# -*- coding: utf-8 -*-
"""
quantshim dev scratch. Licensed under GPL3
written by [email protected]
"""
# Import the libraries we will use here
import datetime
import pytz
import math
import numpy
import pandas
import scipy
import scipy.stats
import zipline
import functools
import collections
import sklearn
import sklearn.naive_bayes
#import sklearn.naive_bayes.BernoulliNB
import sklearn.linear_model
import sklearn.ensemble
import talib
is_offline_Zipline = False
#quantopian shims
class WorstSpreadSlippage(slippage.SlippageModel):
'''will trade at the worst value of the order minute. high if long, low if short.
additionally, supports 'VolumeShareSlippage' functionality, which further biases price/volume'''
def __init__(this, volume_limit=.25, price_impact=0.1, ohlcWeighted=False):
this.volume_limit = volume_limit
this.price_impact = price_impact
this.ohlcWeighted = ohlcWeighted
pass
def __processVolumeShareSlippage(self,event,order, targetPrice):
'''coppied implementation from VolumeShareSlippage.process_order(), found here: https://github.com/quantopian/zipline/blob/4860a966b3a3102fa80d43f393155e53015cc349/zipline/finance/slippage.py
modification: we return the final (price,volume) tuple for our main .process_order() to use, instead of executing the order
RETURNS: final (price,volume) tuple'''
########
max_volume = self.volume_limit * event.volume
# price impact accounts for the total volume of transactions
# created against the current minute bar
remaining_volume = max_volume - self.volume_for_bar
if remaining_volume < 1:
# we can't fill any more transactions
return (0.0,0)
# the current order amount will be the min of the
# volume available in the bar or the open amount.
cur_volume = int(min(remaining_volume, abs(order.open_amount)))
if cur_volume < 1:
return (0.0,0)
# tally the current amount into our total amount ordered.
# total amount will be used to calculate price impact
total_volume = self.volume_for_bar + cur_volume
volume_share = min(total_volume / event.volume,
self.volume_limit)
simulated_impact = volume_share ** 2 \
* math.copysign(self.price_impact, order.direction) \
* targetPrice
#return create_transaction(
# event,
# order,
# # In the future, we may want to change the next line
# # for limit pricing
# event.price + simulated_impact,
# math.copysign(cur_volume, order.direction)
return (targetPrice + simulated_impact,int(math.copysign(cur_volume, order.direction)))
def process_order(this,trade_bar,order):
#worst spread
if order.amount < 0:
targetPrice = trade_bar.low
else:
targetPrice = trade_bar.high
#trade at the open
#targetPrice = trade_bar.open_price
if this.ohlcWeighted:
#midpoint, ohlc weighted
targetPrice = (targetPrice + trade_bar.open_price + trade_bar.high + trade_bar.low + trade_bar.close_price) / 5
pass
price, volume = this.__processVolumeShareSlippage(trade_bar,order,targetPrice)
priceSlippage = trade_bar.close_price - price
volumeSlippage = order.amount - volume
if price == 0.0 or volume == 0:
return
#logger.info(price)
logger.info("ORDER_COMMITTED: {0} shares {1} @ {2} \n\t v={8} o={4} h={5} l={6} c={7} \t (WorstSpreadSlippage: vol= -{9} price= {3:.2f})"
.format(volume,trade_bar.sid.symbol,price,priceSlippage, trade_bar.open_price, trade_bar.high, trade_bar.low, trade_bar.close_price, trade_bar.volume,volumeSlippage))
return slippage.create_transaction(trade_bar,
order,
price,
order.amount)
class TradeAtTheOpenSlippageModel_Simple(slippage.SlippageModel):
def __init__(self, fractionOfOpenCloseRange):
self.fractionOfOpenCloseRange = fractionOfOpenCloseRange
def process_order(self, trade_bar, order):
openPrice = trade_bar.open_price
closePrice = trade_bar.price
ocRange = closePrice - openPrice
ocRange = ocRange * self.fractionOfOpenCloseRange
targetExecutionPrice = openPrice + ocRange
# Create the transaction using the new price we've calculated.
return slippage.create_transaction(
trade_bar,
order,
targetExecutionPrice,
order.amount
)
class TradeAtTheOpenSlippage(slippage.SlippageModel):
'''will trade at the open, good for daily use, kind of not good otherwise.'''
def __init__(this, volume_limit=.25, price_impact=0.1):
this.volume_limit = volume_limit
this.price_impact = price_impact
pass
def __processVolumeShareSlippage(self,event,order, targetPrice):
'''coppied implementation from VolumeShareSlippage.process_order(), found here: https://github.com/quantopian/zipline/blob/4860a966b3a3102fa80d43f393155e53015cc349/zipline/finance/slippage.py
modification: we return the final (price,volume) tuple for our main .process_order() to use, instead of executing the order
RETURNS: final (price,volume) tuple'''
########
max_volume = self.volume_limit * event.volume
# price impact accounts for the total volume of transactions
# created against the current minute bar
remaining_volume = max_volume - self.volume_for_bar
if remaining_volume < 1:
# we can't fill any more transactions
return (0.0,0)
# the current order amount will be the min of the
# volume available in the bar or the open amount.
cur_volume = int(min(remaining_volume, abs(order.open_amount)))
if cur_volume < 1:
return (0.0,0)
# tally the current amount into our total amount ordered.
# total amount will be used to calculate price impact
total_volume = self.volume_for_bar + cur_volume
volume_share = min(total_volume / event.volume,
self.volume_limit)
simulated_impact = volume_share ** 2 \
* math.copysign(self.price_impact, order.direction) \
* targetPrice
#return create_transaction(
# event,
# order,
# # In the future, we may want to change the next line
# # for limit pricing
# event.price + simulated_impact,
# math.copysign(cur_volume, order.direction)
return (targetPrice + simulated_impact,int(math.copysign(cur_volume, order.direction)))
def process_order(this,trade_bar,order):
#if order.amount < 0:
# targetPrice = trade_bar.low
#else:
# targetPrice = trade_bar.high
targetPrice = trade_bar.open_price
price, volume = this.__processVolumeShareSlippage(trade_bar,order,targetPrice)
priceSlippage = trade_bar.close_price - price
volumeSlippage = order.amount - volume
if price == 0.0 or volume == 0:
return
#logger.info(price)
logger.info("ORDER_COMMITTED: {0} shares {1} @ {2} \n\t v={8} o={4} h={5} l={6} c={7} \t (TradeAtTheOpenSlippage: vol= -{9} price= {3:.2f})"
.format(volume,trade_bar.sid.symbol,price,priceSlippage, trade_bar.open_price, trade_bar.high, trade_bar.low, trade_bar.close_price, trade_bar.volume,volumeSlippage))
return slippage.create_transaction(trade_bar,
order,
price,
order.amount)
class CustomSlippage(slippage.SlippageModel):
''' allows customizing slippage if desired, though mostly used for logging your order details to the console'''
def __init__(this, volume_limit=.25, price_impact=0.1, ohlcWeighted=False):
this.volume_limit = volume_limit
this.price_impact = price_impact
this.ohlcWeighted = ohlcWeighted
pass
def __processVolumeShareSlippage(self,event,order, targetPrice):
'''coppied implementation from VolumeShareSlippage.process_order(), found here: https://github.com/quantopian/zipline/blob/4860a966b3a3102fa80d43f393155e53015cc349/zipline/finance/slippage.py
modification: we return the final (price,volume) tuple for our main .process_order() to use, instead of executing the order
RETURNS: final (price,volume) tuple'''
########
max_volume = self.volume_limit * event.volume
# price impact accounts for the total volume of transactions
# created against the current minute bar
remaining_volume = max_volume - self.volume_for_bar
if remaining_volume < 1:
# we can't fill any more transactions
return (0.0,0)
# the current order amount will be the min of the
# volume available in the bar or the open amount.
cur_volume = int(min(remaining_volume, abs(order.open_amount)))
if cur_volume < 1:
return (0.0,0)
