How to update the deprecated python zipline.transforms module?

Question

I wrote a python program using the quantopian zipline package http://www.zipline.io/beginner-tutorial.html. I recently updated the package and have encountered that the zipline.transforms package is deprecated. I was using two functions from the zipline.transforms package, batch_transform() and MovingAverage.

I haven't been able to find a good post demonstrating how to fix this, other than saying to replace batch_transform with the history() function. However, I am unaware how exactly to replace it. I haven't found a post telling how to fix the MovingAverage deprecation.

Here is my code I am using.

from zipline.algorithm import TradingAlgorithm
from zipline.transforms import batch_transform
from zipline.transforms import MovingAverage


class TradingStrategy(TradingAlgorithm):

    def initialize(self, window_length=6):
        self.add_transform(
            MovingAverage, 'kernel', ['price'], window_length=self.window_length)

    @batch_transform
    def get_data(data, context):
        '''
        Collector for some days of historical prices.
        '''
        daily_prices = data.price[STOCKS + [BENCHMARK]]
        return daily_prices

strategy = TradingStrategy()

Could someone provide an example of how to update the code above? I assume there are many people dealing with the issues given how popular quantopian is.

Pulpboard

related: #37697227 – Endblown

here is how history was added: github.com/quantopian/zipline/pull/315/files – Endblown

Answer 1

There doesn't seem to be a direct way to use history instead of batch_transform.

It looks to me that not only were the methods changed, but the way in which they are intended to be used were completely changed also.

The documentation mentions the following:

Every zipline algorithm consists of two functions you have to define:

• initialize(context)
• handle_data(context, data)

---

Here's an example from the docs of using the history method to create some basic moving averages:

def initialize(context):
    context.i = 0
    context.asset = symbol('AAPL')


def handle_data(context, data):
    # Skip first 300 days to get full windows
    context.i += 1
    if context.i < 300:
        return

    # Compute averages
    # data.history() has to be called with the same params
    # from above and returns a pandas dataframe.
    short_mavg = data.history(context.asset, 'price', bar_count=100, frequency="1d").mean()
    long_mavg = data.history(context.asset, 'price', bar_count=300, frequency="1d").mean()

    # Trading logic
    if short_mavg > long_mavg:
        # order_target orders as many shares as needed to
        # achieve the desired number of shares.
        order_target(context.asset, 100)
    elif short_mavg < long_mavg:
        order_target(context.asset, 0)

    # Save values for later inspection
    record(AAPL=data.current(context.asset, 'price'),
           short_mavg=short_mavg,
           long_mavg=long_mavg)

Kulp

Could you provide an example of code using batch_transform that gives you the same result as above, so I can compare the two? – Pulpboard