#importing
##data tools
import pandas as pd
pd.set_option('display.max_columns', 500)
import matplotlib.pyplot as plt
import matplotlib.colors as mcolors
##ML tools
from sklearn.ensemble import RandomForestClassifier
from sklearn import svm
from sklearn.svm import SVC
from sklearn.neural_network import MLPClassifier
from sklearn.metrics import confusion_matrix, classification_report
from sklearn.preprocessing import StandardScaler, LabelEncoder
from sklearn.model_selection import train_test_split
%matplotlib inline

coffeeRaw = pd.read_csv('arabica_data_cleaned.csv')

coffeeRaw.head()

coffee = coffeeRaw[['Aroma','Flavor', 'Aftertaste', 'Acidity', 'Body', 'Balance',
                     'Total.Cup.Points']].copy()
coffee.head()

coffee.info()
coffee.isnull().sum()

#preprocessing into Good/Bad coffee
bins = [-1, 85, 100] #anything scoring 85 or above is good coffee™
labels = ['bad', 'good']
coffee['goodCoffee'] = pd.cut(coffee['Total.Cup.Points'], bins = bins, labels = labels)
coffee = coffee.drop(columns=['Total.Cup.Points'])
coffee['goodCoffee'].unique()

coffee['goodCoffee'].value_counts()

#transform 'good' and 'bad' to 1 & 0
encoder = LabelEncoder()
coffee['goodCoffee'] = encoder.fit_transform(coffee['goodCoffee'])
coffee.head()

# Plot the data
plt.xkcd()
plt.bar(coffee['goodCoffee'].unique(), coffee['goodCoffee'].value_counts(),
        width=0.75, align='center', color=mcolors.XKCD_COLORS)
plt.xticks([0, 1], ['Good', 'Not Good'])
plt.yticks(coffee['goodCoffee'].value_counts())

plt.gca().spines['top'].set_visible(False)
plt.gca().spines['right'].set_visible(False)

plt.ylabel('# of entries')
plt.title('HOW MUCH OF THIS COFFEE SUCKS?')

plt.show()

# and a pie chart
plt.xkcd()
plt.pie(coffee['goodCoffee'].value_counts(), explode=(0, 0.1), labels=(['Not Good', 'Good']), autopct='%1.1f%%',
        shadow=True, startangle=90, colors=mcolors.XKCD_COLORS)
plt.title('HOW MUCH OF THIS COFFEE SUCKS?')

plt.show()

#seperate feature/response vars
X = coffee.drop(columns=['goodCoffee'])
y = coffee['goodCoffee']
#train/test split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

#standardize values
sc = StandardScaler()
X_train = sc.fit_transform(X_train)
X_test = sc.transform(X_test)

#better for mid-size datasets
rfc = RandomForestClassifier(n_estimators = 200)
rfc.fit(X_train, y_train)
pred_rfc = rfc.predict(X_test)

#how well did model perform?
print(classification_report(y_test, pred_rfc))
print(confusion_matrix(y_test, pred_rfc))

#better for smaller datasets
clf = svm.SVC()
clf.fit(X_train, y_train)
pred_clf = clf.predict(X_test)

#how well did model perform?
print(classification_report(y_test, pred_clf))
print(confusion_matrix(y_test, pred_clf))

#better for huge amounts of data
mlpc= MLPClassifier(hidden_layer_sizes=(11, 11, 11), max_iter=500)
mlpc.fit(X_train, y_train)
pred_mlpc = mlpc.predict(X_test)

print(classification_report(y_test, pred_mlpc))
print(confusion_matrix(y_test, pred_mlpc))

from sklearn.metrics import accuracy_score
acc_rfc = accuracy_score(y_test, pred_rfc)
acc_clf = accuracy_score(y_test, pred_clf)
acc_mlpc = accuracy_score(y_test, pred_mlpc)

print(f"Random Forest was {acc_rfc:.2%} accurate")
print(f"SVM was {acc_clf:.2%} accurate")
print(f"Neural Network was {acc_mlpc:.2%} accurate")

#a reminder what our data looks like
coffee[90:100]

#---------------
#A theoretical cup of coffee
new_Aroma = 8
new_Flavor = 7.5
new_Aftertaste = 8
new_Acidity = 8
new_Body = 7.9
new_Balance = 8
#---------------

X_new = [[new_Aroma, new_Flavor, new_Aftertaste, new_Acidity, new_Body, new_Balance]]
X_new = sc.transform(X_new)
y_new = mlpc.predict(X_new)
print(y_new)
if y_new[0] == 1:
    print("it's good!")
else:
    print ("it's not good!")