# tally the current amount into our total amount ordered.
# total amount will be used to calculate price impact
total_volume = self.volume_for_bar + cur_volume
volume_share = min(total_volume / event.volume,
self.volume_limit)
simulated_impact = volume_share ** 2 * math.copysign(self.price_impact, order.direction) * targetPrice
#return create_transaction(
# event,
# order,
# # In the future, we may want to change the next line
# # for limit pricing
# event.price + simulated_impact,
# math.copysign(cur_volume, order.direction)
return (targetPrice + simulated_impact,int(math.copysign(cur_volume, order.direction)))
def process_order(this,trade_bar,order):
####worst spread
#if order.amount < 0:
# targetPrice = trade_bar.low
#else:
# targetPrice = trade_bar.high
####trade at the open
#targetPrice = trade_bar.open_price
####trade at the close
targetPrice = trade_bar.close_price
if this.ohlcWeighted:
#midpoint, ohlc weighted
targetPrice = (targetPrice + trade_bar.open_price + trade_bar.high + trade_bar.low + trade_bar.close_price) / 5
pass
price, volume = this.__processVolumeShareSlippage(trade_bar,order,targetPrice)
priceSlippage = trade_bar.close_price - price
volumeSlippage = order.amount - volume
if price == 0.0 or volume == 0:
return
#construct our pnl once this transaction is comitted (logged below)
pnl = _g.context.portfolio.pnl + (price * order.amount) - (trade_bar.close_price * order.amount)
#logger.info(price)
logger.info("ORDER_COMMITTED: {0} shares {1} @ {2} \n\t v={8} o={4} h={5} l={6} c={7} \t (Slippage: vol= -{9} price= {3:.2f})\n\tpnl={10}"
.format(volume,trade_bar.sid.symbol,price,priceSlippage, trade_bar.open_price, trade_bar.high, trade_bar.low, trade_bar.close_price, trade_bar.volume,volumeSlippage, pnl))
return slippage.create_transaction(trade_bar,
order,
price,
order.amount)
class Logger():
'''shim for exposing the same logging definitions to visualstudio intelisence'''
def __init__(this, logErrors=True, logInfos=True, logWarns=True, logDebugs=True):
this.__logErrors = logErrors
this.__logInfos = logInfos
this.__logWarns = logWarns
this.__logDebugs = logDebugs
this.__recordHistory = {}
this.__lastKnownDay = None
pass
def error(this, message):
if not this.__logErrors: return
log.error(this.__wrapMessage(message))
pass
def info(this, message):
if not this.__logInfos: return
#log.info(this.__wrapMessage(message))
pass
def warn(this, message):
if not this.__logWarns: return
log.warn(this.__wrapMessage(message))
pass
def debug(this, message):
if not this.__logDebugs: return
log.debug(this.__wrapMessage(message))
pass
def __wrapMessage(this,message):
this.__trySpamDailyLogs()
timestamp = _g.context.framework._getDatetime()
#return str(timestamp) + message
time = timestamp.strftime("%H:%M")
#if timestamp.second!=0:
# time += ":{0}".format(timestamp.second)
return str(time) + ": " + str(message)
pass
def debugAccumulateDaily(this,key,message):
'''writes the log once a day to avoid spam. includes timestamp automatically'''
if not this.__logDebugs: return
msg = this.__wrapMessage(message)
this.__storeToDailyLog(key,msg)
def debugOnceDaily(this,key,message):
if not this.__logDebugs: return
this.__storeToDailyLog(key,message)
this.__recordHistory[key] = this.__recordHistory[key][0:1]
this.__trySpamDailyLogs()
pass
def __storeToDailyLog(this,key,message):
if not this.__recordHistory.has_key(key):
this.__recordHistory[key] = []
this.__recordHistory[key].append(message)
pass
def __trySpamDailyLogs(this):
if _g.context.framework.thisFrameDay != this.__lastKnownDay:
#new day, dump our previous logs
this.__lastKnownDay = _g.context.framework.thisFrameDay
for key,values in this.__recordHistory.items():
this.debug("[email protected]{0}=\n{1}".format(key,",".join(values)))
values[:] = [] #clear it
pass
def record(this, name,value, logDaily=False):
this.__trySpamDailyLogs()
if(logDaily == True):
this.__storeToDailyLog(name,"%0.4f" % value)
record(**{name:value})
def recordNormalized(this, name,value,baseline=1,subtract=0, logDaily=False):
'''normalize values to a 0 to 1 range'''
if value - subtract == 0 or baseline == 0:
toRecord = 0
else:
toRecord = (value - subtract) / baseline
this.record(name,toRecord,logDaily=logDaily)
#def getLastRecord(this,name):
# '''returns the last recorded value. only exists if doing daily
# outputs, and during the day. returns None if name not found'''
# return this.__recordHistory.get(name)
pass
global logger
logger = Logger() #(logDebugs=False)
class Shims():
'''SHIM OF QUANTOPIAN INTERNAL REPRESENTATION. here for intelisence only. you SHOULD NOT actually instantiate these.'''
class Position:
'''
The position object represents a current open position, and is contained inside the positions dictionary.
For example, if you had an open AAPL position, you'd access it using context.portfolio.positions[sid(24)].
The position object has the following properties:
amount = 0 #Integer: Whole number of shares in this position.
cost_basis = 0.0 #Float: The volume-weighted average price paid per share in this position.
last_sale_price = 0.0 #Float: Price at last sale of this security.
sid = 0 #Integer: The ID of the security.
'''
def __init__(this):
this.amount = 0 #Integer: Whole number of shares in this position.
this.cost_basis = 0.0 #Float: The volume-weighted average price paid per share in this position.
this.last_sale_price = 0.0 #Float: Price at last sale of this security.
this.sid = 0 #Integer: The ID of the security.
class Context():
def __init__(this , portfolio=zipline.protocol.Portfolio()): #, tradingAlgorithm = zipline.TradingAlgorithm()):
this.portfolio = portfolio
#this.tradingAlgorithm = tradingAlgorithm
pass
pass
class _TradingAlgorithm_QuantopianShim:
'''shim of zipline.TradingAlgorithm for use on quantopian '''
def __init__(this):
#this.logger = Shims._Logger()
#this.logger = log
pass
def order(this,sid,amount,limit_price=None, stop_price=None):
'''
Places an order for the specified security of the specified number of shares. Order type is inferred from the parameters used. If only sid and amount are used as parameters, the order is placed as a market order.
Parameters
sid: A security object.
amount: The integer amount of shares. Positive means buy, negative means sell.
limit_price: (optional) The price at which the limit order becomes active. If used with stop_price, the price where the limit order becomes active after stop_price is reached.
stop_price: (optional) The price at which the order converts to a market order. If used with limit_price, the price where the order converts to a limit order.
Returns
An order id.
'''
if sid is Security:
security = sid
else:
security = this.context.framework.allSecurities[sid]
#logger.info("{0} ordering {1}".format(security.qsec,amount))
orderId = order(security.qsec,amount,limit_price,stop_price)
return orderId
pass
def order_percent(self, sid, percent, limit_price=None, stop_price=None):
"""
Place an order in the specified security corresponding to the given
percent of the current portfolio value.
