Files
catalyst/zipline/optimize/example.py
T
Thomas Wiecki b976c1252b Provides an iterative version of risk metrics.
I wrote this a little while ago as I noticed that a lot of time is spent
computing risk statistics. This is done over the complete history over
and over again while this could be done just by using the previously
computed value (iteratively).

We didn't go forward back then because for minute trade data the
difference was not significant enough. However, now with zipline
standalone I think most people will use daily (because that's
what's available) and it makes a huge difference
(speed-up of a couple of 100%).

Unfortunately, we can't just replace the existing one with an
iterative as for the final cumulative stats the batch is still
better. So that's not as nice, but the performance increase is
big enough for me to issue this PR (zipline is actually painfully
slow with daily data).

There is a unittest that compares that both produce exactly
the same outputs.

Speed measurements (for 500 trading days, daily source):

with iterative:
real 26.617 user 12.909 sys 6.112 pcpu 71.46

prior:
real 44.176 user 31.030 sys 11.381 pcpu 96.00
2012-10-17 23:41:30 -04:00

258 lines
7.5 KiB
Python

#
# Copyright 2012 Quantopian, Inc.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# WARNING: This file is still work in progress and contains rather
# random code snippets.
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from zipline.gens.mavg import MovingAverage
from zipline.algorithm import TradingAlgorithm
from zipline.gens.transform import batch_transform
class DMA(TradingAlgorithm):
"""Dual Moving Average algorithm.
"""
def initialize(self, amount=100, short_window=20, long_window=40):
self.amount = amount
self.events = 0
self.invested = {}
for sid in self.sids:
self.invested[sid] = False
self.add_transform(MovingAverage, 'short_mavg', ['price'],
market_aware=True,
days=short_window)
self.add_transform(MovingAverage, 'long_mavg', ['price'],
market_aware=True,
days=long_window)
def handle_data(self, data):
self.events += 1
for sid in self.sids:
# access transforms via their user-defined tag
if (data[sid].short_mavg['price'] >
data[sid].long_mavg['price']) \
and not self.invested[sid]:
self.order(sid, self.amount)
self.invested[sid] = True
elif (data[sid].short_mavg['price'] <
data[sid].long_mavg['price']) \
and self.invested[sid]:
self.order(sid, -self.amount)
self.invested[sid] = False
class DualMovingAverage(TradingAlgorithm):
"""Dual Moving Average algorithm.
"""
def initialize(self, short_window=200, long_window=400):
self.short_mavg = []
self.long_mavg = []
self.invested = False
self.add_transform(MovingAverage, 'short_mavg', ['price'],
market_aware=True,
days=short_window)
self.add_transform(MovingAverage, 'long_mavg', ['price'],
market_aware=True,
days=long_window)
def handle_data(self, data):
self.short_mavg.append(data['AAPL'].short_mavg['price'])
self.long_mavg.append(data['AAPL'].long_mavg['price'])
if (data['AAPL'].short_mavg['price'] >
data['AAPL'].long_mavg['price']) and not self.invested:
self.order('AAPL', 100)
self.invested = True
elif (data['AAPL'].short_mavg['price'] <
data['AAPL'].long_mavg['price']) and self.invested:
self.order('AAPL', -100)
self.invested = False
def load_close_px(indexes=None, stocks=None):
from pandas.io.data import DataReader
import pytz
from collections import OrderedDict
if indexes is None:
indexes = {'SPX': '^GSPC'}
if stocks is None:
stocks = ['AAPL', 'GE', 'IBM', 'MSFT', 'XOM', 'AA', 'JNJ', 'PEP', 'KO']
start = pd.datetime(1990, 1, 1, 0, 0, 0, 0, pytz.utc)
