Update app.py

app.py

@@ -1,4 +1,4 @@
-# app.py (Updated to V8.4 - Whale Learning Loop Activated)
+# app.py (Updated to V8.5 - Vader Injection Fix)
 import os
 import traceback
 import signal
@@ -89,7 +89,7 @@ async def initialize_services():
     """تهيئة جميع الخدمات بشكل منفصل"""
     global r2_service_global, data_manager_global, llm_service_global
     global learning_hub_global, trade_manager_global, sentiment_analyzer_global
-    global symbol_whale_monitor_global, news_fetcher_global # (V8.1)
+    global symbol_whale_monitor_global, news_fetcher_global, VADER_ANALYZER # (V8.1)
     try:
         # 🔴 --- START OF CHANGE (V7.0) --- 🔴
         print("🚀 بدء تهيئة الخدمات (بنية Sentry الجديدة V5.9)...")
@@ -118,12 +118,10 @@ async def initialize_services():
         state_manager.set_service_initialized('data_manager'); 
         print("   ✅ DataManager مهيأ (ومحرك الأنماط V8 مُحمّل)")
 
-        # 🔴 --- START OF CHANGE (V8.3 - Whale Link) --- 🔴
         # (ربط DataManager بـ WhaleMonitor لحل الاعتمادية الدائرية للسعر)
         if symbol_whale_monitor_global:
             symbol_whale_monitor_global.data_manager = data_manager_global
             print("   ✅ [Whale Link] تم ربط DataManager بـ WhaleMonitor.")
-        # 🔴 --- END OF CHANGE --- 🔴
 
         print("   🔄 تهيئة LLMService (Layer 1 Brain)...");
         llm_service_global = LLMService();
@@ -139,8 +137,13 @@ async def initialize_services():
         news_fetcher_global = NewsFetcher()
         state_manager.set_service_initialized('news_fetcher');
         print("   ✅ NewsFetcher (V8.1) مهيأ")
-        # (تمرير NewsFetcher إلى LLMService - مهم لإعادة التحليل)
+        
+        # 🔴 --- START OF CHANGE (V8.5 - Vader Injection) --- 🔴
+        # (تمرير NewsFetcher و VADER إلى LLMService - مهم لإعادة التحليل)
         llm_service_global.news_fetcher = news_fetcher_global
+        llm_service_global.vader_analyzer = VADER_ANALYZER # (هذا هو السطر المضاف)
+        print("   ✅ [LLM Link] تم ربط NewsFetcher و VaderAnalyzer بـ LLMService.")
+        # 🔴 --- END OF CHANGE --- 🔴
 
         print("   🔄 تهيئة محور التعلم (Hub)...");
         learning_hub_global = LearningHubManager(
@@ -346,8 +349,7 @@ async def run_3_layer_analysis_explorer() -> List[Dict[str, Any]]:
                 success_item['news_score_raw'] = 0.0 # درجة VADER الخام
                 success_item['statistical_news_pnl'] = 0.0 # الدرجة المتعلمة
                 
-                # 🔴 --- (V8.3) تمرير البيانات التي يحتاجها تحليل الحيتان --- 🔴
-                # (تمرير dollar_volume الذي تم حسابه في layer1_rapid_screening)
+                # (V8.3) تمرير البيانات التي يحتاجها تحليل الحيتان
                 original_l1_data = next((c for c in layer1_candidates if c['symbol'] == symbol), None)
                 if original_l1_data:
                     success_item['dollar_volume'] = original_l1_data.get('dollar_volume', 0.0)
@@ -358,22 +360,18 @@ async def run_3_layer_analysis_explorer() -> List[Dict[str, Any]]:
 
         layer2_candidates.sort(key=lambda x: x.get('enhanced_final_score', 0), reverse=True)
         
-        # (V8.1) نأخذ أفضل 10 هنا (بدلاً من 5) لإضافة الحيتان والأخبار
         target_count = min(10, len(layer2_candidates))
         
         final_layer2_candidates = layer2_candidates[:target_count]
 
-        # 🔴 --- START OF CHANGE (V8.3 - Whale Link) --- 🔴
         print(f"\n🐋📰 Layer 1.4 (Optimized): Fetching Whale Data & News for top {len(final_layer2_candidates)} candidates...")
         