Note that percent must expressed as a decimal (0.50 means 50\%).
"""
value = self.context.portfolio.portfolio_value * percent
return self.order_value(sid, value, limit_price, stop_price)
def order_target(self, sid, target, limit_price=None, stop_price=None):
"""
Place an order to adjust a position to a target number of shares. If
the position doesn't already exist, this is equivalent to placing a new
order. If the position does exist, this is equivalent to placing an
order for the difference between the target number of shares and the
current number of shares.
"""
if sid in self.context.portfolio.positions:
current_position = self.context.portfolio.positions[sid].amount
req_shares = target - current_position
return self.order(sid, req_shares, limit_price, stop_price)
else:
return self.order(sid, target, limit_price, stop_price)
def order_target_value(self, sid, target, limit_price=None,
stop_price=None):
"""
Place an order to adjust a position to a target value. If
the position doesn't already exist, this is equivalent to placing a new
order. If the position does exist, this is equivalent to placing an
order for the difference between the target value and the
current value.
"""
if sid in self.context.portfolio.positions:
current_position = self.context.portfolio.positions[sid].amount
current_price = self.context.portfolio.positions[sid].last_sale_price
current_value = current_position * current_price
req_value = target - current_value
return self.order_value(sid, req_value, limit_price, stop_price)
else:
return self.order_value(sid, target, limit_price, stop_price)
def order_target_percent(self, sid, target, limit_price=None,
stop_price=None):
"""
Place an order to adjust a position to a target percent of the
current portfolio value. If the position doesn't already exist, this is
equivalent to placing a new order. If the position does exist, this is
equivalent to placing an order for the difference between the target
percent and the current percent.
Note that target must expressed as a decimal (0.50 means 50\%).
"""
if sid in self.context.portfolio.positions:
current_position = self.context.portfolio.positions[sid].amount
current_price = self.context.portfolio.positions[sid].last_sale_price
current_value = current_position * current_price
else:
current_value = 0
target_value = self.context.portfolio.portfolio_value * target
req_value = target_value - current_value
return self.order_value(sid, req_value, limit_price, stop_price)
pass
#class _TradingAlgorithm_ZiplineShim(zipline.TradingAlgorithm):
# '''auto-generates a context to use'''
# def initialize(this):
# #delay initialize until start of first handle-data, so our
# #portfolio object is available
# #this.__isInitialized = False;
# this.context = Shims.Context()
# this.context.tradingAlgorithm = this
# #this.context.portfolio = this.portfolio
# pass
# def handle_data(this,data):
# this.context.portfolio = this.portfolio
# #if not this.__isInitialized:
# # this.__isInitialized=True
# # this.context.portfolio=this.portfolio
# this.context.framework._update(data)
# pass
# pass
class FrameHistory:
def __init__(this,parent,framework, data):
this.parent = parent
this.framework = framework
this.state = []
this.isActive = this.parent.isActive
#this.maxHistoryFrames = this.framework.maxHistoryFrames
#assert(this.framework.simFrame == this.parent.simFrame, "parent frame
#does not match")
this.initialize(data)
def initialize(this, data):
'''overridable'''
logger.error("FrameHistory.initialize() invoked. You should override this method.")
pass
def constructFrameState(this,data):
'''override and return the frame state, this will be prepended to history
if you return NONE, the frame state (history) is not modified.'''
logger.error("FrameHistory.constructFrameState() invoked. You should override this method.")
pass
def _update(this,data):
this.isActive = this.parent.isActive
if not this.isActive:
return
currentState = this.constructFrameState(data)
if(currentState != None):
currentState.datetime = this.framework._getDatetime()
currentState.simFrame = this.framework.simFrame
this.state.insert(0,currentState)
del this.state[this.framework.maxHistoryFrames:]
class StrategyPosition:
'''allows two or more stratgies to controll their own positions (orders) for securities they care about,
without interfering with the orders of other strategies.
To use: each strategy should set security.myStrategyPositon.targetCapitalSharePercent, which is a percentage of your entire portfolio's value
then execute the order (and/or rebalance) by invoking security.myStrategyPosition.processOrder()
'''
def __init__(this, security, strategyName):
this._security = security
this._strategyName = strategyName
this._lastOrderId = 0
this._lastStopOrderId = 0
this._currentCapitalSharePercent = 0.0
this._currentShares = 0
#for the last trade roundtrip, the aproximate returns. set every time our percent changes to zero
this._lastRoundtripReturns = 0.0
#this is editable
this.targetCapitalSharePercent = 0.0
#price when we decided to order, not actually the fulfillment price
this.__lastOrderPrice = 0.0
this.__currentPeakGains = 0.0
this.__currentPeakGainsDecay = 0.0
this._currentReturns = 0.0 #returns of current open position.
this._totalTrades = 0 #total trades we execute via this strategyPosition. note that due to partial fills, this may be less than actual trades
def processOrder(this, data, rebalanceThreshholdPercent=0.05, maxLosses=None, maxGainsAdditionalDrawdown=None, maxGainsDecay=0.01): #, OBSOLETE_stopLimitPercent=0.0, OBSOLETE_momentumStopLimit = True, OBSOLETE_decayMomentum = 0.001):
''' set rebalanceThreshholdPercent to zero (0.0) to cause the position to readjust even if the targetPercentage doesn't change. this is useful for reinvesting divideds / etc
but is set to 0.05 (5 percent) so we don't spam orders
maxLosses: close if our open position suffers a loss of this percent or more
maxGainsAdditionalDrawdown : close if our open position's gains suffer a decrease of this+maxLosses or more.
maxGainsDecay : over time this will reduce the acceptable gains drawdown (specified by maxGainsAdditionalDrawdown) so that on long-running gains we don't incur such a large drawdown before closing.
'''
#if momentumStopLimit == True (the default) we will stopLimitPercent based on the peak gains, not based on the original purchase price (this is generally a good ideas as it will maximize your gains)
#decayMomentum : = if == 0.01 and using momentumStopLimit==True, we will decay the position's survival chances by 1% per tick until it's finally closed.
