29.05.2016       Выпуск 127 (23.05.2016 - 29.05.2016)       Интересные проекты, инструменты, библиотеки

PyThalesians - финансовая библиотека (тестирование стратегий для рынка)


Экспериментальная функция:

Ниже вы видите текст статьи по ссылке. По нему можно быстро понять ссылка достойна прочтения или нет

Просим обратить внимание, что текст по ссылке и здесь может не совпадать.


PyThalesians is a Python financial library developed by the Thalesians (http://www.thalesians.com). I have used the library to develop my own trading strategies and I've included simple samples which show some of the functionality including an FX trend following model and other bits of financial analysis.

There are many open source Python libraries for making trading strategies around! However, I've developed this one to be as flexible as possible in terms of what types of strategies you can develop with it. In addition, a lot of the library can be used to analyse and plot financial data for broader based analysis, of the type that I've had to face being in markets over the years. Hence, it can be used by a wider array of users.

At present the PyThalesians offers:

  • Backtesting of systematic trading strategies for cash markets (including cross sectional style trading strategies)
  • Sensitivity analysis for systematic trading strategies parameters
  • Seamless historic data downloading from Bloomberg (requires licence), Yahoo, Quandl, Dukascopy and other market data sources
  • Produces beautiful line plots with PyThalesians wrapper (via Matplotlib), Plotly (via cufflinks) and a simple wrapper for Bokeh
  • Analyse seasonality analysis of markets
  • Calculates some technical indicators and gives trading signals based on these
  • Helper functions built on top of Pandas
  • Automatic tweeting of charts
  • And much more!
  • Please bear in mind at present PyThalesians is currently a highly experimental alpha project and isn't yet fully documented
  • Uses Apache 2.0 licence


Below we give some examples of analysis we've done with PyThalesians. Some of these can be run by scripts in the examples folder.

Using PyThalesians to create a simple FX trend following strategy (you can run this backtest using cashbacktest_examples.py)

Using PyThalesians to plot & calculate USD/JPY intraday moves around non-farm payrolls over past 10 years

Using PyThalesians to calculate intraday vol in major FX crosses by time of day

Using PyThalesians to create the Thalesians CTA index (trend following), which replicates Newedge CTA index benchmark

Using PyThalesians with Cufflinks (Plotly wrapper) to plot interactive Plotly chart (using plotly_examples.py) - click the below to get to the interactive chart

Using PyThalesians to plot via Bokeh EUR/USD in the 3 hours following FOMC statements

Using PyThalesians to plot combination of bar/line/scatter for recent equity returns (you can run this analysis using bokeh_examples.py)

Using PyThalesians and PyFolio to plot return statistics of FX CTA strategy (you can run this analysis using strategyfxcta_example.py)

Using PyThalesians to plot with Plotly map of USA unemployment rate by state (using FRED data) (you can run this analysis using histecondata_examples.py)

Using PyThalesians to plot G10 CPI YoY rates (using FRED data) (you can run this analysis using histecondata_examples.py)

Using PyThalesians to plot rolling correlatons in FX (using Bloomberg data) (you can run this analysis using correlation_examples.py)

Using PyThalesians to plot seconds data around last NFP (using Bloomberg data) (you can run this analysis using tick_examples.py)

Using PyThalesians to plot AUD/USD total returns from spot & deposit data (comparing with spot and Bloomberg generated total return index) (you can run this analysis using indicesfx_examples.py)


PyThalesians has been tested on Windows 8 & 10, running Bloomberg Terminal software. I currently run PyThalesians using Anaconda 2.5 (Python 3.5 64bit) on Windows 10. Potentially, it could also work on the Bloomberg Server API (but I have not explicitly tested this). I have also tried running it on Ubuntu and Mac OS X (excluding Bloomberg API)

