def __init__(self, nw, order, dim=2):
self.nw = nw
self.order = order
self.dim = dim
num_states = nw.vcount()
self.P = {
state: {
action: [(action, self.reward(state, action), False)]
for action in self.nw.neighbors(state)
}
for state in range(num_states)
}
dest_idx = self.nw.vs.find(attr='d').index
for state in self.nw.neighbors(dest_idx):
self.P[state][dest_idx] = [(dest_idx, self.reward(state,
dest_idx), True)]
self.num_s = num_states
upper = self.num_s - 1
self.FB = Fourier.FourierBasis(self.dim, [[0, upper], [0, upper]],
order=self.order)
self.c_norm = np.linalg.norm(self.FB.multipliers, 2, 1)
self.c_norm[0] = 1
def histogramFromData():
''' Creates a histogram of the data stored in data.txt '''
# store data in a list
file = open('data.txt')
data = []
for line in file.readlines():
data.append(float(line))
file.close()
# model data as overlapping sine waves
import math
time = []
value = []
for point in range(pointsPerPlot):
t = point * maxTime / pointsPerPlot
v = 0.
for datum in data:
v += math.sin(t * 2 * math.pi * datum)
time.append(t)
value.append(v)
# adjust min and max
global minFreq
global maxFreq
minFreq = min(data)
maxFreq = max(data)
# perform Fourier transform on wave model
import Fourier
frequency, intensity = Fourier.Transform(time,
value,
minFreq=minFreq,
maxFreq=maxFreq)
# plot the histogram
import matplotlib.pyplot as plt
plt.plot(frequency, intensity)
plt.yticks([])
def markTimes(times, color='blue'):
for time in times:
height = intensity[0]
closest = math.fabs(frequency[0] - time)
for i in range(len(frequency)):
if math.fabs(frequency[i] - time) < closest:
closest = math.fabs(frequency[i] - time)
height = intensity[i]
plt.plot([time, time], [0, height], color=color)
## mark my times
#myTimes = [60.43,59.496,58.71,66.56,58,56.97]
#markTimes(myTimes, 'red')
## mark Kim's times
#kimsTimes = [66.831,62.72,62.47,68.69,63.94,62.65]
#markTimes(kimsTimes, 'green')
plt.show()
def paintGL(self):
#print('drawing')
GL.glClear(GL.GL_COLOR_BUFFER_BIT | GL.GL_DEPTH_BUFFER_BIT)
now = time.time() - self.startTime
GL.glLoadIdentity()
GL.glTranslate(0.0, 0.0, -5.0)
GL.glScale(2.0, 2.0, 2.0)
GL.glRotate(1000 * math.sin(now / 10), 0, 1, 0)
GL.glRotate(1000 * math.cos(now / 10), 1, 0, 0)
GL.glRotate(100 * math.tan(now / 10), 0, 0, 1)
GL.glTranslate(-0.5, -0.5, -0.5)
GL.glEnableClientState(GL.GL_VERTEX_ARRAY)
GL.glEnableClientState(GL.GL_COLOR_ARRAY)
GL.glVertexPointerf(self.cubeVtxArray)
GL.glColorPointerf(self.cubeClrArray)
GL.glDrawElementsui(GL.GL_QUADS, self.cubeIdxArray)
GL.glLoadIdentity()
GL.glBegin(GL.GL_LINES)
GL.glColor3f(0, 1, 0)
#draw the grid
for x in range(50):
#print(str((x/2)-12.5))
#horizontal
GL.glVertex3f(25, -1, ((x / 2) - 12.5))
GL.glVertex3f(-25, -1, ((x / 2) - 12.5))
#forward
GL.glVertex3f(((x / 2) - 12.5), -1, 12.5)
GL.glVertex3f(((x / 2) - 12.5), -1, -12.5)
GL.glEnd()
#draw the sound data
data = Fourier.fft_abs(self.ear.stream_read())
GL.glBegin(GL.GL_LINES)
GL.glColor3f(0, 0, 1)
#width is the world space horizontal length of the drawing
width = 2
#delta is the space in between lines
delta = width / self.ear.chunk
for x in range(len(data)):
GL.glVertex3f((delta * x - width / 2), 0.00001 * data[x] - 5, -5)
GL.glVertex3f((delta * x - width / 2), -5, -5)
GL.glEnd()
#self.drawData()
GL.glFlush()
def transform():
global filename, passes
global ui
Averaging.Averaging(filename, passes)
Fourier.Fourier(filename)
freq = []
power = []
with open(filename + '_fourier.csv', 'r') as data:
_ = data.readline()
_ = data.readline()
for i in data.readlines():
line = i.split(',')
freq.append(float(line[0]))
power.append(abs(complex(line[-1])))
ui.fourierGraph.plot(freq, power)
print("Transform finished")
plt.hist(np.array(d16nord))
plt.figure()
plt.hist(np.array(d15nord))
