'''

Created on Jun 8, 2015

@author: jkarnuta

'''

import os

import numpy as np

from scipy.optimize import curve_fit

import re

from csvReader import csvReader

import matplotlib.pyplot as plt

from FileWriter import FileWriter

from PathMethods import figuresFileName

import sys

vi = sys.version_info

if vi[0] != 2 or vi[1] != 6:

print "ERROR: python interpreter of wrong version. Requires 2.6.x"

raise SystemExit

print sys.argv

##-------------------------------------------

from YingRongSpikeInMetArray import DataArray

filepath = sys.argv[1]

denatsFile = sys.argv[2]

pictureFolder = figuresFileName(filepath)

fw = FileWriter(filepath, customFileName="")

##-------------------------------------------

#If a third argument is passed, generate figures

#Figure generation is slow, so it is often optimal to disregard generation

#as fitting is <5 second operation

runFigs = False

try:

sys.argv[3]

runFigs = True

#Check if pictureFolder exists, if not mkdir

if not os.path.exists(pictureFolder):

os.makedirs(pictureFolder)

except IndexError:

pass

denaturantsArray = csvReader(denatsFile).getHeader()

numberDenaturants = len(denaturantsArray)

#Populate data array with data from csv input file

data = DataArray(filepath,numberDenaturants)

#Set up denaturant concentrations from denatsFile (sys.argv[2]

data.setDenConcs(denaturantsArray)

xdata = np.array(data.denConcs)

#Changes the intensities based on the midpointTolerance. To not change intensities, return

def alterIntensities(intensities):

return intensities

#Make custom header (if there is a new line, remove)

newHeader = data.header

newHeader[-1] = re.sub(r"\n", "",newHeader[-1])

newHeader.append("A")

newHeader.append("B")

newHeader.append("dGf")

newHeader.append("sd dGf")

newHeader.append("m")

newHeader.append("sd m")

newHeader.append("C 1/2")

newHeader.append("sd C 1/2")

newHeader.append("b")

newHeader.append("sd b")

newHeader[-1] = newHeader[-1]+"\n"

#Write header to file

fw.writeList(data.header)

def makeModel(x, dGf, m):

return B+(A-B)*np.exp(-(kox*t)/Kfold)

def makeChalfModel(x, chalf, b):

return A + (B-A)/(1+np.exp(-(x-chalf)/(b)))

allattempts = []

failedOnRuntime = []

failedOnCovMatrix = []

numSuccess = 0

numRuntimeError_dGf_m = 0

numCovINF_dGf_m = 0

numRuntimeError_chalf_b = 0

numCovINF_chalf_b = 0

i=0

for i in range(data.size):

##-----------------------------------

ydata = np.array(data.data[i].control.denats)

##-----------------------------------

sequence = data.data[i].sequence

protein = data.data[i].protein

score = str(data.data[i].control.intsum)

rtmin = str(data.data[i].control.rtmin)

#A = ydata[0]

#B = ydata[len(ydata)-1]

A = max(ydata)

B = min(ydata)

ydata = alterIntensities(ydata)

message = [sequence, protein, score, rtmin]

for ele in ydata:

message.append(str(ele))

message.append(str(A))

message.append(str(B))

try:

#use either 3 or 4 parameter curve_fit

try:

popt, pcov = curve_fit(makeModel, xdata, ydata)

if not isinstance(pcov, np.ndarray):

raise RuntimeError

except RuntimeError:

popt, pcov = curve_fit(makeModel, xdata, ydata,0)

except RuntimeError:

print "Max recursion depth exceeded on dGf, m"

print "Could not find fit for RUN: "+str(i+1)

print "Sequence: "+sequence

print "Protein: "+protein

print data.data[i].control.toString()

print ""

numRuntimeError_dGf_m +=1

message.append("NaN")

message.append("NaN")

message.append("NaN")

message.append("NaN")

message.append("NaN")

message.append("NaN")

message.append("NaN")

message.append("NaN")

failedOnRuntime.append(str(i+2))

allattempts.append(message)

message[-1] = message[-1]+"\n"

fw.writeList(message)

continue

if not isinstance(pcov, np.ndarray):

print "Covariance matrix is INF on dGf, m"

print "Could not find fit for RUN: "+str(i+1)

print "Sequence: "+sequence

print "Protein: "+protein

print data.data[i].control.toString()

print ""

numCovINF_dGf_m +=1

message.append("NaN")

message.append("NaN")

message.append("NaN")

message.append("NaN")

message.append("NaN")

message.append("NaN")

message.append("NaN")

message.append("NaN")

failedOnCovMatrix.append(str(i+2))

allattempts.append(message)

