import pandas as pd
from sklearn import tree
from sklearn import ensemble
from sklearn.model_selection import train_test_split
import numpy as np
from sklearn.neighbors import KNeighborsClassifier
from sklearn.model_selection import GridSearchCV

data = pd.read_csv("C:/users/vova/desktop/files/x_train.csv")
ans = pd.read_csv("C:/users/vova/desktop/files/y_train.csv", header = None)
predictions = pd.read_csv("C:/users/vova/desktop/files/x_test.csv")


data["AttmptsPerDay"] = data["totalNumOfAttempts"]/data["numberOfDaysActuallyPlayed"]
predictions["AttmptsPerDay"] = predictions["totalNumOfAttempts"]/predictions["numberOfDaysActuallyPlayed"]


#print(data.head())
X_train, X_test, y_train, y_test = train_test_split(data, ans, test_size = 0.3, random_state = 11)

rf = ensemble.RandomForestClassifier(n_estimators=150, random_state=11)
rf.fit(X_train, y_train.values.ravel())

y_train_predict = rf.predict(X_train)
y_test_predict = rf.predict(X_test)

importances = rf.feature_importances_
indices = np.argsort(importances)[::-1]

print("Feature importances:", importances)

print(np.mean(y_test != list(y_test_predict)))
print(np.mean(y_train != list(y_train_predict)))

answer = rf.predict(predictions)
f = open("C:/users/vova/desktop/y_test.csv", "w")
for i in answer:
    f.write(str(i) + '\n')