Spaces:
Running
Running
Update data_manager.py
Browse files- data_manager.py +52 -98
data_manager.py
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
# ml_engine/data_manager.py
|
| 2 |
-
# (V12.
|
| 3 |
|
| 4 |
import os
|
| 5 |
import asyncio
|
|
@@ -23,7 +23,12 @@ except ImportError:
|
|
| 23 |
from ml_engine.indicators import AdvancedTechnicalAnalyzer
|
| 24 |
from ml_engine.monte_carlo import MonteCarloAnalyzer
|
| 25 |
from ml_engine.ranker import Layer1Ranker
|
| 26 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 27 |
|
| 28 |
logging.getLogger("httpx").setLevel(logging.WARNING)
|
| 29 |
logging.getLogger("httpcore").setLevel(logging.WARNING)
|
|
@@ -32,11 +37,10 @@ logging.getLogger("ccxt").setLevel(logging.WARNING)
|
|
| 32 |
class DataManager:
|
| 33 |
def __init__(self, contracts_db, whale_monitor, r2_service=None):
|
| 34 |
# ==================================================================
|
| 35 |
-
# ⚙️ إعدادات التحكم المركزية (V12
|
| 36 |
# ==================================================================
|
| 37 |
-
self.SCREENING_THRESHOLD = 0.40 #
|
| 38 |
-
self.TITAN_ENTRY_THRESHOLD = 0.90 # عتبة
|
| 39 |
-
self.PROFIT_SAVE_THRESHOLD = 0.30 # عتبة الخروج الذكي (إذا انخفضت ثقة Titan عن هذا الحد)
|
| 40 |
# ==================================================================
|
| 41 |
|
| 42 |
self.contracts_db = contracts_db or {}
|
|
@@ -50,33 +54,40 @@ class DataManager:
|
|
| 50 |
|
| 51 |
self.http_client = None
|
| 52 |
self.market_cache = {}
|
| 53 |
-
self.last_market_load = None
|
| 54 |
|
| 55 |
self.technical_analyzer = AdvancedTechnicalAnalyzer()
|
| 56 |
self.mc_analyzer = MonteCarloAnalyzer()
|
|
|
|
|
|
|
| 57 |
self.layer1_ranker = None
|
|
|
|
| 58 |
|
| 59 |
async def initialize(self):
|
| 60 |
-
"""تهيئة الاتصالات وتحميل
|
| 61 |
self.http_client = httpx.AsyncClient(timeout=30.0)
|
| 62 |
await self._load_markets()
|
| 63 |
|
| 64 |
-
print(" > [DataManager
|
| 65 |
try:
|
| 66 |
-
#
|
| 67 |
self.layer1_ranker = Layer1Ranker(model_path="ml_models/layer1_ranker.lgbm")
|
| 68 |
await self.layer1_ranker.initialize()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 69 |
except Exception as e:
|
| 70 |
-
print(f"⚠️ [DataManager]
|
| 71 |
|
| 72 |
-
print(f"✅ DataManager V12.