if this._currentCapitalSharePercent == 0.0 and this.targetCapitalSharePercent == 0.0:
#no work to do
return 0
currentPrice = data[this._security.qsec].close_price
if this._currentCapitalSharePercent == this.targetCapitalSharePercent and this._currentCapitalSharePercent != 0.0:
#update current returns
this._currentReturns = (currentPrice - this.__lastOrderPrice) / this.__lastOrderPrice * math.copysign(1.0,this._currentCapitalSharePercent)
else:
#target is different so reset our returns as we are about to change our order
this._currentReturns = 0.0
if this._currentCapitalSharePercent == this.targetCapitalSharePercent and maxGainsAdditionalDrawdown != None:
##handle maxGains stoplimits
gainsPercent = this._currentReturns - this.__currentPeakGainsDecay
#if gainsPercent < -maxLosses:
# #loosing, so close out
# logger.debug("loosing, so close out. gainsPercent={0}, maxLosses={1}".format(gainsPercent, maxLosses))
# this.targetCapitalSharePercent = 0.0
#else:
if this._currentReturns > this.__currentPeakGains:
this.__currentPeakGains = this._currentReturns
this.__currentPeakGainsDecay = 0.0 #reset decay
else:
#need to see if our gain exceed our stoplimitGains threshhold
gainsFloorThreshhold = this.__currentPeakGains * maxGainsAdditionalDrawdown
if gainsPercent < gainsFloorThreshhold:
lossesFromPeak = this.__currentPeakGains - gainsPercent
if maxLosses != None and lossesFromPeak < maxLosses:
#we are not yet exceeding maxLosses (from our peak) so don't close out yet
logger.debug("we are not yet exceeding maxLosses (from our peak) so don't close out yet. \t {0} @ {1}, gains={2}".format(this._security.symbol,currentPrice,this._currentReturns))
pass
else:
#loosing from our peak, so close out
logger.debug("loosing from our peak, so close out. gainsPercent={0:.4f}, \t gainsFloorThreshhold={1:.4f}, \t lossesFromPeak={2:.4f}, \t maxLosses={3:.4f} \t this._currentReturns={4:.4f}".format(gainsPercent, gainsFloorThreshhold, lossesFromPeak, maxLosses,this._currentReturns))
this.targetCapitalSharePercent = 0.0
this.__currentPeakGainsDecay += (this.__currentPeakGains * maxGainsDecay)
else:
this.__currentPeakGains = 0.0
this.__currentPeakGainsDecay = 0.0
if this._currentCapitalSharePercent == this.targetCapitalSharePercent and this._currentCapitalSharePercent != 0.0:
#handle maxlosses stoplimit
if maxLosses != None and this._currentReturns < -maxLosses:
logger.debug("maxlosses stoplimit. this._currentReturns={0}, maxLosses={1}".format(this._currentReturns, maxLosses))
this.targetCapitalSharePercent = 0.0
if this.targetCapitalSharePercent == 0.0 and this._currentCapitalSharePercent != 0.0:
#record our expected PnL
this._lastRoundtripReturns = this._currentReturns
this._currentCapitalSharePercent = this.targetCapitalSharePercent
#determine value of percent
targetSharesValue = this._security.framework.context.portfolio.portfolio_value * this._currentCapitalSharePercent
targetSharesTotal = int(math.copysign(math.floor(abs(targetSharesValue / currentPrice)),targetSharesValue))
targetSharesDelta = targetSharesTotal - this._currentShares
if targetSharesTotal != 0:
if abs(targetSharesDelta / (targetSharesTotal * 1.0)) < rebalanceThreshholdPercent:
#logger.debug("{0} ORDER SKIPPED! {1} (change to small) : {2} + {3} => {4} shares".format(this.strategyName,this.security.symbol, this.currentShares, targetSharesDelta, targetSharesTotal))
#our position change was too small so we skip rebalancing
return
#do actual order
if(abs(targetSharesDelta) >= 1): #can not perform an order on less than 1 share
####cancel previous open order, if any #doesn't really work, as even when canceling, some shares may be filled so you'll be left in an uncomplete state
###lastOrder = get_order(this.lastOrderId)
###unfilled = lastOrder.amount - l
###cancel_order(this.lastOrderId)
logger.info("{0} order {1} : {2} + {3} => {4} shares \t \t decisionPrice={5} ".format(this._strategyName,this._security.symbol, this._currentShares, targetSharesDelta, targetSharesTotal,currentPrice))
this._lastOrderId = this._security.framework.tradingAlgorithm.order(this._security.sid,targetSharesDelta,None,None)
this._currentShares = targetSharesTotal
this.__lastOrderPrice = currentPrice
this._totalTrades += 1
this._security.framework._totalTrades += 1
return this._lastOrderId
else:
return 0
class Security:
isDebug = False
class QSecurity:
'''
Quantopian internal security object
If you have a reference to a security object, there are several properties that might be useful:
sid = 0 #Integer: The id of this security.
symbol = "" #String: The ticker symbol of this security.
security_name = "" #String: The full name of this security.
security_start_date = datetime.datetime() #Datetime: The date when this security first started trading.
security_end_date = datetime.datetime() #Datetime: The date when this security stopped trading (= yesterday for securities that are trading normally, because that's the last day for which we have historical price data).
'''
def __init__(this):
this.sid = 0 #Integer: The id of this security.
this.symbol = "" #String: The ticker symbol of this security.
this.security_name = "" #String: The full name of this security.
this.security_start_date = datetime.datetime(1990,1,1) #Datetime: The date when this security first started trading.
this.security_end_date = datetime.datetime(1990,1,1) #Datetime: The date when this security stopped trading (= yesterday for securities that are trading normally, because that's the last day for which we have historical price data).
def __init__(this,sid, framework):
this.sid = sid
this.isActive = False
this.framework = framework
this.security_start_date = datetime.datetime.utcfromtimestamp(0)
this.security_end_date = datetime.datetime.utcfromtimestamp(0)
this.simFrame = -1
this.security_start_price = 0.0
this.security_end_price = 0.0
#this.daily_open_price = [0.0]
#this.daily_close_price = [0.0]
this.symbol = "??? Not yet active so symbol not known"
def getCurrentPosition(this):
if this.simFrame == -1:
return Shims.Position()
return this.framework.context.portfolio.positions[this.qsec]
def update(this,qsec, data):
'''qsec is only given when it's in scope, and it can actually change each timestep
what it does:
- construct new state for this frame
- update qsec to most recent (if any)
'''
#update our tickcounter, mostly for debug
this.simFrame = this.framework.simFrame
#assert(this.simFrame >= 0,"security.update() frame not set")
#update qsec to most recent (if any) 67
this.qsec = qsec
if qsec:
this.isActive = True
this.symbol = qsec.symbol
#assert(qsec.sid == this.sid,"security.update() sids do not match")
if this.security_start_price == 0.0:
this.security_start_price = data[this.sid].close_price
this.security_end_price = data[this.sid].close_price
this.security_start_date = qsec.security_start_date
this.security_end_date = qsec.security_end_date
else:
this.isActive = False
#try:
# this.daily_close_price =
# this.framework.daily_close_price[this.qsec]
# this.daily_open_price = this.framework.daily_open_price[this.qsec]
#except:
# this.daily_close_price = []
# this.daily_open_price = []
#if len(this.daily_close_price) == 0 or len(this.daily_open_price) ==
#0:
# this.isActive = False
class FrameworkBase():
def __init__(this, context, data, maxHistoryFrames=60): #5 days of history
this.maxHistoryFrames = maxHistoryFrames
this.__isFirstTimestepRun = False
this.context = context
this.tradingAlgorithm = Shims._TradingAlgorithm_QuantopianShim() #prepopulate to allow intelisence
this.tradingAlgorithm = context.tradingAlgorithm
this.simFrame = -1 #the current timestep of the simulation
this.framesToday = -1 #number of frames executed today
this.allSecurities = {} #dictionary of all securities, including those not targeted
this.activeSecurities = {}
#stub for intelisence propigation
this.allSecurities["1"]=Security(0,this)
this.allSecurities.clear()
this.thisFrameDay = 0
this.lastFrameDay = 0
this._totalTrades = 0 #total trades we execute via all strategyPositions. note that due to partial fills, this may be less than actual trades
this.isIntradayRunDetected = False #if we are running in intraday, this will be set to true on frame 2. the 2nd bar will be in the same day as the first bar. no better way to detect unfortunately.
#for storing quantopian history
#this.daily_close_price = pandas.DataFrame()
#this.daily_open_price = pandas.DataFrame()
this._initialize(data)
pass
def ensureMinHistory(this, minFrames):
'''increases the history frames if the current is less than your required min.
this is a good way to set your history, as too much history will slow down your sim, and can crash it due to out-of-memory'''
if this.maxHistoryFrames < minFrames:
this.maxHistoryFrames = minFrames
def _initialize(this, data):
'''starts initialiation of the framework
do not override this, or any other method starting with an underscore.
methods without an underscore prefix can and should be overridden.'''
