import numpy as np
import pandas as pd
import matplotlib as mpl
import matplotlib.pyplot as plt
import random
from scipy.stats import rankdata,mannwhitneyu, kruskal
from numpy import sort, size, sqrt, average
import scikit_posthocs as sc



# Importing the dataset
data = pd.read_csv('~/Desktop/michal6.csv')
print("Input Data and Shape")
print(data.shape)
data.head()

# Division of data into data sets
is_guitar = data['Gitarzysta'].tolist()
values = data['M'].tolist()
values2 = data['Technika'].tolist()
Y_T_A = []
for k, v, v2 in zip(is_guitar, values, values2):
    if k == 'Tak' and v2 == 'Tirando':
        Y_T_A.append(v)

is_guitar = data['Gitarzysta'].tolist()
values = data['M'].tolist()
values2 = data['Technika'].tolist()
Y_A_A = []
for k, v, v2 in zip(is_guitar, values, values2):
    if k == 'Tak' and v2 == 'Apoyando':
        Y_A_A.append(v)

is_guitar = data['Gitarzysta'].tolist()
values = data['M'].tolist()
values2 = data['Technika'].tolist()
N_T_A = []
for k, v, v2 in zip(is_guitar, values, values2):
    if k == 'Nie' and v2 == 'Tirando':
        N_T_A.append(v)

is_guitar = data['Gitarzysta'].tolist()
values = data['M'].tolist()
values2 = data['Technika'].tolist()
N_A_A = []
for k, v, v2 in zip(is_guitar, values, values2):
    if k == 'Nie' and v2 == 'Apoyando':
        N_A_A.append(v)


# Preparation for rank assingment

total=Y_T_A+Y_A_A+N_T_A+N_A_A
total2=total
total=sort(total)
ranks = rankdata(total)


# Separation of ranks between categories of interest
Aranks=[]
Branks=[]
Cranks=[]
Dranks=[]
j=0
while j<size(total):
    if total[j] in Y_T_A:
        Aranks.append(ranks[j])
    if total[j] in Y_A_A:
        Branks.append(ranks[j])
    if total[j] in N_T_A:
        Cranks.append(ranks[j])
    if total[j] in N_A_A:
        Dranks.append(ranks[j])
    j = j+1





y={'1': [Y_T_A], '2': [Y_A_A]}
x = pd.DataFrame(y)
x = x.melt(var_name='groups', value_name='values')
print x