|
| 73 |
|
| 74 |
async def _load_markets(self):
|
| 75 |
try:
|
| 76 |
if self.exchange:
|
| 77 |
await self.exchange.load_markets()
|
| 78 |
self.market_cache = self.exchange.markets
|
| 79 |
-
self.last_market_load = datetime.now()
|
| 80 |
except Exception as e:
|
| 81 |
print(f"❌ [DataManager] فشل تحميل الأسواق: {e}")
|
| 82 |
|
|
@@ -84,123 +95,66 @@ class DataManager:
|
|
| 84 |
if self.http_client: await self.http_client.aclose()
|
| 85 |
if self.exchange: await self.exchange.close()
|
| 86 |
|
| 87 |
-
|
| 88 |
-
|
| 89 |
-
|
| 90 |
-
|
| 91 |
-
|
| 92 |
-
'timestamp': datetime.now().isoformat(),
|
| 93 |
-
'bitcoin_price_usd': prices.get('bitcoin'),
|
| 94 |
-
'fear_and_greed_index': sentiment.get('feargreed_value') if sentiment else 50,
|
| 95 |
-
'market_trend': self._determine_market_trend(prices.get('bitcoin'), sentiment)
|
| 96 |
-
}
|
| 97 |
-
except: return {'market_trend': 'UNKNOWN'}
|
| 98 |
-
|
| 99 |
-
async def get_sentiment_live_async(self):
|
| 100 |
-
try:
|
| 101 |
-
async with httpx.AsyncClient(timeout=10) as client:
|
| 102 |
-
resp = await client.get("https://api.alternative.me/fng/")
|
| 103 |
-
if resp.status_code == 200:
|
| 104 |
-
data = resp.json()
|
| 105 |
-
return {"feargreed_value": int(data['data'][0]['value'])}
|
| 106 |
-
except: return None
|
| 107 |
-
|
| 108 |
-
async def _get_prices_live(self):
|
| 109 |
-
if not self.exchange: return {}
|
| 110 |
-
try:
|
| 111 |
-
btc = await self.exchange.fetch_ticker('BTC/USDT')
|
| 112 |
-
return {'bitcoin': btc['last']}
|
| 113 |
-
except: return {}
|
| 114 |
-
|
| 115 |
-
def _determine_market_trend(self, btc_price, sentiment):
|
| 116 |
-
# منطق بسيط لتحديد الاتجاه العام
|
| 117 |
-
if not btc_price: return "NEUTRAL"
|
| 118 |
-
score = 0
|
| 119 |
-
if btc_price > 65000: score += 1 # مثال
|
| 120 |
-
if sentiment and sentiment.get('feargreed_value', 50) > 60: score += 1
|
| 121 |
-
return "BULLISH" if score > 0 else "BEARISH"
|
| 122 |
|
| 123 |
# ==================================================================
|
| 124 |
-
# 🔴 الطبقة 1: الغربلة السريعة
|
| 125 |
# ==================================================================
|
| 126 |
async def layer1_rapid_screening(self) -> List[Dict[str, Any]]:
|
| 127 |
-
"""غربلة أولية سريعة لتقليل عدد العملات التي يحللها Titan"""
|
| 128 |
print(f"🔍 [Layer 1] بدء الغربلة السريعة...")
|
| 129 |
volume_data = await self._get_volume_data_live()
|
| 130 |
if not volume_data: return []
|
| 131 |
|
| 132 |
-
#
|
| 133 |
candidates = volume_data[:150]
|
| 134 |
-
|
| 135 |
-
|
| 136 |
-
# تحليل سريع جداً (فريم 1H فقط)
|
| 137 |
-
batch_size = 30 # دفعة أكبر قليلاً للسرعة
|
| 138 |
-
for i in range(0, len(candidates), batch_size):
|
| 139 |
-
batch = candidates[i:i+batch_size]
|
| 140 |
-
tasks = [self._fetch_ohlcv_live(c['symbol'], '1h', 150) for c in batch]
|
| 141 |
-
results = await asyncio.gather(*tasks, return_exceptions=True)
|
| 142 |
-
|
| 143 |
-
for j, candles in enumerate(results):
|
| 144 |
-
if isinstance(candles, list) and len(candles) >= 100:
|
| 145 |
-
# هنا يمكن إضافة منطق فلترة بسيط (مثلاً: فوق متوسط 50)
|
| 146 |
-
# حالياً سنمرر معظم العملات الجيدة لـ Titan ليقرر هو
|
| 147 |
-
candidates[i+j]['layer1_score'] = 0.5 # درجة مبدئية
|
| 148 |
-
qualified.append(candidates[i+j])
|
| 149 |
-
|
| 150 |
-
print(f"✅ [Layer 1] تأهل {len(qualified)} عملة للتحليل العميق بواسطة Titan.")
