Dashboard Parallel Lists Pattern

Dashboard Parallel Lists Pattern

Experiment Overview

Context

The dashboard crashed with:

[ERROR] __main__: Dashboard error: 'list' object has no attribute 'get'

Root Cause: Code assumed symbol_signals was a dict keyed by symbol:

# WRONG assumption
for symbol, data in symbol_signals.items():
    signal = data.get('signal', 0)
    gate_status = data.get('gate_status', {})

Actual Structure: get_symbol_signals() returns parallel lists:

{
    'symbols': ['AAPL', 'MSFT', 'GOOGL'],
    'signal_values': [1, -1, 0],
    'confidences': [0.72, 0.65, 0.45],
    'gate_statuses': [
        {'final_status': 'READY', ...},
        {'final_status': 'BLOCKED', 'blocking_gate': 'crypto_short'},
        {'final_status': 'HOLD', ...}
    ]
}

Verified Workflow

Correct Pattern: Iterate by Index

def get_model_metrics(symbol_signals: Dict) -> dict:
    """Process symbol signals using parallel list structure."""
    metrics = {
        'total_symbols': 0,
        'ready_count': 0,
        'blocked_count': 0,
        'hold_count': 0,
        'avg_confidence': 0.0,
        'signals': {'buy': 0, 'sell': 0, 'hold': 0}
    }

    if not symbol_signals:
        return metrics

    # Extract parallel lists
    symbols = symbol_signals.get('symbols', [])
    signal_values = symbol_signals.get('signal_values', [])
    confidences = symbol_signals.get('confidences', [])
    gate_statuses = symbol_signals.get('gate_statuses', [])

    if not symbols:
        return metrics

    metrics['total_symbols'] = len(symbols)
    valid_confidences = []

    # Iterate by index across parallel lists
    for i, sym in enumerate(symbols):
        # Safe access with bounds checking
        gate_status = gate_statuses[i] if i < len(gate_statuses) else {}
        signal = signal_values[i] if i < len(signal_values) else 0
        conf = confidences[i] if i < len(confidences) else 0

        status = gate_status.get('final_status', 'HOLD')

        if status == 'READY':
            metrics['ready_count'] += 1
        elif status == 'BLOCKED':
            metrics['blocked_count'] += 1
        else:
            metrics['hold_count'] += 1

        if signal > 0:
            metrics['signals']['buy'] += 1
        elif signal < 0:
            metrics['signals']['sell'] += 1
        else:
            metrics['signals']['hold'] += 1

        if conf > 0:
            valid_confidences.append(conf)

    if valid_confidences:
        metrics['avg_confidence'] = sum(valid_confidences) / len(valid_confidences)

    return metrics

Why Parallel Lists?

The get_symbol_signals() function returns parallel lists because:

1. Performance: Lists are faster to append during signal generation loop
2. Order preservation: Maintains processing order for display
3. Flexible lengths: Different lists can have different lengths if some data missing
4. JSON serialization: Easy to serialize for dashboard communication

Accessing Individual Symbol Data

# Get data for a specific symbol by finding its index
def get_symbol_data(symbol_signals: Dict, target_symbol: str) -> dict:
    symbols = symbol_signals.get('symbols', [])
    try:
        idx = symbols.index(target_symbol)
        return {
            'symbol': target_symbol,
            'signal': symbol_signals.get('signal_values', [])[idx],
            'confidence': symbol_signals.get('confidences', [])[idx],
            'gate_status': symbol_signals.get('gate_statuses', [])[idx],
        }
    except (ValueError, IndexError):
        return None

Failed Attempts (Critical)

Key Insights

Signal Structure from get_symbol_signals()

# In live_trader.py get_symbol_signals() returns:
{
    'symbols': List[str],           # Symbol names
    'signal_values': List[int],     # -1 (sell), 0 (hold), 1 (buy)
    'confidences': List[float],     # Model confidence [0, 1]
    'gate_statuses': List[Dict],    # Gate check results
    'prices': List[float],          # Current prices
    'regime_contexts': List[Dict],  # Markov regime info
}

Gate Status Structure

# Each gate_status dict contains:
{
    'final_status': 'READY' | 'BLOCKED' | 'HOLD',
    'blocking_gate': str | None,    # Which gate blocked (if any)
    'confidence_gate': bool,
    'crypto_short_gate': bool,
    'portfolio_limit_gate': bool,
    'capital_manager_gate': bool,
    'portfolio_risk_gate': bool,
}

Common Error Patterns

Files Modified

scripts/​monitor_dashboard.py:
  - Lines 536-577: get_model_metrics() rewritten for parallel lists
  - Pattern: Extract lists first, iterate by index with bounds checking

References

• scripts/​live_trader.py: get_symbol_signals() function (lines ~800-900)
• scripts/​monitor_dashboard.py: get_model_metrics() function (lines ~536-577)
• alpaca_trading/​visualization/​dashboard.py: Dashboard display logic