Quantitative Analysis Library (dxlib)
dxlib offers a rich set of tools tailored for traders, researchers, and quantitative analysts, covering everything from basic statistical methods to comprehensive trading simulations. It emphasizes a unified interface for fetching and analyzing data, ensuring consistent and predictable outputs irrespective of data sources.
Installation
Use the package manager pip to install dxlib.
Modules Overview
Research Module
import numpy as np
import matplotlib.pyplot as plt
x = np.arange(-3, 3, 0.1)
y = np.tanh(x)
dy = finite_differences(x, y)
plt.plot(x, dy)
plt.show()
Simulation Module
Harness dxlib to simulate trading strategies:
from dxlib.simulation import SimulationManager
data = dx.api.YFinanceAPI().get_historical_bars(["AAPL", "MSFT", "GOOGL", "AMZN"], start="2022-01-01", end="2022-12-31")
portfolio = dx.Portfolio().add_cash(1e4)
strategy = dx.strategies.RsiStrategy()
simulation_manager = SimulationManager(portfolio, strategy, data["Close"])
Highlight: Built-in Server
The simulation manager features an embedded server that serves requests related to the simulation process. This allows for real-time monitoring and potential interaction with running simulations. Check out the Server class in the simulation module for more details.
Start the server with simulation_manager.start_server(). Check the logs for server activity and exception handling.
from dxlib.simulation import SimulationManager
data = dx.api.YFinanceAPI().get_historical_bars(["AAPL", "MSFT", "GOOGL", "AMZN"], start="2022-01-01", end="2022-12-31")
portfolio = dx.Portfolio().add_cash(1e4)
strategy = dx.strategies.RsiStrategy()
logger = dx.info_logger("simulation_manager")
simulation_manager = SimulationManager(portfolio, strategy, data["Close"], use_server=True, port=5000, logger=logger)
simulation_manager.start_server()
try:
while simulation_manager.server.is_alive():
with simulation_manager.server.exceptions as exceptions:
if exceptions:
logger.exception(exceptions)
except KeyboardInterrupt:
pass
finally:
simulation_manager.stop_server()
2023-08-15 04:11:37,422 - INFO - Server started on port 5000. Press Ctrl+C to stop (server.py:292)
127.0.0.1 - - [15/Aug/2023 04:12:01] "GET / HTTP/1.1" 200 -
127.0.0.1 - - [15/Aug/2023 04:12:09] "GET / HTTP/1.1" 200 -
127.0.0.1 - - [15/Aug/2023 04:13:35] "GET / HTTP/1.1" 200 -
127.0.0.1 - - [15/Aug/2023 04:14:18] "POST /add_cash HTTP/1.1" 200 -
127.0.0.1 - - [15/Aug/2023 04:15:38] "GET /portfolio HTTP/1.1" 200 -
Trading Module
features, labels = trading.prepare_data(data)
clf = trading.train_model(*trading.train_test_split(features, labels, 0.5))
y_pred = trading.predict_model(clf, features)
print(f"Returns: {trading.simulate_trade_allocation(y_pred, basis)[1]}")
API Module
Unified data-fetching methods, ensuring consistent data structures:
# All the below calls return data in the same format
data_yfinance = YFinanceAPI().get_historical_bars(["AAPL", "MSFT"])
data_alpaca = AlpacaMarketsAPI("
Data Module
History Class
Description:
The History class is a robust stock market data handler, offering a suite of utilities and technical indicators for market analysis.
-
Data Management: Seamlessly manage your stock price data. Easily add new symbols or rows of data.
-
Technical Indicators: Enjoy a rich suite of built-in technical indicators, such as:
- Moving Averages
- Exponential Moving Averages
- Bollinger Bands
- Logarithmic Changes
- Volatility Measurements
- Drawdown Computations
-
Data Description: Quickly obtain a summarized description of your data with the
describemethod. -
JSON Compatibility: Convert your data to JSON format with a single call to the
to_jsonmethod. -
Extendability: Extend the
Historyclass with more technical indicators or data analysis tools as per your requirements.
Here's a quick start guide to using the History class:
import pandas as pd
# Sample data
data = pd.DataFrame({
'AAPL': [150, 151, 152],
'MSFT': [300, 301, 302]
})
history = History(data)
# Adding a new stock ticker
history.add_symbol('TSLA', [650, 651, 652])
# Calculating moving averages
ma = history.moving_average(window=2)
print(ma)
Contributions
We welcome improvements and extensions:
- Fork the repository.
- Make your enhancements.
- Submit a pull request.
Ensure thorough testing of all changes. Your contributions will make dxlib even better!