#do init here
this.initialize(data)
pass
def initialize(this, data):
'''override this to do your init'''
logger.error("You should override FrameworkBase.initialize()")
pass
def initializeFirstUpdate(this,data):
'''override this. called the first timestep, before update.
provides access to the 'data' object which normal .initialize() does not'''
logger.error("You should override FrameworkBase.initializeFirstUpdate()")
pass
def _update(this,data):
'''invoked by the tradingAlgorithm shim every update. internally we will call abstract_update_timestep_handle_data()
DO NOT OVERRIDE THIS OR ANY METHODS STARTING WITH AN UNDERSCORE
override methods without underscores.
'''
#frame updates
#this.data = data
this.simFrame+=1
this.lastFrameDay = this.thisFrameDay
this.thisFrameDay = this._getDatetime().day
#supdating our history once per day
if(this.thisFrameDay != this.lastFrameDay):
#only update this once per day, hopefully improving performance...
#this.daily_close_price = history(bar_count=180, frequency='1d',
#field='close_price')
#this.daily_open_price = history(bar_count=180, frequency='1d',
#field='open_price')
this.framesToday = 0
else:
this.framesToday += 1
this.isIntradayRunDetected = True
this.__updateSecurities(data)
if not this.__isFirstTimestepRun:
this.__isFirstTimestepRun = True
this.initializeFirstUpdate(data)
this.update(data)
pass
def update(this,data):
'''override and update your usercode here'''
logger.error("You should override FrameworkBase.update()")
pass
def __updateSecurities(this,data):
'''get all qsecs from data, then update the targetedSecurities accordingly'''
#logger.debug("FrameworkBase.__updateSecurities() start.
#allSecLength={0}".format(len(this.allSecurities)))
#convert our data into a dictionary
currentQSecs = {}
newQSecs = {}
for qsec in data:
#if online, qsec is a securities object
sid = qsec.sid
#logger.debug("FrameworkBase.__updateSecurities() first loop, found
#{0}, sid={1}. exists={2}".format(qsec,
#sid,this.allSecurities.has_key(sid) ))
currentQSecs[sid] = qsec
#determine new securities found in data
if not this.allSecurities.has_key(sid):
logger.debug("FrameworkBase.__updateSecurities() new security detected. will construct our security object for it: {0}".format(qsec))
newQSecs[sid] = qsec
#construct new Security objects for our newQSecs
for sid, qsec in newQSecs.items():
#assert(not
#this.allSecurities.has_key(sid),"frameworkBase.updateSecurities
#key does not exist")
#logger.debug("FrameworkBase.__updateSecurities() new security
#found {0}".format(qsec))
security = this._getOrCreateSecurity(qsec, data)
this.allSecurities[security.sid] = security
newQSecs.clear()
#update all security objects, giving a null qsec if one doesn't exist
#in our data dictionary
for sid, security1 in this.allSecurities.items():
qsec = currentQSecs.get(sid)
security1.update(qsec, data)
## determine active securities set.
this.activeSecurities.clear()
for sid,security2 in this.allSecurities.items():
if not security2.isActive:
#logger.debug("FrameworkBase.__updateSecurities() NOT ACTIVE
#{0}".format(security.qsec))
continue
#logger.debug("FrameworkBase.__updateSecurities() ACTIVE
#{0}".format(security.qsec))
this.activeSecurities[sid] = security2
pass
pass
def initializeSecurity(this,security, data):
'''override to do custom init logic on each security.
if you wish to use your own security, return it (it will replace the existing)'''
logger.error("You should override FrameworkBase.initializeSecurity()")
pass
def _getOrCreateSecurities(this,qsecArray, data):
'''pass in an array of quantopian sid/sec tuples (ex: [(24,sid(24)),(3113,sid(3113))])
and returns an array of unique security objects wrapping them. duplicate sids are ignored'''
securities = {}
for qsec in qsecArray:
security = this._getOrCreateSecurity(qsec, data)
securities[security.sid] = security
pass
return securities.values()
def _getOrCreateSecurity(this, qsec, data):
'''pass in a quantopian sec (ex: sid(24)) and returns our security object wrapping it
if the security object
'''
sid = qsec.sid
if this.allSecurities.has_key(sid):
return this.allSecurities[sid]
#does not exist, have to create
newSecurity = Security(sid,this)
#new, so do our framework's custom init logic on this security
maybeNewSec = this.initializeSecurity(newSecurity, data)
if maybeNewSec is not None:
#framework replaced newSec with a different sec
newSecurity = maybeNewSec
this.allSecurities[sid] = newSecurity
return newSecurity
pass
def _getDatetime(this):
'''returns current market time, using US/Eastern timezone'''
return pandas.Timestamp(pandas.Timestamp(get_datetime()).tz_convert('US/Eastern'))
#entrypoints
def handle_data(context=Shims.Context(),data=pandas.DataFrame()):
'''update method run every timestep on quantopian'''
cash = context.portfolio.cash
if cash < context.cash_low:
print 'New cash low ' + str(cash)
context.cash_low = cash
#try:
if context.firstFrame:
#'''init on our first frame'''
context.firstFrame = False
context.tradingAlgorithm = Shims._TradingAlgorithm_QuantopianShim()
context.tradingAlgorithm.context = context
context.framework = constructFramework(context,data)
context.framework._update(data)
#except Exception,e:
# print "Caught:",e
pass
class Global():
pass
global _g
_g = Global()
def initialize(context=Shims.Context()):
'''initialize method used when running on quantopian'''
context.firstFrame = True
context.cash_low = context.portfolio.cash
print 'Starting cash ' + str(context.cash_low)
_g.context = context
#context.spy = sid(8554) #SPY
########## SET UNIVERSE
#if you need set universe, do it here (note that doing this slows the algo
#considerably)
#set_universe(universe.DollarVolumeUniverse(floor_percentile=90.0,ceiling_percentile=100.0))
#context.universe = [#sid(698) #BA
# #sid(8554) #SPY
# #sid(27098) #ISE
# ####################### 9 sector etfs
# sid(19662) # XLY Consumer Discrectionary SPDR Fund
# ,sid(19656) # XLF Financial SPDR Fund
# ,sid(19658) # XLK Technology SPDR Fund
# ,sid(19655) # XLE Energy SPDR Fund
# ,sid(19661) # XLV Health Care SPRD Fund
# ,sid(19657) # XLI Industrial SPDR Fund
# ,sid(19659) # XLP Consumer Staples SPDR Fund
# ,sid(19654) # XLB Materials SPDR Fund
# ,sid(19660) # XLU Utilities SPRD Fund
# ]
#aprox 200 SPY constituents from fetcher
#fetch_csv(
# "https://googledrive.com/host/0BwZ2bMDOKaeDYWYzYzgxNDYtNmQyYi00ZDk5LWE3ZTYtODQ0ZDAzNTBkY2M4/trading/SP500-20131001.csv",
# pre_func=preview,
# date_column='date',
# universe_func=(my_universe))
#aprox 90 hardcoded SPY constituents
#my_static_universe(context)
########## COMMISSION
#use top to decrease uncertainty when testing algorithms
#set_commission(commission.PerShare(cost=0.0))
#set_commission(commission.PerShare(cost=0.005, min_trade_cost=1.00)) #IB
#fixed commission model
#set_commission(commission.PerShare(cost=0.013, min_trade_cost=1.3)) #more
#agressive...