|
| 151 |
-
return qualified
|
| 152 |
|
| 153 |
async def _get_volume_data_live(self):
|
| 154 |
try:
|
| 155 |
tickers = await self.exchange.fetch_tickers()
|
| 156 |
data = []
|
| 157 |
for s, t in tickers.items():
|
|
|
|
| 158 |
if s.endswith('/USDT') and t['quoteVolume'] and t['quoteVolume'] > 100000:
|
| 159 |
data.append({'symbol': s, 'dollar_volume': t['quoteVolume'], 'current_price': t['last']})
|
| 160 |
data.sort(key=lambda x: x['dollar_volume'], reverse=True)
|
| 161 |
return data
|
| 162 |
except: return []
|
| 163 |
|
| 164 |
-
async def _fetch_ohlcv_live(self, symbol, tf, limit):
|
| 165 |
-
try: return await self.exchange.fetch_ohlcv(symbol, tf, limit=limit)
|
| 166 |
-
except: return None
|
| 167 |
-
|
| 168 |
# ==================================================================
|
| 169 |
-
# 🔵 خط أنابيب
|
| 170 |
# ==================================================================
|
| 171 |
async def stream_ohlcv_data(self, symbols: List[Dict], queue: asyncio.Queue):
|
| 172 |
-
"""
|
| 173 |
-
|
| 174 |
-
|
| 175 |
-
"""
|
| 176 |
-
print(f"🚚 [Titan Data Stream] جلب بيانات عميقة لـ {len(symbols)} عملة...")
|
| 177 |
-
|
| 178 |
-
tfs = ['5m', '15m', '1h', '4h', '1d'] # أطر Titan الكاملة
|
| 179 |
-
limit = 500 # عمق كافٍ لحساب كل المؤشرات بدقة (مثل EMA200)
|
| 180 |
|
| 181 |
for sym_data in symbols:
|
| 182 |
sym = sym_data['symbol']
|
| 183 |
-
# جلب متوازي لكل الأطر الزمنية للعملة الواحدة
|
| 184 |
tasks = [self._fetch_ohlcv_live(sym, tf, limit) for tf in tfs]
|
| 185 |
results = await asyncio.gather(*tasks, return_exceptions=False)
|
| 186 |
|
| 187 |
-
|
| 188 |
-
complete_data = True
|
| 189 |
for i, res in enumerate(results):
|
| 190 |
-
if res and isinstance(res, list) and len(res) >= 200:
|
| 191 |
-
|
| 192 |
-
else:
|
| 193 |
-
# إذا فقدنا إطاراً رئيسياً (مثل 5m)، قد لا يعمل Titan بدقة
|
| 194 |
-
if tfs[i] == '5m': complete_data = False
|
| 195 |
|
| 196 |
-
|
| 197 |
-
|
| 198 |
-
|
|
|
|
| 199 |
|
| 200 |
-
|
| 201 |
-
await asyncio.sleep(0.1)
|
| 202 |
|
| 203 |
-
await queue.put(None)
|
| 204 |
-
|
|
|
|
|
|
|
|
|
|
| 205 |
|
| 206 |
-
print("✅ DataManager V12.
