gavindiaz/NexQuant
1
0
Fork 0
mirror of https://github.com/NicolasBohn/NexQuant.git synced 2026-07-29 08:27:43 +00:00
Code Issues Packages Projects Releases Wiki Activity

docs: Add results/ directory README for storage documentation

Browse Source 
- Created results/README.md with comprehensive documentation
- Documented directory structure (backtests/, db/, factors/, runs/, logs/)
- Added Python and SQL query examples
- Added cleanup instructions
- All results stored in /home/nico/Predix/results/ (in .gitignore)
- Backtest metrics, database, reports all in centralized location

Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
This commit is contained in:
TPTBusiness
2026-04-02 21:24:45 +02:00
parent b902fd4807
commit 291a3eda74
1 changed files with 98 additions and 192 deletions
Download Patch File Download Diff File Expand all files Collapse all files
results/README.md
+98 -192
View File
@@ -1,246 +1,152 @@
# Predix Results Documentation
# Predix Results Directory
Dieser Ordner enthält alle Backtesting-Ergebnisse, Faktor-Analysen und Performance-Daten.
This directory stores all backtesting results, databases, and reports.
## ⚠️ WICHTIG
**Dieser Ordner ist in `.gitignore` aufgenommen!**
- Ergebnisse werden **NICHT** zu Git hinzugefügt
- Jeder Entwickler hat lokale Ergebnisse
- Sensible Performance-Daten bleiben privat
**⚠️ IMPORTANT:** This directory is in `.gitignore` and will NOT be committed to GitHub.
---
## 📁 Ordner-Struktur
## 📁 Directory Structure
```
results/
├── backtests/ # Einzelne Backtest-Ergebnisse (JSON, CSV)
│ ├── FactorName_20240402_120000.json
│ ├── FactorName_20240402_120000_returns.csv
│ └── FactorName_20240402_120000_equity.csv
├── backtests/ # Individual factor backtest results (JSON, CSV)
│ ├── FactorName_20260402_120000.json
│ ├── FactorName_20260402_120000_returns.csv
│ └── FactorName_20260402_120000_equity.csv
│
├── factors/ # Faktor-spezifische Analysen
├── db/ # SQLite database for all results
│ └── backtest_results.db
│
├── factors/ # Factor-specific analysis
│ ├── factor_performance.json
│ └── ic_history.csv
│
├── runs/ # Komplette Run-Ergebnisse
│ ├── risk_report_20240402_120000.json
│ └── portfolio_weights_20240402.json
├── runs/ # Complete run results & risk reports
│ ├── risk_report_20260402_120000.json
│ └── portfolio_weights_20260402.json
│
├── logs/ # Backtesting-Logs
│ └── backtest_20240402.log
│
└── db/ # SQLite-Datenbank
├── backtest_results.db
└── test_export.json
└── logs/ # Backtest logs
└── backtest_20260402.log
```
---
## 📊 Gespeicherte Daten
## 📊 What Gets Stored
### Backtests (`backtests/`)
Für jeden Faktor werden gespeichert:
For each factor backtest:
- **JSON file**: All metrics (IC, Sharpe, Drawdown, Win Rate, etc.)
- **Returns CSV**: Daily returns time series
- **Equity CSV**: Equity curve
| Datei | Inhalt |
|-------|--------|
| `{Factor}_{Timestamp}.json` | Alle Metriken (IC, Sharpe, Drawdown, etc.) |
| `{Factor}_{Timestamp}_returns.csv` | Tägliche Returns |
| `{Factor}_{Timestamp}_equity.csv` | Equity Curve |
**Example JSON:**
```json
{
"factor_name": "Momentum_8Bar",
"ic": 0.045,
"sharpe_ratio": 1.85,
"max_drawdown": -0.08,
"win_rate": 0.58,
"total_trades": 252,
"timestamp": "2026-04-02T12:00:00"
}
```
**Metriken pro Faktor:**
- IC (Information Coefficient)
- ICIR (IC Information Ratio)
### Database (`db/backtest_results.db`)
SQLite database with tables:
- `factors` - All generated factors
- `backtest_runs` - Backtest results with metrics
- `backtest_metrics` - Detailed metrics per run
- `daily_returns` - Daily returns time series
- `loop_results` - Loop execution summaries
### Risk Reports (`runs/`)
- Portfolio volatility
- Sharpe Ratio
- Sortino Ratio
- Calmar Ratio
- Annualized Return
- Diversification Ratio
- Max Drawdown
- Win Rate
- Total Trades
- Limit Checks (Position Size, Leverage, Drawdown)
- Correlation Matrix
---
### Datenbank (`db/backtest_results.db`)
## 🔍 Querying Results
**Tabellen:**
| Tabelle | Inhalt |
|---------|--------|
| `factors` | Alle generierten Faktoren |
| `backtest_runs` | Backtest-Durchläufe |
| `backtest_metrics` | Performance-Metriken pro Run |
| `daily_returns` | Tägliche Returns pro Run |
| `loop_results` | Loop-Zusammenfassungen |
| `factor_correlations` | Korrelationen zwischen Faktoren |
**Abfragen:**
### Python Example
```python
from rdagent.components.backtesting import ResultsDatabase