########## SLIPPAGE
#use top to decrease uncertainty when testing algorithms
#set_slippage(slippage.FixedSlippage(spread=0.00))
#set_slippage(slippage.FixedSlippage(spread=0.01))
#set_slippage(WorstSpreadSlippage())
#set_slippage(CustomSlippage(1.0,0.0))
############# Anony values
set_commission(commission.PerTrade(cost=1.0))
set_slippage(TradeAtTheOpenSlippageModel_Simple(0.1))
############## USERCODE BELOW. EDIT BELOW THIS LINE
############## USERCODE BELOW. EDIT BELOW THIS LINE
############## USERCODE BELOW. EDIT BELOW THIS LINE
class BBTechnicalIndicators(FrameHistory):
'''technical indicators relating to bollinger bands'''
class State:
'''State recorded for each frame (minute). number of frames history we store is determined by framework.maxHistoryFrames'''
def __init__(this,parent, security, data):
this.parent = parent
this.security = security
this.history = this.parent.state
bb = this.parent.bbands_data[this.security.sid]
#will be NaN if not enough period
this.upperLimit = bb[0]
this.line = bb[1]
this.lowerLimit = bb[2]
#see https://www.tradingview.com/stock-charts-support/index.php/Bollinger_Bands_%25B_(%25B)
this.percentB = (this.security.standardIndicators.state[0].close_price - this.lowerLimit) / (this.upperLimit - this.lowerLimit)
#see https://www.tradingview.com/stock-charts-support/index.php/Bollinger_Bands_Width_(BBW)
this.bbw = (this.upperLimit - this.lowerLimit) / this.line
#track slope of %B, to determine if rising (positive) or falling (negative)
if len(this.history) > 0:
this.percentBSlope = (this.percentB - this.history[0].percentB)
this.bbwSlope = (this.bbw - this.history[0].bbw)
this.percentBPercentile = scipy.stats.percentileofscore([state.percentB for state in this.history],this.percentB,"mean") / 100.0
this.bbwPercentile = scipy.stats.percentileofscore([state.bbw for state in this.history],this.bbw,"mean") / 100.0
this.lineSlope = (this.line - this.history[0].line)
else:
this.percentBSlope = 0.0
this.bbwSlope = 0.0
this.percentBPercentile = 0.5
this.bbwPercentile = 0.5
this.lineSlope = 0.0
#track momentum
if this.percentB > 0.8:
if len(this.history) > 0:
this.upperMomentumTicks = this.history[0].upperMomentumTicks + 1
else:
this.upperMomentumTicks = 1
else:
this.upperMomentumTicks = 0
if this.percentB < 0.2:
if len(this.history) > 0:
this.lowerMomentumTicks = this.history[0].lowerMomentumTicks + 1
else:
this.lowerMomentumTicks = 1
else:
this.lowerMomentumTicks = 0
#logger.debug(" bb for {0} is {1}".format(this.security.symbol, bb));
#logger.recordNormalized("upperLimit",this.upperLimit,this.security.security_start_price)
#logger.recordNormalized("line",this.line,this.security.security_start_price)
#logger.recordNormalized("lowerLimit",this.lowerLimit,this.security.security_start_price)
#logger.record("upperBar",1.0)
#logger.record("lowerBar",0.0)
#logger.record("upperApproach",0.8)
#logger.record("lowerApproach",0.2)
#logger.record("percentB",this.percentB);
def __repr__(this):
return "{0} @ {1} BBANDS l={2:.2f} \t upper={3:.2f} lower={4:.2f}".format(this.security.symbol, this.datetime, this.line, this.upperLimit, this.lowerLimit)
def initialize(this, data):
this.bbands = ta.BBANDS(timeperiod=20,nbdevup=2, nbdevdn=2,matype=0) #output: Dictionary of sid to tuples, where each tuple is three floats: (upperLimit, line, lowerLimit).
this.bbands_data = this.bbands(data)
pass
def constructFrameState(this,data):
#logger.debug("BBTechnicalIndicators.constructFrameState")
currentState = BBTechnicalIndicators.State(this, this.parent, data)
return currentState
class StandardIndicators(FrameHistory):
'''common technical indicators that we plan to use for any/all strategies
feel free to extend this, or use as a reference for constructing specialized technical indicators'''
class State:
'''State recorded for each frame (minute). number of frames history we store is determined by framework.maxHistoryFrames'''
def __init__(this,parent, security, data):
this.parent = parent
this.security = security
this.history = this.parent.state
#preset for proper intelisence
this.datetime = datetime.datetime.now()
this.open_price = 0.0
this.close_price = 0.0
this.high = 0.0
this.low = 0.0
this.volume = 0
#this.mavg3 = 0.0
#this.mavg7 = 0.0
#this.mavg15 = 0.0
#this.mavg30 = 0.0
#this.mavg45 = 0.0
#this.mavg60 = 0.0
#this.stddev3 = 0.0
#this.stddev7 = 0.0
#this.stddev15 = 0.0
#this.stddev30 = 0.0
#this.stddev45 = 0.0
#this.stddev60 = 0.0
this.datetime = data[this.security.qsec].datetime
this.open_price = data[this.security.qsec].open_price
this.close_price = data[this.security.qsec].close_price
this.high = data[this.security.qsec].high
this.low = data[this.security.qsec].low
this.volume = data[this.security.qsec].volume
#mavg for last x minutes
#this.mavg3 = numpy.mean([state.close_price for state in
#this.history[0:3]])
#this.mavg7 = numpy.mean([state.close_price for state in
#this.history[0:7]])
#this.mavg15 = numpy.mean([state.close_price for state in
#this.history[0:15]])
#this.mavg30 = numpy.mean([state.close_price for state in
#this.history[0:30]])
#this.mavg45 = numpy.mean([state.close_price for state in
#this.history[0:45]])
#this.mavg60 = numpy.mean([state.close_price for state in
#this.history[0:60]])
#this.stddev3 = numpy.std([state.close_price for state in
#this.history[0:3]])
#this.stddev7 = numpy.std([state.close_price for state in
#this.history[0:7]])
#this.stddev15 = numpy.std([state.close_price for state in
#this.history[0:15]])
#this.stddev30 = numpy.std([state.close_price for state in
#this.history[0:30]])
#this.stddev45 = numpy.std([state.close_price for state in
#this.history[0:45]])
#this.stddev60 = numpy.std([state.close_price for state in
#this.history[0:60]])
if len(this.history) < 1:
this.returns = 0.0
this.returns_median_abs = 0.0
else:
#always returns compared to last timestep
this.returns = (this.close_price - this.history[0].close_price) / this.history[0].close_price
if len(this.history) == 1:
this.returns_median_abs = abs(this.returns)
else:
this.returns_median_abs = numpy.median([abs(state.returns) for state in this.history])
try:
#when in intraday mode, stores cumulative returns through the day
this.returns_today = data[this.security.qsec].returns()
except:
this.framework.logger.error("{0} unable to obtain returns() setting returns to zero open={1}. close = {2}".format(this.parent.qsec, this.open_price, this.close_price))
this.returns_today = 0.0
pass
#daily accumulations
if this.security.framework.thisFrameDay != this.security.framework.lastFrameDay or len(this.history) < 1:
this.open_price_today = this.open_price
#if len(this.history) < 1:
# this.open_price_yesterday = this.open_price_today
# this.close_price_yesterday = this.close_price
# this.returns_yesterday = this.returns_today
##new day, so record our start of day values
#else:
# this.open_price_yesterday =
# this.history[0].open_price_today
# this.close_price_yesterday = this.history[0].close_price
# this.returns_yesterday = this.history[0].returns_today
else:
this.open_price_today = this.history[0].open_price_today
#this.open_price_yesterday =
#this.history[0].open_price_yesterday
#this.close_price_yesterday =
#this.history[0].close_price_yesterday
#this.returns_yesterday = this.history[0].returns_yesterday
def __repr__(this):
return "{0} @ {1} c={0}".format(this.security.symbol, this.datetime, this.close_price)
def initialize(this, data):
pass
def constructFrameState(this,data):
#logger.debug("StandardTechnicalIndicators.constructFrameState")
currentState = StandardIndicators.State(this, this.parent, data)
return currentState
class DailyTechnicalIndicators(FrameHistory):
'''standard technical indicators for the entire day
for daily history. the .state[] history does not include the current day, only previous days'''
class State:
'''State recorded for each previous day. number of frames history we store is determined by framework.maxHistoryFrames'''
def __init__(this,parent, security, data):
this.parent = parent
this.security = security
this.history = this.parent.state
#preset for proper intelisence
this.datetime = datetime.datetime.now()
#setting these to default to yesterday's value so that for the
#first day of our simulation we get reasonable values
this.open_price = data[this.security.qsec].mavg(1)
this.close_price = this.open_price
#this.high = 0.0
#this.low = 0.0
#this.volume = 0
#this.mavg3 = this.open_price
#this.mavg7 = this.open_price
#this.mavg15 = this.open_price
#this.mavg30 = this.open_price
#this.mavg45 = this.open_price
#this.mavg60 = this.open_price
#this.stddev3 = this.open_price
#this.stddev7 = this.open_price
#this.stddev15 = this.open_price
#this.stddev30 = this.open_price
#this.stddev45 = this.open_price
#this.stddev60 = this.open_price
if this.security.simFrame != 0:
#assert(this.security.standardIndicators.state[1].simFrame+1 ==
#this.security.simFrame,"expect to be previous day")
this.datetime = this.security.standardIndicators.state[1].datetime
this.open_price = this.security.standardIndicators.state[1].open_price_today
this.close_price = this.security.standardIndicators.state[1].close_price
this.returns = this.security.standardIndicators.state[1].returns_today
#mavg for last x days
#this.mavg3 = data[this.security.qsec].mavg(3)
#this.mavg7 = data[this.security.qsec].mavg(7)
#this.mavg15 = data[this.security.qsec].mavg(15)
#this.mavg30 = data[this.security.qsec].mavg(30)
#this.mavg45 = data[this.security.qsec].mavg(45)
#this.mavg60 = data[this.security.qsec].mavg(60)
#this.stddev3 = data[this.security.qsec].stddev(3)
#this.stddev7 = data[this.security.qsec].stddev(7)
#this.stddev15 = data[this.security.qsec].stddev(15)
#this.stddev30 = data[this.security.qsec].stddev(30)
#this.stddev45 = data[this.security.qsec].stddev(45)
#this.stddev60 = data[this.security.qsec].stddev(60)
def __repr__(this):
return "c={0} mavg7={1} mavg30={2}".format(this.close_price,this.mavg7,this.mavg30)
def initialize(this, data):
pass
def constructFrameState(this,data):
#logger.debug("StandardTechnicalIndicators.constructFrameState")
if this.framework.thisFrameDay == this.framework.lastFrameDay:
#keep previous
currentState = None
else:
currentState = DailyTechnicalIndicators.State(this, this.parent, data)
return currentState
pass
class VolatilityBiasIndicators(FrameHistory):
''' custom indicators used by the volatility bias strategy '''
def initialize(this, data):
pass
def setWindow(this, trendPeriods, weightPeriods, triggerPeriods):
'''set the size of the window our volatilityBias cares about
trendPeriods #state.trend = percent the range is up, exponential weighted by timestep
weightPeriods #state.weight = percent the range is up, linear weighted by timestep
triggerPeriods #state.trigger = percent the range is up, linear weighted by timestep
'''
#internal variables
this.trendPeriods = trendPeriods #the max history we will care about, used for determining the value of the state.trend variable
this.weightPeriods = weightPeriods
this.triggerPeriods = triggerPeriods
this.framework.ensureMinHistory(this.trendPeriods)
this.framework.ensureMinHistory(this.weightPeriods)
this.framework.ensureMinHistory(this.triggerPeriods)
'''technical indicators relating to bollinger bands'''
class State:
'''State recorded for each frame (minute). number of frames history we store is determined by framework.maxHistoryFrames'''
def __init__(this,parent, security, data):
this.parent = parent
this.security = security
this.history = this.parent.state
this.setWeight()
def setWeight(this):
'''computes weight by taking the price range (high-low) for each timestep, and summing them based on linear weight (most recient = more weight)'''
#pretty confident in this port being accurate
security = this.security
stdState = security.standardIndicators.state
stdStateLen = len(stdState)
#set default values
if len(this.history)== 0:
this.weight = 0.0
this.trend = 0.0
this.trigger = 0.0
return
else:
#set initial value equal to previous values
this.weight = this.history[0].weight
this.trend = this.history[0].trend
this.trigger = this.history[0].trigger
#SET state.trend
#trading range linear weighted by timestep
upPortion = 0.0
downPortion = 0.0
span = this.parent.trendPeriods if stdStateLen > this.parent.trendPeriods else stdStateLen
for i in range(0,span):
if stdState[i].close_price > stdState[i].open_price:
upPortion += ((span-i) * (stdState[i].high - stdState[i].low));
else:
downPortion +=((span-i) * (stdState[i].high - stdState[i].low));
if (upPortion + downPortion > 0.0):
factor = upPortion / (upPortion + downPortion)
#trend = percent the range is up, exponential weighted by timestep
this.trend = (this.history[0].trend + factor) / 2.0
#logger.info("sec={0} \t span={1} upP={2} dwnP={3} fac={4} trend={5}".format(this.security.symbol,span,upPortion,downPortion, factor, this.trend))
#SET state.weight
if stdStateLen > this.parent.weightPeriods:
upPortion = 0.0
downPortion = 0.0
span = this.parent.weightPeriods
for i in range(0,span):
if stdState[i].close_price > stdState[i].open_price:
upPortion += ((span-i) * (stdState[i].high - stdState[i].low));
else:
downPortion +=((span-i) * (stdState[i].high - stdState[i].low));
if (upPortion + downPortion > 0.0):
factor = upPortion / (upPortion + downPortion)
#weight = percent the range is up, linear weighted by timestep
this.weight = factor
#SET state.trigger
if stdStateLen > this.parent.triggerPeriods:
upPortion = 0.0
downPortion = 0.0