Major requirements

  • Required: Python 3.4, 3.5
  • Required: pandas, matplotlib, numpy etc.
  • Recommended: Bloomberg Python Open API
    • To use Bloomberg you will need to have a licence
    • Use experimental Python 3.4 version from Bloomberg http://www.bloomberglabs.com/api/libraries/
    • Also download C++ version of Bloomberg API and extract into any location
    • For Python 3.5 - need to compile blpapi source using Microsoft Visual Studio 2015 yourself
      • Install Microsoft Visual Studio 2015 (Community Edition is free)
      • Before doing do be sure to add environment variables for the Bloomberg DLL (blpapi3_64.dll) to PATH variable
      • Make sure BLPAPI_ROOT root is set as an environmental variable in Windows
      • python setup.py build
      • python setup.py install
    • For Python 3.4 - prebuilt executable can be run, which means we can skip the build steps above
      • Might need to tweak registry to avoid "Python 3.4 not found in registry error" (blppython.reg example) when using this executable
    • Alternatively to access Bloomberg, the software also supports the old COM API (but I'm going to remove it because very slow)
  • Recommended: Plotly for funky interactive plots (https://github.com/plotly/python-api) and
  • Recommended: Cufflinks a nice Plotly wrapper when using Pandas dataframes (Jorge Santos project now supports Python 3 https://github.com/jorgesantos/cufflinks - so I recommend using that rather than my fork)
  • Recommended: PyFolio for statistical analysis of trading strategy returns (https://github.com/quantopian/pyfolio/)
  • Recommended: multiprocessor_on_dill because standard multiprocessing library pickle causes issues (from https://github.com/sixty-north/multiprocessing_on_dill)


Once installed please make sure you edit pythalesians.util.constants file for the following variables:

  • Change the root path variable - this will ensure that the logging (and a number of other features works correctly). Failure to do so will result in the project not starting
  • Change the default Bloomberg settings (Which API to use? What server address to use?)
  • Write in API keys for Quandl, Twitter, Plotly etc.
  • Latest version can be installed using setup.py or pip (see below)
pip install git+https://github.com/thalesians/pythalesians.git

Examples for PyThalesians

After installation, the easiest way to get started is by looking at the example scripts. I am hoping to add some Jupyter notebooks, illustrating how to use the library too. The example scripts show how to:

  • Download market data from many different sources, Bloomberg, Yahoo, Quandl, Dukascopy etc
  • Plot line charts, with different styles

About the Thalesians

The Thalesians are a think tank of dedicated professionals with an interest in quantitative finance, economics, mathematics, physics and computer science, not necessarily in that order. We run quant finance events in London, New York, Budapest, Prague and Frankfurt (join our Meetup.com group at http://events.thalesians.com). We also publish research on systematic trading and also consult in the area. One of our clients is RavenPack, a major news analytics vendor.

Major contributors to PyThalesians

  • Saeed Amen - Saeed is managing director and co-founder of the Thalesians. He has a decade of experience creating and successfully running systematic trading models at Lehman Brothers and Nomura. Independently, he runs a systematic trading model with proprietary capital. He is the author of Trading Thalesians – What the ancient world can teach us about trading today (Palgrave Macmillan). He graduated with a first class honours master’s degree from Imperial College in Mathematics and Computer Science.

Supporting PyThalesians project

If you find PyThalesians useful (and in particular if you are commercial company) please consider supporting the project through sponsorship or by using our consultancy/research services in systematic trading. If you would like to contribute to the project, also let me know: it's a big task to try to build up this library on my own!

For the UK election Plot.ly code - please visit https://github.com/plotly/IPython-plotly/tree/master/notebooks/ukelectionbbg

Future Plans for PyThalesians

We plan to add the following features:

  • Have a proper setup mechanism (eg. via pip), at present needs (partial) manual deployment
  • Add Plotly & Seaborn wrappers for plotting (partially there)
  • Improve support for Bokeh plotting (partially)
  • Add more plots from Matlibplot
  • Add Reuters as a historic data source
  • Add ability to stream data from Bloomberg and Reuters
  • Use event driven code to generate trading signals (to be used live and historically)
  • Add more interesting trading analysis tools
  • Add support for live trading via Interactive Brokers
  • Integrate support for zipline as an alternative trading system
  • Improve support for PyFolio
  • Support Python 2.7+

More generally, we want to:

  • Make existing code more robust
  • Increase documentation and examples

Release Notes

  • 0.1a (highly experimental alpha version) - 01 Jul 2015
  • Basic implementation of plotting for line charts
  • Basic downloading of market data like Bloomberg/Yahoo etc. via generic wrapper

