Zipline Installation Problems/ Trading Evolved First Backtest Error

Question

I have been trying to get Zipline properly installed. I have followed the instructions in Ch 7. of Andreas Clenow book and ran into problems with the first backtest program. To eliminate that it was an error in my code I downloaded the books code that is supposed to work.

%matplotlib inline

from zipline import run_algorithm
from zipline.api import order_target_percent,symbol

from datetime import datetime
import pytz

import matplotlib.pyplot as plt

#debug
import pandas as pd

def initialize(context):
    context.stock = symbol('AAPL')
    context.index_average_window = 100
    
def handle_data(context, data):
    equities_hist = data.history(context.stock, "close", 
                                 context.index_average_window, "1d")
    if equities_hist[-1] > equities_hist.mean():
        stock_weight = 1.0
    else:
        stock_weight = 0.0
    order_target_percent(context.stock, stock_weight)

def analyze(context, perf):
    fig = plt.figure(figsize=(12, 8))
    ax = fig.add_subplot(311)
    ax.set_title('Strategy Results')
    ax.semilogy(perf['portfolio_value'], linestyle='-', 
                label='Equity Curve', linewidth=3.0)
    ax.legend()
    ax.grid(False)
    ax = fig.add_subplot(312)
    ax.plot(perf['gross_leverage'], 
            label='Exposure', linestyle='-', linewidth=1.0)
    ax.legend()
    ax.grid(True)
    ax = fig.add_subplot(313)
    ax.plot(perf['returns'], label='Returns', linestyle='-.', linewidth=1.0)
    ax.legend()
    ax.grid(True)

start_date=datetime(1996,1,1,tzinfo=pytz.UTC)

end_date=datetime(2018,12,31,tzinfo=pytz.UTC)

results = run_algorithm(
    start=start_date, 
    end=end_date, 
    initialize=initialize, 
    analyze=analyze, 
    handle_data=handle_data, 
    capital_base=10000, 
    data_frequency='daily',
    bundle='quandl') 

I get the following error messages

AssertionError                            Traceback (most recent call last)
<ipython-input-8-b9def796232e> in <module>
      9     capital_base=10000,
     10     data_frequency='daily',
---> 11     bundle='quandl'
     12 ) 

~\Anaconda3202007\envs\zip35\lib\site-packages\zipline\utils\run_algo.py in run_algorithm(start, end, initialize, capital_base, handle_data, before_trading_start, analyze, data_frequency, bundle, bundle_timestamp, trading_calendar, metrics_set, benchmark_returns, default_extension, extensions, strict_extensions, environ, blotter)
    405         environ=environ,
    406         blotter=blotter,
--> 407         benchmark_spec=benchmark_spec,
    408     )
    409 

~\Anaconda3202007\envs\zip35\lib\site-packages\zipline\utils\run_algo.py in _run(handle_data, initialize, before_trading_start, analyze, algofile, algotext, defines, data_frequency, capital_base, bundle, bundle_timestamp, start, end, output, trading_calendar, print_algo, metrics_set, local_namespace, environ, blotter, benchmark_spec)
    201                 trading_calendar=trading_calendar,
    202                 capital_base=capital_base,
--> 203                 data_frequency=data_frequency,
    204             ),
    205             metrics_set=metrics_set,

~\Anaconda3202007\envs\zip35\lib\site-packages\zipline\finance\trading.py in __init__(self, start_session, end_session, trading_calendar, capital_base, emission_rate, data_frequency, arena)
     36                  arena='backtest'):
     37 
---> 38         assert type(start_session) == pd.Timestamp
     39         assert type(end_session) == pd.Timestamp
     40 

AssertionError:

As the code worked for the author and i have yet to get any working code out of zipline i assume it is an install problem. Following the zipline install instructions at this link https://pythonprogramming.net/zipline-local-install-python-programming-for-finance/ I checked the C packages that supposed to be needed in Anaconda Navigator and discovered that wrapt,cython and cordereddict were not installed and installed them via AN.

I get the same exact assertion error.

So I thought I would try a different simpler piece of code from the Py4fi link above for testing the install.

%load_ext zipline
from zipline.api import order, record, symbol


def initialize(context):
    pass

def handle_data(context, data):
    order(symbol('AAPL'), 10)
    record(AAPL=data.current(symbol('AAPL'), 'price'))

%zipline --bundle quantopian-quandl --start 2000-1-1 --end 2012-1-1 -o backtest.pickle

In this case I get the following error:

NoBenchmark                               Traceback (most recent call last)
~\Anaconda3202007\envs\zip35\lib\site-packages\zipline\utils\run_algo.py in _run(handle_data, initialize, before_trading_start, analyze, algofile, algotext, defines, data_frequency, capital_base, bundle, bundle_timestamp, start, end, output, trading_calendar, print_algo, metrics_set, local_namespace, environ, blotter, benchmark_spec)
    215                 'algo_filename': getattr(algofile, 'name', '<algorithm>'),
--> 216                 'script': algotext,
    217             }

RunAlgoError: No ``benchmark_spec`` was provided, and ``zipline.api.set_benchmark`` was not called in ``initialize``.

BTW - I am running on Windows.

Any advice would be appreciated.

Answer 1

Change the start_date from '1996-1-1' to a later date, e.g. '2004-4-1'. Change the end_date from '2018-12-31'to an earlier date, e.g. '2018-3-27'.

Then the two "RuntimeWarning: invalid value encountered in true_divide..." disappear.

Answer 2

Zipline is running its backtest loading stock prices from Quandl => bundle='quandl'.

You need to create a free account at Quandl and set your API Key:

import os
os.environ['QUANDL_API_KEY'] = 'you_api_key_here'

Then execute:

!zipline bundles
!zipline ingest -b quantopian-quandl

Answer 3

To get this to work I had to do 3 things but I am still getting runtime warnings but the plots look good.

Is it possible that these errors are the result of not having all the right packages installed?

No. 1 - To address the Assertion Error – I found the following in another post. # changed to pd.Timestamp as zipline was thowing errors #Import pandas import pandas as pd

# Set start and end date
# start_date=datetime(1996, 1, 1, tzinfo=pytz.UTC)
# end_date=datetime(2018, 12, 31, tzinfo=pytz.UTC)

start_date = pd.Timestamp('1996-1-1', tz='utc')
# changed to 303272018 as data does not exist past this
end_date = pd.Timestamp('2018-03-27', tz='utc')

No. 2 - Then I was back to the benchmark error, which I added set_benchmark to the initialize function to eliminate the error

# added this to get benchmark spec error to go away
# tried SPY but apparently this is not a symbol in the quandl bundle
zipline.api.set_benchmark(symbol('AAPL'))

No. 3. - Had to change the end date to 0/3/27/2018 as quandl data is not going to 12/31/2018

# changed to 303272018 as data does not exist past this
end_date = pd.Timestamp('2018-03-27', tz='utc')

Warning Messages

C:\Users\tbrug\Anaconda3202007\envs\zip35\lib\site-packages\empyrical\stats.py:711: RuntimeWarning: invalid value encountered in true_divide out=out, C:\Users\tbrug\Anaconda3202007\envs\zip35\lib\site-packages\empyrical\stats.py:797: RuntimeWarning: invalid value encountered in true_divide np.divide(average_annual_return, annualized_downside_risk, out=out)

Answer 4

I run into the same issues as you.

Just ignore the warnings. You can include the following at the top of your notebook:

import warnings
warnings.filterwarnings('ignore')