from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import confusion_matrix
from sklearn import preprocessing as pp
from sklearn import datasets, linear_model
from sklearn.metrics import classification_report, confusion_matrix
from sklearn.model_selection import train_test_split

#data pre-processing
le = pp.LabelEncoder()
metadata['dx_type'].iloc[:] = le.fit_transform(metadata['dx_type']).astype(float)
metadata['sex'].iloc[:] = le.fit_transform(metadata['sex']).astype(float)
metadata['localization'].iloc[:] = le.fit_transform(metadata['localization']).astype(float)

labels = metadata['dx']
features = metadata.drop(['lesion_id', 'image_id', 'dx'], axis=1)

#train + test
X_train, X_test, y_train, y_test = train_test_split(features, labels, test_size=0.2, random_state=42)
X_train = X_train.fillna(-1.0)
X_test = X_test.fillna(-1.0)
y_train = y_train.fillna(-1.0)
y_test = y_test.fillna(-1.0)

#random forest classifier
model = RandomForestClassifier()
model.fit(X_train, y_train)
ypred = model.predict(X_test)

#save labels, ypred
reverse = dict()
label_map = y_test.index
for l in label_map:
    reverse[y_test.loc[l]] = l
labels_named = [reverse[number] for number in y_test]
ypred_named = [reverse[number] for number in ypred]

pred_df = pd.DataFrame()
pred_df["y_true"] = labels_named
pred_df["y_pred"] = ypred_named
pred_df.to_csv("../Figures/prat_classifications.csv")

#confusion matrix
print('Confusion Matrix')
print(confusion_matrix(y_test, ypred))
print('\n\n\nClassification Report')
print(classification_report(y_test, ypred, target_names=metadata['dx'].unique()))
num_right = np.sum(y_test == ypred)
print(f"\nAccuracy {num_right / ypred.shape[0]}")