def crossover(quotes, d1, d2, params={}):
# define a function that takes a quotes object, so it is an array with elements [day,o,h,l,c]
# and returns a dictionary with keys of buy and sell times
# values are not used (1 here), so that gives us room to pack in more information with this function
print d1, d2
timesclose = quotes[:,[0,4]]
p1, p2 = params["p1"], params["p2"]
p1, p2 = min(p1,p2), max(p1,p2)
if(p2 - p1 <= 1): return { }, { }
emas = [ ind.ematimes(timesclose,i) for i in [p1, p2] ]
crossovertimes = ind.crossovertimes(emas)
for t in crossovertimes:
print u.inum2tuple(t[0])
dCrossovers = {} # turn into dict for fast lookup
for time,rising in crossovertimes: dCrossovers[time] = rising
dBuy = { }
dSell = { }
hasBought = False
boughtPrice = -1.0
for i,day in enumerate(quotes[:,0]):
# only trade within specified window
if( day > u.tuple2inum(d2) or day < u.tuple2inum(d1) ): continue
if(day in dCrossovers):
rising = dCrossovers[day]
if(rising):
dBuy[day] = 1
hasBought = True
boughtPrice = quotes[i][4]
else:
if(hasBought): # don't sell if we already sold
dSell[day] = 1
hasBought = False
# if price is x% lower than what we bought for, get the hell out and sell a certain %
if( hasBought and (0.0 < quotes[i][4] / boughtPrice < 0.90) ):
dSell[day] = 1
hasBought = False
return dBuy, dSell
def crossover(quotes):
timesclose = quotes[:,[0,4]]
emas = [ ind.ematimes(timesclose,i) for i in [5, 10] ]
crossovertimes = ind.crossovertimes(emas)
dCrossovers = {} # turn into dict for fast lookup
for time,rising in crossovertimes: dCrossovers[time] = rising
dBuy = { }
hasBought = False
boughtPrice = -1.0
for i,day in enumerate(quotes[:,0]):
#print day
if(day in dCrossovers):
rising = dCrossovers[day]
if(rising):
dBuy[day] = 1
hasBought = True
boughtPrice = quotes[i][4]
return dBuy
import numpy as np
import getStocks as gs
import utils as u
import indicators as ind
import tradeReport as tr
#symbol = "WMT"
symbol = "NFLX"
# symbol = "F"
stock = gs.getStock(symbol, (2011, 6, 1), (2015, 6, 1)) # for calculating
quotes = u.dictToList(stock) # [day,o,h,l,c]
timesclose = quotes[:,[0,4]]
emas = []
for i in [1,2]: emas.append( ind.ematimes(timesclose,10*i) )
crossovertimes = ind.crossovertimes(emas)
dCrossovers = {} # turn into dict for fast lookup
for time,rising in crossovertimes: dCrossovers[time] = rising
# CROSSOVER STRATEGY
price = 0
ledger = tr.Ledger(1000)
for quote in quotes:
day,price,h,l,c = quote
if(day in dCrossovers):
rising = dCrossovers[day]
if(rising): ledger.buyStock(symbol, price)
stock = gs.getStock("F", (2010, 1, 1), (2011, 10, 5))["days"] # for calculating
d1,d2 = (2010,8,1),(2011,3,5) # for plotting
quotes = []
timesclose = []
for day in sorted(stock.keys()):
vals = stock[day]
o, h, l, c = vals["o"], vals["h"], vals["l"], vals["c"]
quotes.append( [day,o,h,l,c] )
timesclose.append([day,c])
timesclose = np.array(timesclose)
bbands = ind.bbtimes(timesclose, 20)
emas = [ ind.ematimes(timesclose,i) for i in [5,10] ]
quotes = np.array(quotes)
rsis = ind.rsitimes(quotes[:,[0,1]],14)
# bbands discards early dates since it can't do a moving average for those
# we reshape the other arrays to match up with bbands' size
quotes = quotes[-len(bbands):]
nclusters=5
ncandles=10 # remember to change to "saveTo" below when you change these values
times, clusters, clusterCenters = cl.clusterCandles(quotes, nclusters=nclusters, ncandles=ncandles, saveTo="test.txt")
clusterQuotes, vlines = cl.clusterCentersForPlotting(clusterCenters)
clustersAndProfits = cl.clustersAndTheirNextProfits(clusters,clusterCenters)
