"""Statistické testování průměrné délky slova v textech: 
normalita rozdělení a její ověření, t-test, Mann-Whitney-Wilcox test"""
# je treba nainstalovat scipy a matplotlib
# py  -m pip install scipy


# text_01
with open('texty/text_01.txt', encoding='UTF-8') as f:
    obsah = f.read()

text_bez_interpunkce = obsah.translate(str.maketrans({',':'', '.':'', '!':'', '?':'', '(':'', ')':'', '„':'', '“':'', '–':'' }))
# print(text_bez_interpunkce)

text_bez_interpunkce_lc = text_bez_interpunkce.lower()
# print(text_bez_interpunkce_lc)

text_01 = text_bez_interpunkce_lc.split() 
print(text_01)


# text_02
with open('texty/text_02.txt', encoding='UTF-8') as f:
    obsah = f.read()

text_bez_interpunkce = obsah.translate(str.maketrans({',':'', '.':'', '!':'', '?':'', '(':'', ')':'', '„':'', '“':'', '–':'' }))
# print(text_bez_interpunkce)

text_bez_interpunkce_lc = text_bez_interpunkce.lower()
# print(text_bez_interpunkce_lc)
text_02 = text_bez_interpunkce_lc.split() 
print(text_02)


# delky slov v poctu pismen
delky_slov_01 = []
for slovo in text_01:
    delky_slov_01.append(len(slovo))
print(delky_slov_01)

delky_slov_02 = []
for slovo in text_02:
    delky_slov_02.append(len(slovo))
print(delky_slov_02)

# vypocitame prumer, median, modus a SD u obou textu

import statistics
import numpy


# pro tokeny
statistics.mean(delky_slov_01)
statistics.stdev(delky_slov_01)
statistics.median(delky_slov_01)
statistics.mode(delky_slov_01)


mean_L_text_01 = statistics.mean(delky_slov_01)
mean_L_text_02 = statistics.mean(delky_slov_02)
median_L_text_01 = statistics.median(delky_slov_01)
median_L_text_02 = statistics.median(delky_slov_02)
mode_L_text_01 = statistics.mode(delky_slov_01)
mode_L_text_02 = statistics.mode(delky_slov_02)
sd_L_text_01 = numpy.std(delky_slov_01)
sd_L_text_02 = numpy.std(delky_slov_02)




# testovani noramlity rozdeleni, Shapiro test
from scipy import stats

# HO rika, ze data jsou normalne rozdelena, pokud p-value menzi ne 0.05, zamitame HO
shapiro_01 = stats.shapiro(delky_slov_01)
print(shapiro_01)
# k samotnemu p-value a statisticke hodnote se dostaneme takto
shapiro_01.pvalue
shapiro_01.statistic

shapiro_02 = stats.shapiro(delky_slov_02)
print(shapiro_02)
shapiro_02.pvalue

# pro grafickou jednoduchou kontrolu
# Quantile-Quantile Plot
# https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.probplot.html

import pylab 
stats.probplot(delky_slov_02, dist="norm", plot=pylab)
pylab.show()

# histogram
import matplotlib.pyplot as plt
# https://matplotlib.org/stable/api/_as_gen/matplotlib.pyplot.hist.html
plt.hist(delky_slov_02)
plt.show()


#ukazeme si, jak by to vypadalo na datech z normalni distribuce
import numpy
# genereujeme si normalne rozdelena data, loc je prumer, scale je SD
# https://numpy.org/doc/stable/reference/random/generated/numpy.random.normal.html
normdata = numpy.random.normal(loc=5, scale=2, size=100)

stats.shapiro(normdata)

stats.probplot(normdata, dist="norm", plot=pylab)
pylab.show()

plt.hist(normdata)
plt.show()


# Mann-Whitney U test
# https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.mannwhitneyu.html
mwwu_test = stats.mannwhitneyu(delky_slov_01, delky_slov_02)
print(mwwu_test)
mwwu_test.pvalue

# pokud data normalne rozdelena - t-test (Two sample t-test (unpaired or independent t-test))
# https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.ttest_ind.html
stats.ttest_ind(delky_slov_01, delky_slov_02)



# obrazky
import matplotlib.pyplot
import numpy


# barplot
from collections import Counter

freq_lengths_01 = dict(Counter(delky_slov_01))
print(freq_lengths_01)

lengths = list(freq_lengths_01.keys())
frequencies = list(freq_lengths_01.values())

plt.bar(lengths, frequencies)

matplotlib.pyplot.xlabel('délky')
matplotlib.pyplot.ylabel('frekvence')

matplotlib.pyplot.show()




# scatterplot
matplotlib.pyplot.scatter(lengths, frequencies)
matplotlib.pyplot.show()


# boxplot

# vytvore obrazku
fig, axes = matplotlib.pyplot.subplots(nrows= 1, ncols= 2)

axes[0].boxplot(delky_slov_01)
axes[0].set_title('text_01')
axes[0].set_ylabel('délka')
axes[0].set_yticks(range(0,26,5))


axes[1].boxplot(delky_slov_02)
axes[1].set_title('text_02')
axes[1].set_yticks(range(0,26,5))

matplotlib.pyplot.show()



# UKOL: otestujte rozdily delek slov pro typy