# Libraries

import pandas as pd
import numpy as np

from sklearn.datasets import make_classification
from sklearn.linear_model import LogisticRegression, LinearRegression
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score, confusion_matrix, classification_report, roc_auc_score, roc_curve
from sklearn.metrics import log_loss

import matplotlib.pyplot as plt

# classifier = LogisticRegression()
# classifier.fit(x, y)

# Generate a random n-class classification problem
n = 10000
X, y = make_classification(n_samples=n, n_features=1, n_informative=1, n_redundant=0, n_clusters_per_class=1)

# Create a dataframe of the feature and class
df = pd.DataFrame({'Feature': X.flatten(), 'Class': y.flatten()})
print('Dataframe Head')
print(df.head())
print()

print('Model Summary')
import statsmodels.api as sm
logit=sm.Logit(df['Class'], df['Feature'])
result=logit.fit()
print(result.summary2())

X = df.drop('Class', axis=1)
y = df['Class']

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20)

model = LogisticRegression()
model.fit(X_train,y_train)
predictions = model.predict(X_test)

print('Test size')
print(len(predictions))
print()

print('Iterations: ', model.n_iter_)
print('Intercept: ', model.intercept_)
print('Coefficient: ', model.coef_)
print()

print('Accuracy Score')
print(accuracy_score(ny, predictions))
print()

print('Confustion matrix')
print(confusion_matrix(ny,predictions))
tn, fp, fn, tp = confusion_matrix(ny,predictions).ravel()
print('tn: {} fp: {} fn: {} tp: {}'.format(tn, fp, fn, tp))
print()

print('Classification Report')
print(classification_report(ny,predictions))
print()

print('Misclassified count and location')
misclassified = np.flatnonzero(ny != predictions)
print(len(misclassified), misclassified)
print()
print(df.loc[misclassified,:])

# https://stackoverflow.com/questions/48185090/how-to-get-the-log-likelihood-for-a-logistic-regression-model-in-sklearn
# Every quantity described as "loss", implies "the lower the better"

# http://scikit-learn.org/stable/modules/generated/sklearn.metrics.log_loss.html
# -log P(yt|yp) = -(yt log(yp) + (1 - yt) log(1 - yp))

print(log_loss(y_test, predictions))
log_likelihood_elements = y_test*np.log(predictions) + (1-y_test)*np.log(1-predictions)
print(log_likelihood_elements)
print(-np.sum(log_likelihood_elements)/len(y_test))

# https://towardsdatascience.com/building-a-logistic-regression-in-python-step-by-step-becd4d56c9c8
logit_roc_auc = roc_auc_score(y_test, predictions)
fpr, tpr, thresholds = roc_curve(y_test, model.predict_proba(X_test)[:,1])
plt.figure()
plt.plot(fpr, tpr, label='Logistic Regression (area = %0.2f)' % logit_roc_auc)
plt.plot([0, 1], [0, 1],'r--')
plt.xlim([0.0, 1.0])
plt.ylim([0.0, 1.05])
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.title('Receiver operating characteristic')
plt.legend(loc="lower right")
plt.show()

# http://scikit-learn.org/stable/auto_examples/linear_model/plot_logistic.html#sphx-glr-auto-examples-linear-model-plot-logistic-py
# Code source: Gael Varoquaux
# License: BSD 3 clause

# Plot the logistic and linear models
plt.figure(figsize=(12, 4))
plt.clf()
plt.scatter(df['Feature'], df['Class'], color='black', s=2)
X_line = np.linspace(-3, 3, 100)

def model_func(x):
    return 1 / (1 + np.exp(-x))

loss = model_func(X_line * model.coef_ + model.intercept_).ravel()
plt.plot(X_line, loss, color='red', linewidth=1)

ols = LinearRegression()
ols.fit(X_train,y_train)
plt.plot(X_line, ols.coef_ * X_line + ols.intercept_, linewidth=1)
plt.axhline(.5, color='.5')

plt.ylabel('Class')
plt.xlabel('Feature')
plt.xticks(range(-3, 3))
plt.yticks([0, 0.5, 1])
plt.legend(('Logistic Regression Model', 'Linear Regression Model'), loc="lower right")
plt.tight_layout()
plt.show()