end = pd.datetime(2000, 1, 1, 0, 0, 0, 0, pytz.utc)
data = OrderedDict()
for stock in stocks:
print stock
stkd = DataReader(stock, 'yahoo', start, end).sort_index()
data[stock] = stkd
for name, ticker in indexes.iteritems():
print name
stkd = DataReader(ticker, 'yahoo', start, end).sort_index()
data[name] = stkd
df = pd.DataFrame({key: d['Close'] for key, d in data.iteritems()})
df.index = df.index.tz_localize(pytz.utc)
df.save('close_px.dat')
return df
def run((short_window, long_window)):
data = pd.load('close_px.dat')
#data = load_close_px()
myalgo = DMA([0, 1],
amount=100,
short_window=short_window,
long_window=long_window)
stats = myalgo.run(data)
stats['sw'] = short_window
stats['lw'] = long_window
return stats
def explore_params():
sws, lws = np.mgrid[10:20:5, 10:20:5]
stats_all = map(run, zip(sws.flatten(), lws.flatten()))
stats = pd.concat(stats_all)
returns = stats.groupby(['sw', 'lw']).sum()
plt.contourf(sws, lws, returns.returns.reshape(sws.shape))
plt.xlabel('Short window length')
plt.ylabel('Long window length')
plt.savefig('DMA_contour.png')
plt.show()
def get_opt_holdings_qp(univ_rets, track_rets):
from cvxopt import matrix
from cvxopt.solvers import qp
# set up the QP for CVXOPT
# .5 x' P x + q'x
# P = 2 * R'R
# q = - 2 * bmk'R
R = univ_rets.values
b = track_rets.values
P = matrix(2 * np.dot(R.T, R))
q = matrix(-2 * np.dot(R.T, b))
result = qp(P, q)
if result['status'] != 'optimal':
raise Exception('optimum not reached by QP')
return pd.Series(np.array(result['x']).ravel(), index=univ_rets.columns)
def opt_portfolio(cov, budget, min_return):
from cvxopt import matrix
from cvxopt.solvers import qp
n = len(cov)
cov = matrix(2 * cov)
q = matrix(np.zeros(n))
h = matrix(budget) # G*x < h
# coneqp
result = qp(cov, q, h=h)
if result['status'] != 'optimal':
raise Exception('optimum not reached by QP')
return pd.Series(np.array(result['x']).ravel())
def calc_te(weights, univ_rets, track_rets):
port_rets = (univ_rets * weights).sum(1)
return (port_rets - track_rets).std()
def plot_returns(port_returns, bmk_returns):
plt.figure()
cum_port = ((1 + port_returns).cumprod() - 1)
cum_bmk = ((1 + bmk_returns).cumprod() - 1)
# cum_port = port_returns.cumsum()
# cum_bmk = bmk_returns.cumsum()
cum_port.plot(label='Portfolio returns')
cum_bmk.plot(label='Benchmark')
plt.title('Portfolio performance')
plt.legend(loc='best')
#print run((10, 20))
import statsmodels.api as sm
@batch_transform
def ols_transform(data, spreads):
p0 = data.price['PEP']
p1 = sm.add_constant(data.price['KO'])
beta, intercept = sm.OLS(p0, p1).fit().params
spread = (data.price['PEP'] - (beta * data.price['KO'] + intercept))[-1]
if len(spreads) > 10:
z_score = (spread - np.mean(spreads[-10:])) / np.std(spreads[-10:])
else:
z_score = np.nan
spreads.append(spread)
return z_score
class Pairtrade(TradingAlgorithm):
def initialize(self):
self.spreads = []
self.invested = False
self.ols_transform = ols_transform(refresh_period=10, days=10)
def handle_data(self, data):
zscore = self.ols_transform.handle_data(data, self.spreads)
if zscore == np.nan:
return
if zscore >= 2.0 and not self.invested:
self.order('PEP', int(100 / data['PEP'].price))
self.order('KO', -int(100 / data['KO'].price))
elif zscore <= -2.0 and not self.invested:
self.order('KO', -int(100 / data['KO'].price))
self.order('PEP', int(100 / data['PEP'].price))
elif abs(zscore) < .5 and self.invested:
pass
def run_pairtrade():
data = load_close_px()
data.save('close_px.dat')
#data = pd.load('close_px.dat')
myalgo = Pairtrade()
stats = myalgo.run(data)
return stats