AI Hedge Fund - Core Signals System

🎯 Overview

A real-time market intelligence system that detects institutional breakouts and reversals using dark pool (DP) data confirmation. The system prioritizes institutional flow validation over retail signals to avoid traps and identify high-probability moves.

🏗️ Architecture

Core Components

core/
├── core_signals_runner.py          # Main orchestration system
├── data/                           # Data connectors
│   ├── chartexchange_api_client.py # ChartExchange DP data API
│   ├── real_breakout_reversal_detector_yahoo_direct.py # Yahoo Direct scraper
│   └── production_rate_limit_solver.py # Rate limiting utilities
├── detectors/                      # Signal detection
│   ├── live_high_resolution_detector.py
│   └── real_breakout_reversal_detector_yahoo_direct.py
└── filters/                        # Signal filtering
    └── dp_aware_signal_filter.py  # DP-aware signal validation

Data Flow

1. Price Data: Yahoo Direct API (no rate limits)
2. DP Data: ChartExchange API (Tier 3, 1000 req/min)
3. Signal Detection: Breakout/reversal detection with rolling windows
4. DP Filtering: Institutional confirmation required
5. Output: Timestamped signals with confidence scores

🚀 Quick Start

Prerequisites

pip install requests numpy yfinance

Configuration

Create configs/chartexchange_config.py:

CHARTEXCHANGE_API_KEY = "your_api_key_here"
CHARTEXCHANGE_TIER = 3

Run Core System

# Test the system
python3 tests/test_core_system.py

# Run live signals (RTH only)
python3 core/core_signals_runner.py

📊 APIs & Data Sources

1. ChartExchange API (Primary DP Data)

Endpoints:

• /data/dark-pool-levels/ - DP support/resistance levels
• /data/dark-pool-prints/ - Real-time DP transactions

Usage:

from core.data.chartexchange_api_client import ChartExchangeAPI
from configs.chartexchange_config import CHARTEXCHANGE_API_KEY

api = ChartExchangeAPI(CHARTEXCHANGE_API_KEY, tier=3)
dp_levels = api.get_dark_pool_levels('SPY', days_back=1)
dp_prints = api.get_dark_pool_prints('SPY', days_back=1)

Rate Limits:

• Tier 1: 60 req/min
• Tier 2: 250 req/min
• Tier 3: 1000 req/min

2. Yahoo Direct API (Price Data)

Features:

• No rate limits
• Real-time quotes
• Historical data
• Options chains

Usage:

from core.detectors.real_breakout_reversal_detector_yahoo_direct import YahooDirectDataProvider

provider = YahooDirectDataProvider()
data = provider.get_market_data('SPY')

3. DP-Aware Signal Filter

Purpose: Only trigger signals when institutional flow confirms the move

Key Features:

• DP level proximity detection
• Institutional battleground avoidance
• Risk level calculation
• Signal tightening based on DP agreement

Usage:

from core.filters.dp_aware_signal_filter import DPAwareSignalFilter

filter = DPAwareSignalFilter()
signals = await filter.filter_signals_with_dp_confirmation('SPY')

🔍 Signal Detection Logic

Breakout Detection

• Rolling Windows: 5, 10, 20, 30 minutes
• Thresholds: 0.2% - 1.5% price movement
• Volume Confirmation: Above average volume
• DP Confirmation: Price above DP resistance levels

Reversal Detection

• Support Levels: DP support level proximity
• Mean Reversion: Price below DP support
• Volume Confirmation: Above average volume
• DP Confirmation: Price near DP support levels

DP Structure Analysis

• Support Levels: Price levels with high DP volume below current price
• Resistance Levels: Price levels with high DP volume above current price
• Battlegrounds: High-volume DP levels (>1M shares or >10K contracts)
• Strength Score: Normalized DP volume (0-1 scale)

📈 Signal Output Format

{
  "ticker": "SPY",
  "action": "BUY",
  "entry_price": 664.39,
  "confidence": 0.75,
  "risk_level": "HIGH",
  "dp_agreement": 0.30,
  "breakout": false,
  "mean_reversion": true,
  "dp_factors": ["near_support"],
  "timestamp": "2025-10-17T23:00:12Z"
}

