Zipline Bundles

This repository contains some zipline data bundles, which are used to download and extract historical price data for backtesting in zipline platform. zipline is a backtesting framework written in python, which can be used to test, analyze and visualize trading strategies. It currently powers Quantopian, a free rich community centered hosted platform for trading strategy development and many more.

Quick start

You have already cloned the repository and switched the working directory to the repository root. Then, in an environment where zipline is installed and working, adding all data bundles can be done by

python install.py

Note that the installer complains if there already exist python modules with the same name. To force the installer to overwrite the existing modules, add -f. Apart from zipline itself and its dependencies, there are additional dependencies used by each bundle. To install them all run

pip install -r requirements.txt

You can check if the installation is complete by running:

zipline bundles

You should see new bundles are added to the list:

csvdir <no ingestions>
iex <no ingestions>
quandl <no ingestions>
quantopian-quandl <no ingestions>
yahoo_csv <no ingestions>
yahoo_direct <no ingestions>

To test the installation, a simple strategy backtest can be executed over price data read by one of the bundles. For this, yahoo_csv bundle is used, which reads data from a directory containing csv files that are downloaded from yahoo finance. First, ingest price data stored in data directory:

YAHOO_CSVDIR=./data/ zipline ingest -b yahoo_csv

Then backtest buy and hold strategy over the ingested data:

zipline run -f tests/buy_and_hold.py -b yahoo_csv --start 2019-07-02 --end 2020-07-02

The cumulative return of the strategy will be depicted in a plot after backtesting.

Bundles

The following bundles are currently defined by the repository.

Bundle	Data Source	Dependency	Module
`yahoo_csv`	csv files downloaded from yahoo finance	none	none
`yahoo_direct`	yahoo finance	`yahoofinancials`	yahoo.py
`iex`	IEX cloud	`iexfinance`	iex.py
`binance_daily`	daily price from binance exchange	`python-binance`	binance.py
`binance_min`	per minute price from binance exchange	`python-binance`	binance.py

`yahoo_csv`

This bundle takes data from CSV files downloaded from yahoo finance. Each file contains price data of a single asset and shall be named as assert_name.csv. The bundle reads all the csv files located in a directory given by environment variable YAHOO_CSVDIR:

YAHOO_CSVDIR=/path/to/csvdir zipline ingest -b yahoo_csv

`yahoo_direct`

It directly downloads price data from yahoo finance. The bundle extracts asset names from environment variable YAHOO_SYM_LST, which holds a comma separated list of asset names. For example:

YAHOO_SYM_LST=SPY,AAPL zipline ingest -b yahoo_direct

ingests price data of assets SPY and AAPL. The start and the end date of ingestion can be set into variables start_date and end_date, respectively. These variables are passed to function get_downloader where the bundle is registered in $HOME/.zipline/extension.py. Here is how the registration may look like:

register('yahoo_direct', # bundle's name
         direct_ingester('YAHOO',
                         every_min_bar=False,
                         symbol_list_env='YAHOO_SYM_LST', # the environment variable holding the comma separated list of assert names
                         downloader=yahoo.get_downloader(start_date='2010-01-01',
                                                         end_date='2020-01-01'
                         ),
         ),
         calendar_name='NYSE',
)

In addition to the start and the end date, the environment variable name holding price data can be set here. direct_ingester can additionally takes callable filter_cb. It takes as a parameter a data frame that is just retured from the downloader and returns a new data frame. It is useful when the downloaded price data needs additional prepossessing.

`iex`

It downloads price data from IEX cloud. Its usage is fairly similar to that of yahoo_direct. Fetching price data from IEX cloud however requires passing a valid API token, which is stored in environment variable IEX_TOKEN. Moreover, the environment variable storing asset names is called IEX_SYM_LST.

`binance_daily` and `binance_min`

Both collect data from binance cryptocurrency exchange with daily and minutely frequency. The list of symbols are taken from environment variable BINANCE_SYM_LST. Moreover, the API key and the secret key are supposed to set in environment variables BINANCE_API_KEY and BINANCE_SECRET_KEY. For example, the following command ingests the daily price of bitcoin and ethereum in USD.

BINANCE_API_KEY=your_api_key BINANCE_SECRET_KEY=your_secret_key BINANCE_SYM_LST=BTCUSDT,ETHUSDT zipline ingest -b binance_daily

Manual installation

install.py takes the following steps to add the bundles:

copy extension.py into ~/.zipline/,
add ingester.py as well as the proper module for each bundle listed in above table into package zipline.data.bundles, i.e. copy the modules into where the package is located. Package location differs depending on the way zipline is installed. One way to find out the location in an environment with zipline installed is to run the following code:
```
python -c 'import zipline.data.bundles as bdl; print(bdl.__path__)'
```

If only a subset of bundles is needed, one way is to keep their registration in extension.py, their dependency in requirements.txt and their related modules in variable src_ing inside install.py. Then, use the installation script!

Adding new bundles

It is possible to define new data bundles using the structures provided by this repository. The process is explained in this post in more detail.

wshobson/zipline_bundles