class ButterflyHistogram():
def __init__(self, strike_intervals):
self.intervals = strike_intervals
self.prices = [0 for x in self.intervals][:-1]
self.price_sum = 0
self.probs = [1.0/len(self.intervals) for x in self.intervals]
self.gpbase = GNUPlotBase()
self.gpbase.xmin = strike_intervals[0][0]*.9
self.gpbase.xmax = strike_intervals[-1][-1]*1.1
self.gpbase.ymin = 0
self.gpbase.ymax = 1
self.datafile = 'butterhist.dat'
self.line = '%0.2f %0.3f\n'
self.xticks = [x[0] for x in strike_intervals
if strike_intervals.index(x) % 2 == 0]
self.xticks.append(strike_intervals[-1][-1])
self.yticks = [x/10.0 for x in range(1, 10)]
def update_price(self, price, index):
if price > 0: self.prices[index] = price
else: self.prices[index] = 0
self.calculate_probabilities()
def sum_prices(self):
self.price_sum = 0
for x in self.prices: self.price_sum += x
def calculate_probabilities(self):
self.sum_prices()
if self.price_sum == 0:
f = open(self.datafile, 'w')
f.write('0 0\n')
f.close()
return None
f = open(self.datafile, 'w')
self.probs[0] = 0.5*self.prices[0]/self.price_sum
f.write(self.line % (self.mid(self.intervals[0]), self.probs[0]))
for i in range(1, len(self.probs)-1):
self.probs[i] = self.mid(self.prices[i-1:i+1])/self.price_sum
f.write(self.line % (self.mid(self.intervals[i]), self.probs[i]))
self.probs[-1] = 0.5*self.prices[-1]/self.price_sum
f.write(self.line % (self.mid(self.intervals[-1]), self.probs[-1]))
f.close()
def mid(self, interval):
return (interval[1] + interval[0])/2.0
def plot_histogram(self, fname=None, timestamp=None):
commands = []
commands.append(self.gpbase.set_output(fname=fname))
commands.append(self.gpbase.gen_ticks(self.xticks, self.yticks))
commands.append(self.gpbase.gen_header(ylabel='Probability',
timestamp=timestamp))
commands.append('plot "%s" with boxes' % self.datafile)
return '\n'.join(commands)
if __name__ == "__main__":
datfname = sys.argv[1]
jpgfname = sys.argv[2]
start = sys.argv[3]
end = sys.argv[4]
increment = sys.argv[5]
dt = datfname.split('.dat')[0].split('_')[2]
if float(increment) % 1 == 0:
start = int(start)
end = int(end)
increment = int(increment)
xtics = range(start, end+increment, increment)
elif float(increment) == 0.5:
start = int(float(start)*10.0)
end = int(float(end)*10.0)
increment = 5
strikes = range(start, end+increment, increment)
xticks = [x/10.0 for x in strikes]
yticks = [x/10.0 for x in range(0, 11)]
gpb = GNUPlotBase()
gpb.xmin = start*.9
gpb.xmax = end*1.1
gpb.ymin = 0
gpb.ymax = 1
commands = []
commands.append(gpb.set_output(fname=jpgfname))
commands.append(gpb.gen_header(ylabel='Probability', timestamp=dt))
commands.append(gpb.gen_ticks(xticks, yticks))
commands.append('plot "%s" with boxes' % datfname)
print '\n'.join(commands)