Community

What does this error coercing to Unicode need string or buffer NoneType found mean ?

I am creating a Django webapp and inside it, I am using a custom class made by me using numpy and pandas libraries. The output of which is a numpy nd-array. On returning or running the function this error crops up every now and then. Can anyone help? I am unable to figure out what went wrong.
The code for views.py is here

from django.shortcuts import render
from django.conf import settings
from django.core.files.storage import FileSystemStorage
import numpy as np
from . import sequential_procedure as sp

# Create your views here.
def homepage(request):
    return render(request, 'sequential_analysis_app/homepage.html')

def file_upload(request):
    if request.method == 'POST' and request.FILES['input_file']:
        budget=request.POST['budget']
        input_file = request.FILES['input_file']
        fs = FileSystemStorage()
        filename = fs.save(input_file.name, input_file)
        uploaded_file_url = fs.url(filename)
        num_iterations=5
        #result = np.zeros((10, 3))[0]#optimum matrix containing both optimum sample size & respective deviation
        try:
            print sp.performSequentialAnalysis(input_file,budget)
        except Exception, e:
            result = str(e)

        return render(request, 'sequential_analysis_app/homepage.html', {
            'uploaded_file_url': uploaded_file_url,
            'result':result       
        })

    return render(request, 'sequential_analysis_app/homepage.html')

Code for the custom class is here:

    #import required libraries

from openpyxl import load_workbook
import numpy as np
import pandas as pd

def performSequentialAnalysis(input_file, budget):
    #read from excel file
    xlsx_file = pd.ExcelFile(input_file)
    #getting data stored in 'Sheet1;
    data = xlsx_file.parse('Sheet1')

    total_gene_ids = len(data['Gene'])

    #print("total_gene_ids : {}".format(total_gene_ids))
    mu = data['mu'].as_matrix() # Mean for each group 
    sigma = data['sigma'].as_matrix() # Standard Deviation for each group
    C = data['Group-weightage'].as_matrix() # Coeficients for each group
    a = data['Gene-cost'].as_matrix() # Sampling cost for each group
    A=budget
    num_iterations=1
    current_iteration=1

    denominator_matrix = np.zeros((total_gene_ids, 1))
    test_result_matrix = np.zeros((total_gene_ids, num_iterations))

    #Theoritical calulation of sample value per group
    optimum = np.zeros((total_gene_ids, 3))#optimum matrix containing both optimum sample size & respective deviation

    #print C
    #print sigma
    #print sigma[0]

    for i in range(0,total_gene_ids):
        #print (A)*(np.absolute(C[i])*sigma[i])
        num = (A)*np.absolute(C[i])*sigma[i]
        for l in range(0,total_gene_ids):
            denominator_matrix[l]=np.absolute(C[l])*(sigma[l])*(np.sqrt(a[l]*a[i]))
        optimum[i] = num/np.sum(denominator_matrix)
    #print optimum

    while current_iteration <= num_iterations:
        #print("Current iteration : {}".format(current_iteration))      
        m=2#size of pilot data for each group
        counter = 0
        pilot_data = np.zeros((1, m))
        sd_of_each_group = np.zeros((total_gene_ids, 1)) #stores sd of each group when it reaches its optimum sample size
        #pilot_data generated successfully
        numerator=0
        denominator=0
        calculated_value = np.zeros((total_gene_ids, 1))#stores num/denom value for each group after each iteration
        state = np.zeros((total_gene_ids, 1))#this vector is used to determine whether that particular group achieved optimum value(i.e. 1) or not (i.e. "0")
        #generating pilot_data
        for i in range(0,total_gene_ids):
            temp = np.random.normal(mu[i],sigma[i], m)
            #print("Generated normal rand_num : {}".format(temp))       
            #print("Pilot_data : {}".format(pilot_data))        
            if (pilot_data[0]==0).all():
                pilot_data = temp
            else:
                pilot_data = np.vstack((pilot_data, temp))

        while counter < total_gene_ids:#this counter shows #groups that have achieved optimum value
            for i in range(0,total_gene_ids):
                if state[i]==0:
                    numerator = (A)*abs(C[i])*(np.std(pilot_data[i]))
                    for l in range(0,total_gene_ids):
                        if state[l]==1:
                            #print("IF:l : {} | {} ".format(l,pilot_data)) 
                            denominator_matrix[l]=abs(C[l])*(sd_of_each_group[l])*np.sqrt(a[l]*a[i])
                        else:#if(state[l]==0)
                            #print("Else:l : {} | {}".format(l,pilot_data)) 
                            denominator_matrix[l]=abs(C[l])*(np.std(pilot_data[l]))*np.sqrt(a[l]*a[i])
                calculated_value[i] = numerator/sum(denominator_matrix)
                ###condition check
            for i in range(0,total_gene_ids):
                b = calculated_value[i]
                if (m>= b and state[i]==0):
                    state[i]=1
                    #optimum[i,1]=m
                    #optimum[i,2]=sd(pilot_data[i,])
                    test_result_matrix[i,(current_iteration-1)] = m
                    sd_of_each_group[i]=np.std(pilot_data[i])
                    counter = counter+1


