ridgeRegression

from algorithmClass import Algorithm
import sklearn
from sklearn.linear_model import LinearRegression, Ridge, Lasso
from sklearn.preprocessing import scale
import numpy as np
from sklearn.metrics import mean_squared_error, make_scorer, explained_variance_score, mean_absolute_error
import mysql.connector
import pandas as pd
from sklearn.decomposition import PCA
from sklearn.preprocessing import normalize
import pickle

class RidgeRegressor(Algorithm):
    def split_data(self):
        return

    def split_features(self):
        self.features = self.data
        
        

        # extract and remove the fanduel points (response variable)
        self.target = [features.pop(14) for features in self.features]

        for i in range(12):
            temp = [features.pop(0) for features in self.features]

        self.features = scale(self.features)

    def run_pca(self):
        #applies Principal Component Analysis on data with 201 components, returns dimension-reduced training set

        filename = 'PCA_model.sav'
        pca = pickle.load(open(filename, 'rb'))
        pca.fit(self.features)
        self.features = pca.transform(self.features)
        print(len(self.features[0]))

    def predict(self):
        self.split_features()
        # self.run_pca()
        filename = 'Ridge100_model.sav'
        self.model = pickle.load(open(filename, 'rb'))
        self.predictions = self.model.predict(
            self.features)  # predict for features
        return self.predictions

    def mse(self):
        return mean_squared_error(self.target, self.predictions)

    def compare(self):
        for i in range(len(self.predictions)):
            print self.predictions[i], self.target[i]
    







def main():

    cnx = mysql.connector.connect(user="wsa",
                                  host="34.68.250.121",
                                  database="NCAAWomens",
                                  password="LeBron>MJ!")
    cursor = cnx.cursor(buffered=True)

    getStuff = []
    getStuff = "SELECT * from performancePlayer left join teamVsPosition on ((performancePlayer.positionID = teamVsPosition.positionID) and (performancePlayer.opponent = teamVsPosition.team) and (performancePlayer.seasonID = teamVsPosition.seasonID)) left join playerBasicAverages on ((performancePlayer.playerID = playerBasicAverages.playerID) and (performancePlayer.seasonID = playerBasicAverages.seasonID)) left join playerAdvancedAverages on ((performancePlayer.playerID = playerAdvancedAverages.playerID) and (performancePlayer.seasonID = playerAdvancedAverages.seasonID)) where performancePlayer.opponentTeamID is not null and performancePlayer.opponentTeamID < 15 and performancePlayer.minutes > 5 and playerAdvancedAverages.gamesPlayed > 5 and playerBasicAverages.minutes > 5"
    cursor.execute(getStuff)

    dataSet = [list(feature) for feature in cursor.fetchall()] 
    print dataSet[0]
    print len(dataSet)
    print len(dataSet[0])
    
    ridge = RidgeRegressor(dataSet)
    predictions = ridge.predict()
    ridge.compare()
    print predictions
    print ridge.mse() 
    
    
    return



if __name__=="__main__":
    main()