span = this.parent.triggerPeriods
for i in range(0,span):
if stdState[i].close_price > stdState[i].open_price:
upPortion += ((span-i) * (stdState[i].high - stdState[i].low));
else:
downPortion +=((span-i) * (stdState[i].high - stdState[i].low));
if (upPortion + downPortion > 0.0):
factor = upPortion / (upPortion + downPortion)
#trigger = percent the range is up, linear weighted by timestep
this.trigger = factor
return this.weight
def __repr__(this):
return "{0} @ {1} VOLBIAS weight={2:.2f} \t trend={3:.2f} trigger={4:.2f}".format(this.security.symbol, this.datetime, this.weight, this.trend, this.trigger)
def constructFrameState(this,data):
#logger.debug("VolatilityBiasIndicators.constructFrameState")
currentState = VolatilityBiasIndicators.State(this, this.parent, data)
return currentState
class VolatilityBiasStrategy():
def __init__(this, framework, data):
this.framework = framework
pass
def initialize(this,data):
this.trendPeriods = 53
this.weightPeriods = 33
this.triggerPeriods = 4
#this.universe = this.framework._getOrCreateSecurities([
# sid(12915) # MDY SPDR S&P MIDCAP 400 ETF TRUST
# ,sid(19654) # XLB Materials Select Sector SPDR
# ,sid(19655) # XLE Energy Select Sector SPDR
# ,sid(19656) # XLF Financial Select Sector SPDR
# ,sid(19657)# XLI Industrial Select Sector SPDR
# ,sid(19658)#XLK Industrial Select Sector SPDR
# ,sid(19659) # XLP Consumer Staples Select Sector SPDR
# ,sid(19660)# XLU Utilities Select Sector SPDR
# ,sid(19661)# XLV Utilities Select Sector SPDR
# ,sid(19662) # XLY Consumer Discretionary Select Sector SPDR
# ,sid(25485) # AGG ISHARES CORE U.S. AGGREGATE BONDS
# ],data)
this.universe = this.framework._getOrCreateSecurities([
sid(19920) # QQQ
, sid(2174) # DIA
, sid(24705) # ISHARES MSCI EMERGING MARKETS "EEM"
, sid(22972) # ISHARES MSCI EAFE ETF "EFA"
, sid(24744) # GUGGENHEIM S&P 500 EQUAL WEIGH "RSP",
, sid(19654) # Materials Select Sector SPDR "XLB"
, sid(19655) # Energy Select Sector SPDR "XLE"
, sid(19656) # Financial Select Sector SPDR "XLF"
, sid(19657) # Industrial Select Sector SPDR "XLI"
, sid(19658) # Technology Select Sector SPDR "XLK"
, sid(19659) # Consumer Staples Select Sector SPDR "XLP"
, sid(19660) # Utilities Select Sector SPDR "XLU"
, sid(19661) # Healthcare Select Sector SPDR "XLV"
, sid(19662) # Consumer Discretionary Select Sector SPDR "XLY"
, sid(22739) # VANGUARD TOTAL STOCK MARKET ETF "VTI"
, sid(25901) # VANGUARD SMALL-CAP VALUE ETF "VBR"
, sid(25485) # ISHARES CORE U.S. AGGREGATE BONDS "AGG"
, sid(2) # Alcoa "AA"
, sid(679) # Amex "AXP"
, sid(698) # BOEING CO "BA"
, sid(700) # BANK OF AMERICA CORP "BAC"
, sid(734) # BAXTER INTERNATIONAL INC "BAX"
, sid(1267) # CATERPILLAR INC "CAT"
,sid(1900) # CISCO SYSTEMS INC "CSCO"
, sid(23112) # CHEVRON CORPORATION "CVX"
, sid(2119) # DU PONT DE NEMOURS E I &CO "DD"
, sid(2190) # WALT DISNEY CO-DISNEY COMMON "DIS"
, sid(8347) # EXXON MOBIL CORPORATION "XOM"
, sid(3149) # GENERAL ELECTRIC CO "GE"
, sid(3496) # HOME DEPOT INC "HD"
, sid(3735) # HEWLETT-PACKARD CO "HPQ"
, sid(3766) # INTL BUSINESS MACHINES CORP "IBM"
,sid(3951) # INTEL CORP "INTC"
, sid(4151) # JOHNSON AND JOHNSON "JNJ"
, sid(25006) # JPMORGAN CHASE & CO COM STK "JPM"
, sid(4283) # COCA-COLA CO "KO"
, sid(4707) # MCDONALDS CORP "MCD"
, sid(4922) # 3M COMPANY "MMM"
, sid(5029) # MERCK & CO IN C "MRK"
,sid(5061) # MICROSOFT CORP "MSFT"
, sid(5923) # PFIZER INC "PFE"
, sid(5938) # PROCTER & GAMBLE CO "PG"
, sid(6653) # AT&T INC.COM "T"
, sid(24845) # Travelers "TRV"
, sid(7792) # UNITEDHEALTH GROUP INC "UNH"
, sid(7883) # UNITED TECHNOLOGIES CORP "UTX"
, sid(21839) # VERIZON COMMUNICATIONS "VZ"
, sid(8229) # WAL-MART STORES INC "WMT"
],data)
def update(this, data):
if this.framework.simFrame < (this.weightPeriods -1):
#ensure we have our history populated
return
totalWeight = 0
entries = [] #securities we will open positions with
exits = [] #open positions we will close
securitiesToEnumerate = this.framework.activeSecurities.items()
for sid,security in securitiesToEnumerate:
enter = False
exit = False
if security.isActive==False:
continue
volBiasState = security.volBiasIndicators.state[0]
if volBiasState.weight - volBiasState.trigger >= 0.40 and volBiasState.trend > 0.50:
enter = True
elif security.volatilityBiasStrategyPosition._currentCapitalSharePercent > 0.0:
if volBiasState.trigger > 0.50 and volBiasState.trigger < 0.80:
enter = True
else:
exit = True
else:
exit = True
if enter:
totalWeight += volBiasState.weight
entries.append(security)
if exit:
exits.append(security)
pass
pass
enterCount = len(entries)
if enterCount > 0:
for security in entries:
#rebalance based on weights
security.volBiasIndicators.state[0].weight /= totalWeight
security.volatilityBiasStrategyPosition.targetCapitalSharePercent = security.volBiasIndicators.state[0].weight
for security in exits:
security.volatilityBiasStrategyPosition.targetCapitalSharePercent = 0.0
pass
for sid,security in securitiesToEnumerate:
#execute trades for this timestep
security.volatilityBiasStrategyPosition.processOrder(data)
logger.record("Weight",this.universe[0].volBiasIndicators.state[0].weight)
logger.record("Trigger",this.universe[0].volBiasIndicators.state[0].trigger)
logger.record("Trend",this.universe[0].volBiasIndicators.state[0].trend)
class ExampleFramework(FrameworkBase):
def initialize(this, data):
this.volatilityBiasStrategy = VolatilityBiasStrategy(this,data)
this.volatilityBiasStrategy.initialize(data)
#this.spy = this._getOrCreateSecurity(sid(8554), data) #SPY
pass
def initializeFirstUpdate(this, data):
#this.ensureMinHistory(360)
pass
def initializeSecurity(this,security, data):
security.standardIndicators = StandardIndicators(security,this, data)
#security.dailyIndicators = DailyTechnicalIndicators(security,this)
security.volBiasIndicators = VolatilityBiasIndicators(security,this, data)
#set our securities to use the same window sizes based on a global config
security.volBiasIndicators.setWindow(this.volatilityBiasStrategy.trendPeriods, this.volatilityBiasStrategy.weightPeriods, this.volatilityBiasStrategy.triggerPeriods)
security.volatilityBiasStrategyPosition = StrategyPosition(security,"volatilityBiasStrategyPosition")
pass
def update(this, data):
#update all security indicators
for sid,security in this.activeSecurities.items():
security.standardIndicators._update(data)
#security.dailyIndicators._update(data)
security.volBiasIndicators._update(data)
pass
this.volatilityBiasStrategy.update(data)
#logger.record("tradesD100",this._totalTrades / 100.0)
pass
############## CONFIGURATION BELOW
def constructFramework(context,data):
'''factory method to return your custom framework/trading algo'''
return ExampleFramework(context,data)