🎛️ Configuration Parameters

DP Filter Settings

# DP structure thresholds
dp_support_threshold = 0.7      # 70% of volume at support
dp_resistance_threshold = 0.7    # 70% of volume at resistance  
battleground_threshold = 0.8     # 80% institutional ratio = battleground
breakout_confirmation_threshold = 0.25  # 25% above magnet for breakout
mean_reversion_threshold = 0.15  # 15% below DP support for mean reversion

# Signal tightening parameters
min_dp_agreement = 0.3          # Minimum DP agreement (lowered for testing)
min_composite_confidence = 0.75 # 75% composite signal confidence
max_risk_level = "MEDIUM"       # Maximum risk level allowed

Detection Windows

# Rolling window sizes (minutes)
windows = [5, 10, 20, 30]

# Breakout thresholds (%)
breakout_thresholds = [0.2, 0.5, 1.0, 1.5]

# Volume multipliers
volume_multipliers = [1.5, 2.0, 2.5, 3.0]

🚨 Risk Management

Risk Levels

• LOW: Strong DP confirmation, low volatility
• MEDIUM: Moderate DP confirmation, normal volatility
• HIGH: Weak DP confirmation, high volatility

DP Agreement Factors

• near_support: Price within 3% of DP support level
• near_resistance: Price within 3% of DP resistance level
• battleground: Price near institutional battleground
• volume_confirmation: Above-average volume
• momentum_alignment: Price momentum aligns with DP structure

📊 Monitoring & Logging

Log Files

• logs/core_signals.csv - Signal history
• logs/core_signals.jsonl - Detailed signal logs

Key Metrics

• Signal frequency
• DP agreement scores
• Risk level distribution
• Win/loss ratios
• Drawdown periods

🔧 Troubleshooting

Common Issues

1. No DP Data: Check ChartExchange API key and tier
2. Rate Limits: Verify API tier limits
3. No Signals: Check DP agreement thresholds
4. Import Errors: Verify folder structure and paths

Debug Mode

import logging
logging.basicConfig(level=logging.DEBUG)

📚 Examples

Basic Signal Detection

import asyncio
from core.filters.dp_aware_signal_filter import DPAwareSignalFilter

async def detect_signals():
    filter = DPAwareSignalFilter()
    signals = await filter.filter_signals_with_dp_confirmation('SPY')
    
    for signal in signals:
        print(f"{signal.action} {signal.ticker} @ ${signal.entry_price:.2f}")
        print(f"Risk: {signal.risk_level}, DP Agreement: {signal.dp_agreement:.2f}")

asyncio.run(detect_signals())

DP Structure Analysis

import asyncio
from core.filters.dp_aware_signal_filter import DPAwareSignalFilter

async def analyze_dp_structure():
    filter = DPAwareSignalFilter()
    structure = await filter.analyze_dp_structure('SPY')
    
    print(f"Support Levels: {len(structure.dp_support_levels)}")
    print(f"Resistance Levels: {len(structure.dp_resistance_levels)}")
    print(f"Battlegrounds: {len(structure.institutional_battlegrounds)}")
    print(f"DP Strength: {structure.dp_strength_score:.2f}")

asyncio.run(analyze_dp_structure())

🎯 Performance

Current Status

• DP Data: ✅ Real ChartExchange integration
• Price Data: ✅ Yahoo Direct (no rate limits)
• Signal Detection: ✅ Breakout/reversal logic
• DP Filtering: ✅ Institutional confirmation
• Risk Management: ✅ Multi-level risk assessment

Test Results

✅ DP structure analyzed successfully
   Support levels: 56
   Resistance levels: 99  
   Battlegrounds: 5
   DP strength: 1.00

✅ Signals detected!
  1. BUY @ $664.39 - Risk: HIGH

🔮 Roadmap

Phase 1: Core System ✅

• DP data integration
• Signal detection
• DP filtering
• Risk management

Phase 2: Enhancement

• Multi-timeframe analysis
• Sector rotation detection
• Options flow integration
• News sentiment analysis

Phase 3: Production

• Live trading integration
• Portfolio management
• Performance analytics
• Alert systems

📞 Support

For issues or questions:

1. Check the troubleshooting section
2. Review log files for errors
3. Verify API configurations
4. Test with tests/test_core_system.py

---

Built for institutional-grade market intelligence 🚀