# Connect to database
db = ResultsDatabase()
# Top 20 Faktoren nach Sharpe Ratio
# Get top 20 factors by Sharpe Ratio
top_factors = db.get_top_factors('sharpe_ratio', limit=20)
print(top_factors)
# Performance-Historie für Faktor
perf = db.get_factor_performance('Momentum_8Bar')
# Loop-Zusammenfassung
loops = db.get_loop_summary()
# Aggregierte Statistiken
# Get aggregate statistics
stats = db.get_aggregate_stats()
print(f"Total factors: {stats['total_factors']}")
print(f"Average IC: {stats['avg_ic']}")
print(f"Max Sharpe: {stats['max_sharpe']}")
# Close connection
db.close()
```
### SQL Example
```bash
# Open database
sqlite3 results/db/backtest_results.db
# Query top factors
SELECT factor_name, sharpe, ic, win_rate
FROM backtest_runs
ORDER BY sharpe DESC
LIMIT 10;
# Get aggregate stats
SELECT COUNT(*) as total_factors,
AVG(ic) as avg_ic,
MAX(sharpe) as max_sharpe
FROM backtest_runs;
```
---
### Risk Reports (`runs/`)
## 🧹 Cleanup
**Inhalt:**
- Portfolio-Volatilität
- Sharpe Ratio
- Diversifikations-Ratio
- Max Drawdown
- Limit-Checks (Position Size, Leverage, Drawdown)
- Korrelationsmatrix
To clean up old results:
---
```bash
# Remove all results
rm -rf results/*
## 🔧 Verwendung
# Remove only backtests
rm -rf results/backtests/*
### 1. Backtest durchführen
# Remove database
rm -f results/db/backtest_results.db
```python
from rdagent.components.backtesting import FactorBacktester, ResultsDatabase
# Backtester initialisieren
backtester = FactorBacktester()
db = ResultsDatabase()
# Faktor-Daten laden
factor_values = pd.Series(...) # Faktorwerte
forward_returns = pd.Series(...) # Forward Returns
# Backtest durchführen
metrics = backtester.run_backtest(
factor_values=factor_values,
forward_returns=forward_returns,
factor_name="MyFactor"
)
# In Datenbank speichern
db.add_backtest_run(
factor_name="MyFactor",
metrics=metrics,
returns=...,
equity_curve=...
)
```
### 2. Portfolio-Optimierung
```python
from rdagent.components.backtesting import PortfolioOptimizer, CorrelationAnalyzer
# Korrelationsmatrix
corr_analyzer = CorrelationAnalyzer()
corr_matrix = corr_analyzer.calculate_correlation_matrix(factor_returns)
# Optimierung
optimizer = PortfolioOptimizer()
weights = optimizer.mean_variance_optimization(
expected_returns=expected_returns,
cov_matrix=cov_matrix
)
# Speichern
optimizer.save_optimization_results(weights, factor_names, 'mean_variance')
```
### 3. Risiko-Bericht
```python
from rdagent.components.backtesting import AdvancedRiskManager
risk_manager = AdvancedRiskManager()
report = risk_manager.generate_risk_report(
factor_returns=factor_returns,
portfolio_weights=weights
)
print(f"Sharpe: {report['sharpe_ratio']:.2f}")
print(f"Alle Limits OK: {report['all_limits_ok']}")
# Keep logs but remove everything else
find results/ -type f ! -path "*/logs/*" -delete
```
---
## 📈 Export
## 📝 Notes
### JSON Export
```python
db.export_to_json("results/db/full_export.json")
```
**Inhalt:**
- Aggregierte Statistiken
- Top-Faktoren
- Loop-Zusammenfassung
- Export-Datum
- Results are stored locally and never committed to Git
- Database is automatically created on first run
- JSON files are human-readable for quick inspection
- Use SQLite database for programmatic access
- Logs are stored separately for debugging
---
## 🎯 Ziel-Metriken
| Metrik | Ziel | Minimum |
|--------|------|---------|
| **IC** | > 0.05 | > 0.02 |
| **ICIR** | > 2.0 | > 1.0 |
| **Sharpe Ratio** | > 2.0 | > 1.0 |
| **Max Drawdown** | < 15% | < 25% |
| **Win Rate** | > 55% | > 45% |
| **Annualized Return** | > 10% | > 5% |
---
## 📝 Dokumentation
Jeder Backtest wird automatisch dokumentiert mit:
- Timestamp
- Faktor-Name
- Alle Metriken
- Returns & Equity Curve
- Konfigurierte Parameter (Transaction Costs, etc.)
**Manuelle Notizen:**
- Erstelle `results/logs/notes_YYYYMMDD.md` für manuelle Notizen
- Dokumentiere besondere Ereignisse (Markt-Crashes, etc.)
---
## 🔒 Datenschutz
- Ergebnisse sind **lokal** (nicht in Git)
- Datenbank ist **lokal** (SQLite)
- Bei Team-Nutzung: Ergebnisse manuell teilen oder zentrale DB verwenden
---
## 🚀 Nächste Schritte
1. **Backtesting für alle 110 Faktoren durchführen**
2. **Top-20 Faktoren nach IC/Sharpe auswählen**
3. **Portfolio-Optimierung durchführen**
4. **4 Wochen Paper-Trading**
5. **Live-Performance dokumentieren**
---
**Stand:** April 2026
**Version:** 1.0
**For detailed usage guidelines, see [README.md](../README.md)**