Coding log

  • 27 May 2016 - Added ability to plot strategy signal at point in time
  • 19 May 2016 - Updated Quandl wrapper to use new Quandl API
  • 02 May 2016 - Tidied up BacktestRequest, added SPX seasonality example
  • 28 Apr 2016 - Updated cashbacktest (for Pandas 0.18)
  • 21 Apr 2016 - Got rid of deprecated Pandas methods in EventStudy
  • 18 Apr 2016 - Fixed some incompatibility issues with Pandas 0.18
  • 06 Apr 2016 - Added more trade statistics output
  • 01 Apr 2016 - Speeded up joining operations, noticeable when fetching high freq time series
  • 21 Mar 2016 - Added IPython notebook to demonstrate how to backtest simple FX trend following trading strategy
  • 19 Mar 2016 - Tested with Python 3.5 64 bit (Anaconda 2.5 on Windows 10)
  • 17 Mar 2016 - Refactored some of graph/time series functions and StrategyTemplate
  • 11 Mar 2016 - Fixed warnings in matplotlib 1.5
  • 09 Mar 2016 - Added more TradeAnalysis features (for sensitivity analysis of trading strategies)
  • 01 Mar 2016 - Added IPython notebook to demonstrate how to download market data and plot
  • 27 Feb 2016 - Fixed total returns FX example
  • 20 Feb 2016 - Added more parameters for StrategyTemplate
  • 13 Feb 2016 - Edited time series filter methods
  • 11 Feb 2016 - Added example to plot BoJ interventions against USDJPY spot
  • 10 Feb 2016 - Updated project description
  • 01 Feb 2016 - Added LightEventsFactory to make it easier to deal with econ data events (stored as HDF5 files)
  • 20 Jan 2016 - Added kurtosis measure for trading strategy results, fixed Quandl issue
  • 19 Jan 2016 - Changed examples folder name
  • 15 Jan 2016 - Added risk on/off FX correlation example
  • 05 Jan 2016 - Added total return (spot) indices construction for FX and example
  • 26 Dec 2015 - Fixed problem with econ data downloaders
  • 24 Dec 2015 - Added datafactory templates for creating custom indicators
  • 19 Dec 2015 - Refactored Dukascopy downloader
  • 10 Dec 2015 - Various bug fixes
  • 22 Nov 2015 - Increased vol targeting features for doing backtesting
  • 07 Nov 2015 - Added feature to download tick data from Bloomberg (with example)
  • 05 Nov 2015 - Added intraday event study class (and example)
  • 02 Nov 2015 - Added easy wrapper for doing rolling correlations (and example)
  • 28 Oct 2015 - Added more sensitivity analysis for trading strategies
  • 26 Oct 2015 - Various bug fixes for Bloomberg Open API downloader
  • 14 Oct 2015 - Added capability to do parallel downloading of market data (thread/multiprocessing library), with an example for benchmarking and bug fixes for Bloomberg downloader
  • 25 Sep 2015 - Refactored examples into different folders / more seasonality examples
  • 19 Sep 2015 - Added support for Plotly choropleth map plots & easy downloading of economic data via FRED/Bloomberg/Quandl
  • 12 Sep 2015 - Added basic support for PyFolio for statistical analysis of strategies
  • 04 Sep 2015 - Added StrategyTemplate for backtesting (with example) & bug fixes
  • 21 Aug 2015 - Added stacked charts (with matplotlib & bokeh) & several bug fixes
  • 15 Aug 2015 - Added bar charts (with matplotlib & bokeh) & added more time series filter functions
  • 09 Aug 2015 - Improved Bokeh support
  • 07 Aug 2015 - Added Plotly support (via Jorge Santos Cufflinks wrapper)
  • 04 Aug 2015 - Added ability to download from FRED and example for downloading from FRED.
  • 29 Jul 2015 - Added backtesting functions (including simple FX trend following strategy) and various bug fixes/comments.
  • 24 Jul 2015 - Added functions for doing simple seasonality studies and added examples.
  • 17 Jul 2015 - Created example to show how to use technical indicators.
  • 13 Jul 2015 - Changed location of conf, renamed examples folder to pythalesians_examples. Can now be installed using setup.py.
  • 10 Jul 2015 - Added ability to download Dukascopy FX tick data (data is free for personal use - check Dukascopy terms & conditions). Note that past month of data is generally not made available by Dukascopy

End of note

Разместим вашу рекламу

Пиши: mail@pythondigest.ru

Нашли опечатку?

Выделите фрагмент и отправьте нажатием Ctrl+Enter.

Система Orphus