|
|
|
|
| 1 |
# ml_engine/data_manager.py
|
| 2 |
+
# (V12.1 - Titan + Hybrid Support Pipeline)
|
| 3 |
|
| 4 |
import os
|
| 5 |
import asyncio
|
|
|
|
| 23 |
from ml_engine.indicators import AdvancedTechnicalAnalyzer
|
| 24 |
from ml_engine.monte_carlo import MonteCarloAnalyzer
|
| 25 |
from ml_engine.ranker import Layer1Ranker
|
| 26 |
+
# 🔥 إعادة استيراد محرك الأنماط لدعم النظام الهجين
|
| 27 |
+
try:
|
| 28 |
+
from ml_engine.patterns import ChartPatternAnalyzer
|
| 29 |
+
except ImportError:
|
| 30 |
+
print("⚠️ [DataManager] لم يتم العثور على ml_engine/patterns.py")
|
| 31 |
+
ChartPatternAnalyzer = None
|
| 32 |
|
| 33 |
logging.getLogger("httpx").setLevel(logging.WARNING)
|
| 34 |
logging.getLogger("httpcore").setLevel(logging.WARNING)
|
|
|
|
| 37 |
class DataManager:
|
| 38 |
def __init__(self, contracts_db, whale_monitor, r2_service=None):
|
| 39 |
# ==================================================================
|
| 40 |
+
# ⚙️ إعدادات التحكم المركزية (V12 Hybrid Thresholds)
|
| 41 |
# ==================================================================
|
| 42 |
+
self.SCREENING_THRESHOLD = 0.40 # غربلة أولية
|
| 43 |
+
self.TITAN_ENTRY_THRESHOLD = 0.90 # عتبة Titan الصارمة
|
|
|
|
| 44 |
# ==================================================================
|
| 45 |
|
| 46 |
self.contracts_db = contracts_db or {}
|
|
|
|
| 54 |
|
| 55 |
self.http_client = None
|
| 56 |
self.market_cache = {}
|
|
|
|
| 57 |
|
| 58 |
self.technical_analyzer = AdvancedTechnicalAnalyzer()
|
| 59 |
self.mc_analyzer = MonteCarloAnalyzer()
|
| 60 |
+
|
| 61 |
+
# نماذج الطبقة الأولى والثانية المساندة
|
| 62 |
self.layer1_ranker = None
|
| 63 |
+
self.pattern_analyzer = None # 🔥 تمت إعادته لتجنب AttributeError
|
| 64 |
|
| 65 |
async def initialize(self):
|
| 66 |
+
"""تهيئة الاتصالات وتحميل جميع النماذج المساندة"""
|
| 67 |
self.http_client = httpx.AsyncClient(timeout=30.0)
|
| 68 |
await self._load_markets()
|
| 69 |
|
| 70 |
+
print(" > [DataManager] تهيئة النماذج المساندة (Ranker + Patterns)...")
|
| 71 |
try:
|
| 72 |
+
# 1. الكاشف المصغر (Ranker)
|
| 73 |
self.layer1_ranker = Layer1Ranker(model_path="ml_models/layer1_ranker.lgbm")
|
| 74 |
await self.layer1_ranker.initialize()
|
| 75 |
+
|
| 76 |
+
# 2. محرك الأنماط (Patterns) - للنظام الهجين
|
| 77 |
+
if ChartPatternAnalyzer:
|
| 78 |
+
self.pattern_analyzer = ChartPatternAnalyzer(r2_service=self.r2_service)
|
| 79 |
+
await self.pattern_analyzer.initialize()
|
| 80 |
+
|
| 81 |
except Exception as e:
|
| 82 |
+
print(f"⚠️ [DataManager] تحذير أثناء تهيئة النماذج المساندة: {e}")
|
| 83 |
|
| 84 |
+
print(f"✅ DataManager V12.1 initialized (Hybrid Ready).")
|
| 85 |
|
| 86 |
async def _load_markets(self):
|
| 87 |
try:
|
| 88 |
if self.exchange:
|
| 89 |
await self.exchange.load_markets()
|
| 90 |
self.market_cache = self.exchange.markets
|
|
|
|
| 91 |
except Exception as e:
|
| 92 |
print(f"❌ [DataManager] فشل تحميل الأسواق: {e}")
|
| 93 |
|
|
|
|
| 95 |
if self.http_client: await self.http_client.aclose()
|
| 96 |
if self.exchange: await self.exchange.close()
|
| 97 |
|
| 98 |
+
# تنظيف الذاكرة عند الإغلاق
|
| 99 |
+
if self.pattern_analyzer and hasattr(self.pattern_analyzer, 'clear_memory'):
|
| 100 |
+
self.pattern_analyzer.clear_memory()
|
| 101 |
+
if self.layer1_ranker and hasattr(self.layer1_ranker, 'clear_memory'):
|
| 102 |
+
self.layer1_ranker.clear_memory()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 103 |
|
| 104 |
# ==================================================================
|
| 105 |
+
# 🔴 الطبقة 1: الغربلة السريعة
|
| 106 |
# ==================================================================
|
| 107 |
async def layer1_rapid_screening(self) -> List[Dict[str, Any]]:
|
|
|
|
| 108 |
print(f"🔍 [Layer 1] بدء الغربلة السريعة...")