         # (دالة مساعدة لجلب الحيتان)
         async def get_whale_data_for_candidate(candidate):
             symbol = candidate.get('symbol', 'UNKNOWN')
-            # (جلب حجم التداول اليومي من المرشح)
             symbol_daily_volume = candidate.get('dollar_volume', 0.0)
             
             try:
-                # (تمرير حجم التداول إلى الدالة)
                 data = await data_manager_global.get_whale_data_for_symbol(
                     symbol, 
                     daily_volume_usd=symbol_daily_volume
@@ -385,7 +383,6 @@ async def run_3_layer_analysis_explorer() -> List[Dict[str, Any]]:
             except Exception as e:
                 print(f"   ❌ [Whale Fetch] {symbol} - Error: {e}")
                 candidate['whale_data'] = {'data_available': False, 'error': str(e)}
-        # 🔴 --- END OF CHANGE --- 🔴
 
         # (دالة مساعدة لجلب الأخبار وتحليل VADER)
         async def get_news_data_for_candidate(candidate):
@@ -396,16 +393,14 @@ async def run_3_layer_analysis_explorer() -> List[Dict[str, Any]]:
                 return
 
             try:
-                # 1. جلب نص الأخبار (باستخدام NewsFetcher المطور)
                 news_text = await news_fetcher_global.get_news_for_symbol(symbol)
                 candidate['news_text'] = news_text
                 
-                # 2. حساب درجة VADER الخام
                 if "No specific news found" in news_text or not news_text:
-                    candidate['news_score_raw'] = 0.0 # محايد
+                    candidate['news_score_raw'] = 0.0
                 else:
                     vader_score = VADER_ANALYZER.polarity_scores(news_text)
-                    candidate['news_score_raw'] = vader_score.get('compound', 0.0) # النتيجة (-1 إلى +1)
+                    candidate['news_score_raw'] = vader_score.get('compound', 0.0)
                 
             except Exception as e:
                 print(f"   ❌ [News Fetch] {symbol} - Error: {e}")
@@ -424,17 +419,14 @@ async def run_3_layer_analysis_explorer() -> List[Dict[str, Any]]:
         print("   🔄 Re-calculating enhanced scores with new Whale & Statistical News data...")
         for candidate in final_layer2_candidates:
             try:
-                # (V8.2) جلب درجة VADER الخام
                 raw_vader_score = candidate.get('news_score_raw', 0.0)
                 
-                # (V8.2) جلب الربح/الخسارة الإحصائي المرتبط بهذه الدرجة من محور التعلم
                 if learning_hub_global:
                     statistical_pnl = await learning_hub_global.get_statistical_news_score(raw_vader_score)
                     candidate['statistical_news_pnl'] = statistical_pnl
                 else:
-                    candidate['statistical_news_pnl'] = 0.0 # (الوضع الآمن)
+                    candidate['statistical_news_pnl'] = 0.0
                 
-                # (هذه الدالة في processor.py تم تعديلها لتقبل statistical_news_pnl)
                 new_score = ml_processor._calculate_enhanced_final_score(candidate)
                 candidate['enhanced_final_score'] = new_score
             except Exception as e:
@@ -442,7 +434,6 @@ async def run_3_layer_analysis_explorer() -> List[Dict[str, Any]]:
         
         final_layer2_candidates.sort(key=lambda x: x.get('enhanced_final_score', 0), reverse=True)
         print("   ✅ Top scores updated (with Stat. News + Whale) and re-sorted.")
-        # 🔴 --- END OF CHANGE (V8.2) --- 🔴
 
 
         print(f"\n🔬 Layer 1.5: Running Advanced MC (GARCH+LGBM) on top {len(final_layer2_candidates)} candidates...")
@@ -476,7 +467,6 @@ async def run_3_layer_analysis_explorer() -> List[Dict[str, Any]]:
         await r2_service_global.save_candidates_async(final_layer2_candidates)
 
         print("\n🧠 Layer 1.6: LLM Strategic Analysis (Explorer Brain)...")
-        # (V8.1) نختار أفضل 5 الآن *بعد* اكتمال كل التحليلات (حيتان + أخبار + MC متقدم)
         top_5_for_llm = final_layer2_candidates[:5]
         print(f"   (Sending Top {len(top_5_for_llm)} candidates to LLM)")
         
@@ -492,8 +482,6 @@ async def run_3_layer_analysis_explorer() -> List[Dict[str, Any]]:
 
                 candidate['sentiment_data'] = await data_manager_global.get_market_context_async()
 
