import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

def abline(slope, intercept):
    """Plot a line from slope and intercept"""
    axes = plt.gca()
    x_vals = np.array(axes.get_xlim())
    y_vals = intercept + slope * x_vals
    plt.plot(x_vals, y_vals, '--')



df=pd.read_csv('ccMobTrial4.CSV', sep=',',names=['L','p','Fail','Samples'])
dfcoll = (pd.DataFrame(columns=['L','p','Fail','Samples', 'LogicalFailureRate']))
uniqueLValues = df['L'].unique()
uniquepValues = df['p'].unique()
for L in uniqueLValues:
	dftemp = df[df['L'] == L]
	for p in uniquepValues:
		dftemp2 = dftemp[dftemp['p'] == p]
		failval = dftemp2['Fail'].sum()
		sampeval = dftemp2['Samples'].sum()
		dfcoll = dfcoll.append({'L' : L, 'p' : p, 'Fail' : failval, 'Samples' : sampeval, 'LogicalFailureRate' : failval/sampeval}, ignore_index = True)
#print(dfcoll)


fig = plt.figure()
ps = []
alphas = []
gl = []
for L in uniqueLValues:
	if p > 0.00:
		frame = dfcoll[dfcoll['L'] == L]
		frame = frame[frame['p'] < 0.05]
		#plt.scatter (frame['L'], np.log(frame['LogicalFailureRate']))
		slope, intercept = np.polyfit(np.log(frame['p']), np.log(frame['LogicalFailureRate']), 1)
		#print(p, ":" , slope/np.log(p))
		#ps.append((np.log(p)))
		#alphas.append(slope)
		gl.append(slope)
plt.scatter(np.log(uniqueLValues), np.log(gl))
gamma, logalpha = np.polyfit(np.log(uniqueLValues), np.log(gl), 1)
print("gamma = ", gamma)
print("logalpha = " , (logalpha))
abline(gamma, logalpha)
plt.show()