|
| 109 |
volume_data = await self._get_volume_data_live()
|
| 110 |
if not volume_data: return []
|
| 111 |
|
| 112 |
+
# توسيع النطاق لـ 150 عملة لزيادة فرص Titan
|
| 113 |
candidates = volume_data[:150]
|
| 114 |
+
print(f"✅ [Layer 1] تم تمرير {len(candidates)} عملة للتحليل الهجين.")
|
| 115 |
+
return candidates
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 116 |
|
| 117 |
async def _get_volume_data_live(self):
|
| 118 |
try:
|
| 119 |
tickers = await self.exchange.fetch_tickers()
|
| 120 |
data = []
|
| 121 |
for s, t in tickers.items():
|
| 122 |
+
# فلترة العملات ذات السيولة الضعيفة جداً (< 100k)
|
| 123 |
if s.endswith('/USDT') and t['quoteVolume'] and t['quoteVolume'] > 100000:
|
| 124 |
data.append({'symbol': s, 'dollar_volume': t['quoteVolume'], 'current_price': t['last']})
|
| 125 |
data.sort(key=lambda x: x['dollar_volume'], reverse=True)
|
| 126 |
return data
|
| 127 |
except: return []
|
| 128 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 129 |
# ==================================================================
|
| 130 |
+
# 🔵 خط أنابيب البيانات الهجين (Hybrid Data Stream)
|
| 131 |
# ==================================================================
|
| 132 |
async def stream_ohlcv_data(self, symbols: List[Dict], queue: asyncio.Queue):
|
| 133 |
+
"""جلب بيانات عميقة (500 شمعة) لكافة الأطر المطلوبة"""
|
| 134 |
+
tfs = ['5m', '15m', '1h', '4h', '1d']
|
| 135 |
+
limit = 500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 136 |
|
| 137 |
for sym_data in symbols:
|
| 138 |
sym = sym_data['symbol']
|
|
|
|
| 139 |
tasks = [self._fetch_ohlcv_live(sym, tf, limit) for tf in tfs]
|
| 140 |
results = await asyncio.gather(*tasks, return_exceptions=False)
|
| 141 |
|
| 142 |
+
ohlcv = {}
|
|
|
|
| 143 |
for i, res in enumerate(results):
|
| 144 |
+
if res and isinstance(res, list) and len(res) >= 200:
|
| 145 |
+
ohlcv[tfs[i]] = res
|
|
|
|
|
|
|
|
|
|
| 146 |
|
| 147 |
+
# يجب توفر معظم الأطر لعمل النظام الهجين بدقة
|
| 148 |
+
if len(ohlcv) >= 4 and '5m' in ohlcv:
|
| 149 |
+
sym_data['ohlcv'] = ohlcv
|
| 150 |
+
await queue.put([sym_data])
|
| 151 |
|
| 152 |
+
await asyncio.sleep(0.1) # تجنب حظر API
|
|
|
|
| 153 |
|
| 154 |
+
await queue.put(None)
|
| 155 |
+
|
| 156 |
+
async def _fetch_ohlcv_live(self, symbol, tf, limit):
|
| 157 |
+
try: return await self.exchange.fetch_ohlcv(symbol, tf, limit=limit)
|
| 158 |
+
except: return None
|
| 159 |
|
| 160 |
+
print("✅ DataManager V12.1 (Hybrid Support) loaded.")
|