Spaces:
Configuration error
Configuration error
Niranjan-Murugarasu024
feat: TradeSense AI Chairman Edition production release ready for Hugging Face Spaces
0125724 | """ | |
| TradeSense AI — Grounded AI Research Copilot & Evidence Synthesizer | |
| Generates institutional research commentary strictly synthesized from StructuredEvidence payloads. | |
| Zero hallucination / zero financial fabrication guarantee. | |
| """ | |
| from typing import Dict, Any | |
| from backend.app.schemas.contracts import StructuredEvidence, SignalType | |
| class GroundedCopilotEngine: | |
| """ | |
| Synthesizes a structured, natural-language equity research note strictly from a StructuredEvidence object. | |
| Never invents unprovided metrics or unsupported price targets. | |
| """ | |
| def generate_research_note(cls, evidence: StructuredEvidence) -> str: | |
| """ | |
| Generates deterministic, evidence-grounded research commentary. | |
| """ | |
| forecast = evidence.forecast | |
| backtest = evidence.backtest | |
| risk = evidence.risk | |
| # Auto-detect Indian instrument vs US | |
| from backend.data.providers.yfinance import YFinanceProvider | |
| is_in = YFinanceProvider.is_indian_symbol(evidence.ticker) | |
| c = "₹" if is_in else "$" | |
| # Section 1: Executive Stance & Quantitative Signal | |
| stance_header = f"### 📊 TradeSense Quantitative Research Note: {evidence.ticker} (Horizon: {evidence.horizon_sessions} Sessions)\n\n" | |
| stance_header += f"**Quantitative Signal**: `{evidence.signal.value}` · **Factor Concurrence**: {evidence.factors_positive}/{evidence.factors_total} Positive Factors\n\n" | |
| stance_header += f"**Core Assessment**: {evidence.signal_reason}\n\n" | |
| # Section 2: Probabilistic Forecast Distribution | |
| fc_text = "#### 🔮 Forecast Distribution & Probability Bounds\n" | |
| fc_text += f"- **Current Reference Price**: `{c}{forecast.reference_price:,.2f}` (as of {evidence.as_of_session.strftime('%Y-%m-%d')})\n" | |
| fc_text += f"- **Expected Horizon Value**: `{c}{forecast.expected_price:,.2f}` ({forecast.expected_return_pct:+.2f}%)\n" | |
| fc_text += f"- **80% Prediction Interval**: `{c}{forecast.interval_80[0]:,.2f}` – `{c}{forecast.interval_80[1]:,.2f}`\n" | |
| fc_text += f"- **95% Prediction Interval**: `{c}{forecast.interval_95[0]:,.2f}` – `{c}{forecast.interval_95[1]:,.2f}`\n" | |
| fc_text += f"- **Active Model**: {forecast.active_model_id}\n\n" | |
| # Section 3: Walk-Forward Out-of-Sample Empirical Evidence | |
| bt_text = "#### 📈 Walk-Forward Backtesting & Baseline Edge Matrix\n" | |
| bt_text += f"- **Out-of-Sample RMSE**: `{c}{backtest.oos_rmse:,.2f}` (vs. Naive Baseline RMSE `{c}{backtest.baseline_naive_rmse:,.2f}`)\n" | |
| bt_text += f"- **Baseline Improvement**: **{backtest.baseline_improvement_rmse_pct:+.1f}%** reduction in forecast error over Naive persistence\n" | |
| bt_text += f"- **Directional Accuracy**: **{backtest.directional_accuracy_pct:.1f}%** over expanding test windows\n" | |
| bt_text += f"- **Friction-Adjusted Net Sharpe**: **{backtest.net_sharpe:.2f}** (net of execution friction & statutory drag)\n" | |
| bt_text += f"- **Empirical 80% Coverage**: {backtest.empirical_coverage_80_pct:.1f}% (Calibration error: {backtest.calibration_error_80_pct:.1f}%)\n\n" | |
| # Section 4: Quantitative Risk & Geometry | |
| risk_text = "#### 🛡️ Quantitative Risk & Volatility Sizing\n" | |
| risk_text += f"- **14-Session ATR**: `{c}{risk.atr_14:,.2f}` ({risk.atr_pct:.2f}% of price) · Regime: {risk.volatility_regime.replace('_', ' ').capitalize()} Volatility\n" | |
| risk_text += f"- **Key Support / Resistance**: Support at `{c}{risk.support_level:,.2f}` · Resistance at `{c}{risk.resistance_level:,.2f}`\n" | |
| risk_text += f"- **ATR Suggested Stop-Loss**: `{c}{risk.suggested_stop_loss:,.2f}`\n" | |
| risk_text += f"- **Suggested Targets**: Target 1 at `{c}{risk.suggested_take_profit_1:,.2f}` (1:2 R:R) · Target 2 at `{c}{risk.suggested_take_profit_2:,.2f}` (1:3 R:R)\n" | |
| risk_text += f"- **Risk-Reward Score**: **{risk.risk_reward_ratio:.2f}:1**\n\n" | |
| # Section 5: Cryptographic Lineage & Disclaimer | |
| disclaimer_text = "---\n" | |
| disclaimer_text += f"*SHA-256 Forecast Audit Hash*: `{forecast.sha256_audit_hash}`\n\n" | |
| disclaimer_text += "> **Disclaimer**: TradeSense AI is a quantitative research platform providing algorithmic factor analysis. It does not guarantee market returns or provide individualized investment advice." | |
| return stance_header + fc_text + bt_text + risk_text + disclaimer_text | |