np.mean(d16nord)
np.mean(d15nord)
np.std(d16nord)
np.std(d15nord)
np.median(d16nord)
np.median(d15nord)
import Fourier
plt.figure()
plt.plot(np.array(var_nord16), lw = 2)
plt.plot(Fourier.fourierExtrapolation(var_nord16, 0), lw = 2, color = 'black')
data = pd.read_excel("H:/Energy Management/04. WHOLESALE/02. REPORT PORTAFOGLIO/2016/06. MI/DB_Borse_Elettriche_PER MI.xlsx", sheetname = 'DB_Dati')
data = data.set_index(data['Date'])
data = data.ix[data.index.month <= 9]
pnord = data['MGP NORD [€/MWh]']
pnord = pnord.ix[:pnord.shape[0]-1]
cnord = data['MGP CNOR [€/MWh]']
cnord = cnord.ix[:pnord.shape[0]-1]
csud = data['MGP CSUD [€/MWh]']
csud = csud.ix[:csud.shape[0]-1]
sud = data['MGP SUD [€/MWh]']
sud = sud.ix[:sud.shape[0]-1]
sici = data['MGP SICI [€/MWh]']
sici = sici.ix[:sici.shape[0]-1]
sard = data['MGP SARD [€/MWh]']
TransformedDataDFT = []
TransformedDataFFT = []
Sampels = []
TimeBeforeFFT = []
TimeAfterFFT = []
TimeDifferenceFFT = []
TimeBeforeDFT = []
TimeAfterDFT = []
TimeDifferenceDFT = []
NumberOfData = []
for i in range(5):
Sampels.append([random.randint(1, 10000) for _ in range(2**(4 + i * 3))])
NumberOfData.append(len(Sampels[i]))
for i in range(5):
TimeBeforeFFT.append(time.time())
TransformedDataFFT.append(Fourier.FFT(Sampels[i]))
TimeAfterFFT.append(time.time())
TimeDifferenceFFT.append(TimeAfterFFT[i] - TimeBeforeFFT[i])
TimeBeforeDFT.append(time.time())
TransformedDataDFT.append(Fourier.DFT(Sampels[i]))
TimeAfterDFT.append(time.time())
TimeDifferenceDFT.append(TimeAfterDFT[i] - TimeBeforeDFT[i])
MeanSquareError.append(
np.square(np.subtract(TransformedDataDFT[i],
TransformedDataFFT[i])).mean())
print(MeanSquareError)
for i in range(len(MeanSquareError)):
Error.append(abs(MeanSquareError[i]))
plt.plot(NumberOfData, Error)
plt.ylim(-2, 2)
######################################################################
dpun = np.diff(np.array(data[data.columns[12]].dropna().resample('D').mean()))
import statsmodels.api
plt.figure()
plt.plot(statsmodels.api.tsa.periodogram(dpun))
per = statsmodels.api.tsa.periodogram(dpun)
np.where(per > 50)[0]
per[per > 50]
import Fourier
reconstructed = Fourier.fourierExtrapolation(dpun, 0, 16)
plt.figure()
plt.plot(dpun)
plt.plot(reconstructed, color = 'red')
np.mean(dpun - reconstructed)
np.std(dpun - reconstructed)
from pandas.tools import plotting
plt.figure()
plotting.lag_plot(pd.DataFrame(dpun))
plt.figure()
plt.plot(statsmodels.api.tsa.acf(dpun))
scipy.stats.mstats.mquantiles(dv_p, prob=[0.025, 0.975])
scipy.stats.mstats.mquantiles(dv_f, prob=[0.025, 0.975])
###############
divp = np.diff(vol_p)
divf = np.diff(vol_f)
plt.figure()
plt.plot(divp)
plt.plot(np.array(vol_p))
plt.figure()
plt.plot(divf, color='magenta')
plt.plot(np.array(vol_f), color='black')
ddvp = np.diff(dv_p)
ddvf = np.diff(dv_f)
plt.figure()
plt.plot(ddvp)
plt.plot(np.array(dv_p))
plt.figure()
plt.plot(ddvf, color='magenta')
plt.plot(np.array(dv_f), color='black')
fdvp = Fourier.fourierExtrapolation(
dv_p, 0, 25) ### best one to see the 'right' process for the volatility
plt.figure()
plt.plot(np.array(dv_p))
plt.plot(fdvp, color='black', lw=2)
scipy.stats.mstats.mquantiles(dv_f, prob = [0.025, 0.975]) ############### divp = np.diff(vol_p) divf = np.diff(vol_f) plt.figure() plt.plot(divp) plt.plot(np.array(vol_p)) plt.figure() plt.plot(divf, color = 'magenta') plt.plot(np.array(vol_f), color = 'black') ddvp = np.diff(dv_p) ddvf = np.diff(dv_f) plt.figure() plt.plot(ddvp) plt.plot(np.array(dv_p)) plt.figure() plt.plot(ddvf, color = 'magenta') plt.plot(np.array(dv_f), color = 'black') fdvp = Fourier.fourierExtrapolation(dv_p,0, 25) ### best one to see the 'right' process for the volatility plt.figure() plt.plot(np.array(dv_p)) plt.plot(fdvp, color = 'black', lw = 2)
