Add Deep Learning (LSTM) + Quick Wins automation features

Major improvements:
- Deep Learning integration with PyTorch LSTM (dual models: main 1-50 + euro 1-12)
- Hybrid predictor: 40% RandomForest + 60% Deep Learning
- LaunchAgent for automatic weekly tip generation (Mon/Thu 21:00)
- Health-Check system with auto-recovery and Telegram alerts
- Fixed health checks for Eurojackpot-specific paths and file names
- Model caching and intelligent retraining logic
- Updated CSV data and generated tips
- Performance reports for recent draws

Technical details:
- PyTorch used instead of TensorFlow (Python 3.14 compatibility)
- Separate LSTM models for main numbers (1-50) and euro numbers (1-12)
- Apple Silicon MPS acceleration support
- Sequence learning with 20-draw history
- Health-check adapted for eurojackpot_ml_models/ and learning_log.json

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

scripts/automation/auto_update_and_learn.py

@@ -81,7 +81,9 @@ class AutoUpdateAndLearn:
             df = pd.read_csv(self.data_file, sep=';')
             df['datum'] = pd.to_datetime(df['datum'], format='%Y-%m-%d')
 
-            latest_draw = df.iloc[0]  # Neueste Ziehung (sortiert absteigend)
+            # Sortiere nach Datum absteigend und nimm neueste Ziehung
+            df = df.sort_values('datum', ascending=False)
+            latest_draw = df.iloc[0]  # Neueste Ziehung
             latest_date = latest_draw['datum']
 
             print(f"   📅 Neueste Ziehung in Daten: {latest_date.strftime('%Y-%m-%d')}")
@@ -148,7 +150,7 @@ class AutoUpdateAndLearn:
         print("\n🎯 EVALUIERE LETZTE TIPPS")
         print("=" * 70)
 
-        # Finde neueste Tipps-Datei
+        # Finde Tipps-Datei die VOR der Ziehung generiert wurde
         if not os.path.exists(self.tips_dir):
             print("   ⚠️  Keine Tipps zum Evaluieren")
             return {}
@@ -162,8 +164,30 @@ class AutoUpdateAndLearn:
             print("   ⚠️  Keine Tipps-Dateien gefunden")
             return {}
 
-        latest_tips_file = os.path.join(self.tips_dir, tip_files[0])
-        print(f"   📁 Evaluiere: {tip_files[0]}")
+        # Finde Tip-Datei die vor der Ziehung erstellt wurde
+        draw_date = new_draw['date']
+        selected_tip_file = None
+
+        for tip_file in tip_files:
+            # Parse Timestamp aus Dateiname: weekly_tips_YYYYMMDD_HHMMSS.csv
+            try:
+                parts = tip_file.replace('.csv', '').split('_')
+                tip_date_str = parts[-2]  # YYYYMMDD
+                tip_date = pd.to_datetime(tip_date_str, format='%Y%m%d')
+
+                # Nehme erste Datei die vor der Ziehung war
+                if tip_date < draw_date:
+                    selected_tip_file = tip_file
+                    break
+            except:
+                continue
+
+        if not selected_tip_file:
+            # Fallback: nehme älteste Datei
+            selected_tip_file = tip_files[-1]
+
+        latest_tips_file = os.path.join(self.tips_dir, selected_tip_file)
+        print(f"   📁 Evaluiere: {selected_tip_file}")
 
         try:
             tips_df = pd.read_csv(latest_tips_file)