-                # (V8.2) النموذج الضخم سيحصل على 'news_text' (النص الخام)
-                # ولن يرى 'statistical_news_pnl' (الدرجة المتعلمة)
                 llm_analysis = await llm_service_global.get_trading_decision(candidate) 
 
                 if llm_analysis and llm_analysis.get('action') in ['WATCH']:
@@ -527,19 +515,17 @@ async def run_3_layer_analysis_explorer() -> List[Dict[str, Any]]:
 
         top_watchlist = watchlist_candidates
         
-        # --- (V8-MODIFICATION) إنشاء وحفظ سجل التدقيق ---
         print("📊 إنشاء سجل تدقيق لمحرك الأنماط V8...")
         audit_log = {
             "log_id": f"audit_{int(datetime.now().timestamp())}",
             "timestamp": datetime.now().isoformat(),
-            "model_key": "lgbm_pattern_model_combined.pkl", # (من النموذج الناجح)
-            "scaler_key": "scaler_combined.pkl", # (من النموذج الناجH)
-            "model_accuracy": 0.5870, # (من ملف evaluation_results.txt)
+            "model_key": "lgbm_pattern_model_combined.pkl", 
+            "scaler_key": "scaler_combined.pkl",
+            "model_accuracy": 0.5870, 
             "predictions": []
         }
 
-        # (اجمع كل التوقعات التي حدثت في هذه الدورة)
-        for candidate in final_layer2_candidates: # (استخدام المرشحين النهائيين)
+        for candidate in final_layer2_candidates: 
             pattern_analysis = candidate.get('pattern_analysis', {})
             audit_entry = {
                 "symbol": candidate.get('symbol', 'N/A'),
@@ -547,14 +533,12 @@ async def run_3_layer_analysis_explorer() -> List[Dict[str, Any]]:
                 "pattern_detected": pattern_analysis.get('pattern_detected', 'N/A'),
                 "confidence": pattern_analysis.get('pattern_confidence', 0),
                 "predicted_direction": pattern_analysis.get('predicted_direction', 'neutral'),
-                "error": "None" # (نفترض عدم وجود خطأ، وإلا لما كان مرشحاً)
+                "error": "None" 
             }
             audit_log["predictions"].append(audit_entry)
 
-        # (حفظ السجل في R2)
         if r2_service_global:
             await r2_service_global.save_analysis_audit_log_async(audit_log)
-        # --- (نهاية الإضافة) ---
         
         print(f"✅ Explorer analysis complete. Sending {len(top_watchlist)} candidates to Sentry.")
         return top_watchlist
@@ -592,7 +576,6 @@ async def re_analyze_open_trade_async(trade_data):
             if not ohlcv_data_list: print(f"⚠️ Failed to get re-analysis data for {symbol}"); return None
             ohlcv_data = ohlcv_data_list[0]
 
-            # 🔴 --- START OF CHANGE (V8.3 - Whale Link) --- 🔴
             # (جلب حجم التداول اليومي الحالي لإعادة التحليل)
             print(f"   🔄 [Re-Analyze] Fetching current daily volume for {symbol}...")
             symbol_daily_volume = await data_manager_global.get_symbol_daily_volume(symbol)
@@ -601,7 +584,6 @@ async def re_analyze_open_trade_async(trade_data):
                 symbol,
                 daily_volume_usd=symbol_daily_volume
             )
-            # 🔴 --- END OF CHANGE --- 🔴
 
             ml_processor = MLProcessor(market_context, data_manager_global, learning_hub_global)
 
@@ -621,7 +603,6 @@ async def re_analyze_open_trade_async(trade_data):
             processed_data['ohlcv'] = processed_data['raw_ohlcv']
             processed_data['sentiment_data'] = market_context
             
-            # 🔴 --- START OF CHANGE (V8.2) --- 🔴
             # (جلب الأخبار ودرجة VADER الخام لتمريرها إلى Reflector لاحقاً)
             if news_fetcher_global and VADER_ANALYZER:
                 try:
@@ -637,9 +618,7 @@ async def re_analyze_open_trade_async(trade_data):
             else:
                 processed_data['news_text'] = "News analysis disabled."
                 processed_data['news_score'] = 0.0
-            # 🔴 --- END OF CHANGE --- 🔴
 
-            # (LLMService سيستخدم NewsFetcher الخاص به لجلب أحدث الأخبار لتحليلها)
             re_analysis_decision = await llm_service_global.re_analyze_trade_async(trade_data, processed_data)
 
             if re_analysis_decision:
@@ -659,10 +638,7 @@ async def run_bot_cycle_async():
         if not await state_manager.wait_for_initialization():
             print("❌ Services not fully initialized - skipping cycle"); return
 
-        # 🔴 --- START OF CHANGE (V6.9) --- 🔴
-        # (إضافة: تأخير بسيط لضمان تحرير القفل من الدورة السابقة)
         await asyncio.sleep(1.0) 
-        # 🔴 --- END OF CHANGE --- 🔴
         
         print("🔄 Starting Explorer cycle (Layer 1)...");
         await r2_service_global.save_system_logs_async({"explorer_cycle_started": True})