            #steps to update m & pilot_data & restart process again for non-optimized groups
            m=m+1 #increamenting m
            pilot_data = np.hstack((pilot_data,np.zeros((total_gene_ids, 1))))
            for i in range(0,total_gene_ids):#updating the pilot data according to value of m
                if state[i]==0:
                    pilot_data[i,(m-1)] = np.random.normal(mu[i], sigma[i], 1)


        current_iteration = current_iteration + 1

    for i in range(0,total_gene_ids): 
        optimum[i,1] = np.mean(test_result_matrix[i])
        optimum[i,2] = np.std(test_result_matrix[i])
    #print(optimum)
    #print(test_result_matrix)

    # Create a Pandas dataframe from the data.
    df = pd.DataFrame({'Gene': data['Gene']})

    # Create a Pandas Excel writer using XlsxWriter as the engine.
    writer = pd.ExcelWriter('output.xlsx', engine='xlsxwriter')

    # Convert the dataframe to an XlsxWriter Excel object.
    df.to_excel(writer, sheet_name='Result')

    # Close the Pandas Excel writer and output the Excel file.
    writer.save()

    return 'success'

'''
def performSequentialAnalysisSimulation(input_file, budget, num_iteration):
    #read from excel file
    xlsx_file = pd.ExcelFile(input_file)
    #getting data stored in 'Sheet1;
    data = xlsx_file.parse('Sheet1')

    total_gene_ids = len(data['Gene'])

    #print("total_gene_ids : {}".format(total_gene_ids))

    mu = data['mu'].as_matrix() # Mean for each group 
    sigma = data['sigma'].as_matrix() # Standard Deviation for each group
    C = data['Group-weightage'].as_matrix() # Coeficients for each group
    a = data['Gene-cost'].as_matrix() # Sampling cost for each group
    A=budget
    current_iteration=1

    denominator_matrix = np.zeros((total_gene_ids, 1))
    test_result_matrix = np.zeros((total_gene_ids, num_iterations))

        #Theoritical calulation of sample value per group
    optimum = np.zeros((total_gene_ids, 3))#optimum matrix containing both optimum sample size & respective deviation

    #print C
    #print sigma
    #print sigma[0]

    for i in range(0,total_gene_ids):
        #print (A)*(np.absolute(C[i])*sigma[i])
        num = (A)*np.absolute(C[i])*sigma[i]
        for l in range(0,total_gene_ids):
            denominator_matrix[l]=np.absolute(C[l])*(sigma[l])*(np.sqrt(a[l]*a[i]))
        optimum[i] = num/np.sum(denominator_matrix)
    #print optimum

    while current_iteration <= num_iterations:
        print("Current iteration : {}".format(current_iteration))       
        m=2#size of pilot data for each group
        counter = 0
        pilot_data = np.zeros((1, m))
        sd_of_each_group = np.zeros((total_gene_ids, 1)) #stores sd of each group when it reaches its optimum sample size
        #pilot_data generated successfully
        numerator=0
        denominator=0
        calculated_value = np.zeros((total_gene_ids, 1))#stores num/denom value for each group after each iteration
        state = np.zeros((total_gene_ids, 1))#this vector is used to determine whether that particular group achieved optimum value(i.e. 1) or not (i.e. "0")
        #generating pilot_data
        for i in range(0,total_gene_ids):
            temp = np.random.normal(mu[i],sigma[i], m)
            #print("Generated normal rand_num : {}".format(temp))       
            #print("Pilot_data : {}".format(pilot_data))        
            if (pilot_data[0]==0).all():
                pilot_data = temp
            else:
                pilot_data = np.vstack((pilot_data, temp))

        while counter < total_gene_ids:#this counter shows #groups that have achieved optimum value
            for i in range(0,total_gene_ids):
                if state[i]==0:
                    numerator = (A)*abs(C[i])*(np.std(pilot_data[i]))
                    for l in range(0,total_gene_ids):
                        if state[l]==1:
                            #print("IF:l : {} | {} ".format(l,pilot_data)) 
                            denominator_matrix[l]=abs(C[l])*(sd_of_each_group[l])*np.sqrt(a[l]*a[i])
                        else:#if(state[l]==0)
                            #print("Else:l : {} | {}".format(l,pilot_data)) 
                            denominator_matrix[l]=abs(C[l])*(np.std(pilot_data[l]))*np.sqrt(a[l]*a[i])
                calculated_value[i] = numerator/sum(denominator_matrix)
                ###condition check
            for i in range(0,total_gene_ids):
                b = calculated_value[i]
                if (m>= b and state[i]==0):
                    state[i]=1
                    #optimum[i,1]=m
                    #optimum[i,2]=sd(pilot_data[i,])
                    test_result_matrix[i,(current_iteration-1)] = m
                    sd_of_each_group[i]=np.std(pilot_data[i])
                    counter = counter+1


            #steps to update m & pilot_data & restart process again for non-optimized groups
            m=m+1 #increamenting m
            pilot_data = np.hstack((pilot_data,np.zeros((total_gene_ids, 1))))
            for i in range(0,total_gene_ids):#updating the pilot data according to value of m
                if state[i]==0:
                    pilot_data[i,(m-1)] = np.random.normal(mu[i], sigma[i], 1)


        current_iteration = current_iteration + 1

    for i in range(0,total_gene_ids): 
        optimum[i,1] = np.mean(test_result_matrix[i])
        optimum[i,2] = np.std(test_result_matrix[i])
    print(optimum)
        #print(test_result_matrix)
    return optimum

''' 

def performSequentialAnalysisSimulation(input_file, budget, num_iteration):
    #return 'Input File:',input_file," | Budget:",budget, " | num_iteration:",num_iteration
    return input_file`