| 263 |
async def _pattern_recognition_strategy(self, symbol_data, market_context):
|
|
|
|
|
|
|
|
|
|
| 264 |
try:
|
| 265 |
score = 0.0
|
| 266 |
pattern_analysis = symbol_data.get('pattern_analysis')
|
|
|
|
|
|
|
| 267 |
if pattern_analysis and pattern_analysis.get('pattern_confidence', 0) > 0.6:
|
|
|
|
| 268 |
if pattern_analysis.get('predicted_direction') == 'up':
|
| 269 |
score += pattern_analysis.get('pattern_confidence', 0) * 0.8
|
| 270 |
else:
|
|
|
|
| 271 |
indicators = symbol_data.get('advanced_indicators', {})
|
| 272 |
if '1h' in indicators:
|
| 273 |
tf_indicators = indicators['1h']
|
| 274 |
if (tf_indicators.get('rsi', 50) > 60 and
|
| 275 |
tf_indicators.get('macd_hist', 0) > 0):
|
| 276 |
score += 0.3
|
|
|
|
| 277 |
return min(score, 1.0)
|
| 278 |
+
except Exception: return None
|
|
|
|
|
|
|
| 279 |
|
| 280 |
async def _hybrid_ai_strategy(self, symbol_data, market_context, base_scores):
|
|
|
|
|
|
|
|
|
|
| 281 |
try:
|
| 282 |
score = 0.0
|
|
|
|
|
|
|
| 283 |
monte_carlo_prob = symbol_data.get('monte_carlo_probability')
|
| 284 |
if monte_carlo_prob is not None:
|
| 285 |
+
score += monte_carlo_prob * 0.4
|
|
|
|
|
|
|
|
|
|
| 286 |
|
| 287 |
breakout_score = base_scores.get('breakout_momentum', 0)
|
| 288 |
volume_score = base_scores.get('volume_spike', 0)
|
| 289 |
whale_score = base_scores.get('whale_tracking', 0)
|
| 290 |
pattern_score = base_scores.get('pattern_recognition', 0)
|
| 291 |
|
| 292 |
+
if breakout_score > 0.7 and volume_score > 0.6: score += 0.3
|
| 293 |
+
if breakout_score > 0.6 and whale_score > 0.7: score += 0.4
|
| 294 |
+
if pattern_score > 0.7 and volume_score > 0.5: score += 0.2
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 295 |
if breakout_score > 0.7 and whale_score > 0.7 and volume_score > 0.7:
|
| 296 |
+
score = 1.0
|
|
|
|
| 297 |
return max(0.0, min(score, 1.0))
|
| 298 |
+
except Exception: return None
|
|
|
|
|
|
|
| 299 |
|
| 300 |
+
print("✅ ML Module: Strategy Engine loaded (V3 - Integrated LearningHub for weights)")
|