from google.colab import drive
drive.mount('/content/drive')

!pip3 install shap
!pip3 install dice-ml

import pandas as pd
from pathlib import Path
import numpy as np
from joblib import dump, load
import shap
import warnings
warnings.filterwarnings("ignore")
# Sklearn imports
from sklearn.compose import ColumnTransformer
from sklearn.model_selection import train_test_split
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import OneHotEncoder
from sklearn.metrics import accuracy_score, roc_auc_score, precision_recall_fscore_support
from sklearn.base import TransformerMixin, BaseEstimator
from keras.utils import np_utils
from keras.wrappers.scikit_learn import KerasClassifier
from keras.models import Sequential
from keras.layers import Activation, Dense, Dropout
from tensorflow.keras.layers.experimental import RandomFourierFeatures
from keras.regularizers import l2
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import MinMaxScaler
import keras
from numpy.random import seed
seed(1)

# Tensorflow import
import tensorflow as tf
tf.random.set_seed(2)

# DiCE imports
import dice_ml
from dice_ml.utils import helpers  # helper functions

%load_ext autoreload
%autoreload 2

# custom transformer for sklearn pipeline
class LabelTransform(TransformerMixin, BaseEstimator):
    def transform(self, X):
      return X

    def fit(self, X, y=None):
      y= np_utils.to_categorical(y)
      return self

def create_model(init='uniform', optimizer='Adam'):
    # define model
    model = Sequential()
    model.add(Dense(64, input_dim=len(train_X.columns), kernel_initializer=init, activation='tanh')) 
    model.add(Dropout(0.2))
    model.add(Dense(100, input_dim=len(train_X.columns))) 
    RandomFourierFeatures(
           output_dim=200, kernel_initializer="gaussian"
       ),
    
    model.add(Dense(2, kernel_initializer=init,kernel_regularizer='l2'))
    model.compile(loss=keras.losses.hinge, optimizer=tf.keras.optimizers.Adam(lr=0.001), metrics=['accuracy'])
  
    return model

np.random.seed(0)

# Fix the probelem that SHAP is not compatible with Pipeline
with open('/usr/local/lib/python3.7/dist-packages/shap/explainers/_permutation.py', 'r') as f:
  filedata = f.read()

# Replace the target string
filedata = filedata.replace('return explanation._old_format()', 'return explanation')

# Write the file out again
with open('/usr/local/lib/python3.7/dist-packages/shap/explainers/_permutation.py', 'w') as file:
  file.write(filedata)