message[-1] = message[-1]+"\n"

fw.writeList(message)

continue

perr = np.sqrt(np.diag(pcov))

dGf_val = str(popt[0])

m_val = str(popt[1])

dGf_err = str(perr[0])

m_err = str(perr[1])

message.append(dGf_val)

message.append(dGf_err)

message.append(m_val)

message.append(m_err)

try:

try:

ch, ch_cov = curve_fit(makeChalfModel, xdata, ydata)

if not isinstance(pcov, np.ndarray):

raise RuntimeError

except RuntimeError:

ch, ch_cov = curve_fit(makeChalfModel, xdata, ydata, p0=[1,0.592154/float(m_val)])

except RuntimeError:

print "Max recursion depth exceeded on C 1/2"

print "Could not find fit for RUN: "+str(i+1)

print "Sequence: "+sequence

print "Protein: "+protein

print data.data[i].control.toString()

print ""

numRuntimeError_chalf_b +=1

message.append("NaN")

message.append("NaN")

message.append("NaN")

message.append("NaN")

failedOnRuntime.append(

str(i+2))

allattempts.append(message)

message[-1] = message[-1]+"\n"

fw.writeList(message)

continue

if not isinstance(ch_cov, np.ndarray):

print "Covariance matrix is INF on C 1/2"

print "Could not find fit for RUN: "+str(i+1)

print "Sequence: "+sequence

print "Protein: "+protein

print data.data[i].control.toString()

print ""

numCovINF_chalf_b +=1

message.append("NaN")

message.append("NaN")

message.append("NaN")

message.append("NaN")

failedOnCovMatrix.append(str(i+2))

allattempts.append(message)

message[-1] = message[-1]+"\n"

fw.writeList(message)

continue

chalf_val = str(ch[0])

chalf_err = str(np.sqrt(np.diag(ch_cov))[0])

b_val = str(ch[1])

b_err = str(np.sqrt(np.diag(ch_cov))[1])

message.append(chalf_val)

message.append(chalf_err)

message.append(b_val)

message.append(b_err)

message[-1] = message[-1]+"\n"

allattempts.append(message)

fw.writeList(message)

print "Run #"+str(i+1)

print "dGf: "+dGf_val+" +/- "+dGf_err

print " m: "+m_val+" +/- "+m_err

print "C12: "+chalf_val+"+/-"+chalf_err

print " b: "+b_val+"+/-"+b_err

print ""

numSuccess += 1

fw.close()

if runFigs:

for i in range(data.size):

numberToView = i+2

if numberToView == -1:

raise SystemExit

print "Drawing Fig"+str(numberToView)+".png"

plottedYData = [float(x) for x in allattempts[numberToView - 2][4:4+numberDenaturants]]

A = float(allattempts[numberToView-2][4+numberDenaturants])

B = float(allattempts[numberToView-2][5+numberDenaturants])

m_val = float(allattempts[numberToView-2][8+numberDenaturants])

plottedXData = np.linspace(xdata[0], xdata[-1], num = 100)

params = (float(allattempts[numberToView-2][6+numberDenaturants]), m_val)

chalf_val = float(allattempts[numberToView-2][10+numberDenaturants])

ylim =[min(plottedYData)*0.8, max(plottedYData)*1.2]

fittedModel = makeModel(plottedXData,*params)

chalfModel= makeChalfModel(plottedXData, float(chalf_val), float(b_val))

plt.figure()

plt.title(allattempts[numberToView-2][0])

plt.scatter(xdata, plottedYData, label = "Data", marker = 'o', alpha = 0.5)

plt.plot(plottedXData, fittedModel ,"r",label = "dGf, m model")

plt.plot(plottedXData, chalfModel, "g", label = "C 1/2, b Model")

plt.legend()

plt.ylim(ylim)

plt.savefig(pictureFolder +"Fig"+str(numberToView)+".png",bbox_inches='tight')

print "finished rasterizing images"

else:

"Figures not being generated for this run."

print ""

print "Run Statistics: "

print "Number Successful: "+str(numSuccess) + " ("+str((float(numSuccess)/data.size)*100)+"%)"

print "Number Failed at Runtime on dGf, m: "+str(numRuntimeError_dGf_m)

print "Number with INF Covariance Matrix on dGf, m: "+str(numCovINF_dGf_m)

print "Number Failed at Runtime on chalf, b: "+str(numRuntimeError_chalf_b)

print "Number with INF Covariance Matrix on chalf, b: "+str(numCovINF_chalf_b)

print ""

print "Failed on runtime: "+str(failedOnRuntime)

print ""

print "Failed on Cov Matrix: "+str(failedOnCovMatrix)