# -*- coding: utf-8 -*-

#conformal_skin_prediction.py

__doc__ = """ Description:

"""

##########################################

# Import modules:

import sys

import argparse

import pandas

from pandas import DataFrame

import rdkit

from rdkit import Chem

from rdkit.Chem import Descriptors

import numpy as np

from numpy.core.numeric import asanyarray

from numpy import mean

import nonconformist

from nonconformist.icp import IcpRegressor

from nonconformist.nc import NormalizedRegressorNc

from nonconformist.nc import RegressorNc, abs_error, abs_error_inv

import sklearn

from sklearn.ensemble import RandomForestRegressor

from sklearn.svm import SVR

from sklearn.neighbors import KNeighborsRegressor

from sklearn.metrics import mean_squared_error

from sklearn.metrics import r2_score

from sklearn.metrics import mean_absolute_error

from sklearn.decomposition import PCA

from sklearn import preprocessing

from math import sqrt

import random

import matplotlib

import matplotlib.pyplot as plt

import seaborn as sns

import dill

import copy

# -----------------------------------------------------------------------------

parser = argparse.ArgumentParser()

parser.add_argument('-verbose', help='Get verbose output', action='store_true')

parser.add_argument('-i','-infile', help='Define the input file. Default:baba_jan_2015_smiles.csv', default='C:\Users\Martin\Documents\Magnus_Lundborg\Nya_data/baba_jan_2015_smiles.csv')

parser.add_argument('-c','-conformal_model', action ='store_true', help='Create a new conformal predictions model')

parser.add_argument('-m', '-model_type', default = 'RF', help = 'Define model algorithm, RF (default) or SVM')

parser.add_argument('-t', '-test_set', default='random', help = 'Use: random (default), full_model, existing, reference')

parser.add_argument('-p', '-predict_model', action='store_true', help="Predict values for one or many smile-codes using single conformal model")

parser.add_argument('-d','-data_file_name', help='Output file name. File with descriptor data', default='data_with_descriptors.csv')

parser.add_argument('-num_models', default='100', help='Number of models to create. Default = 100')

parser.add_argument('-reference', default=1, help='Use this reference as test_set. Default= 1')

parser.add_argument('-significance', default=0.2, help='Set the significance level of the conformal prediction. Default = 0.2')

parser.add_argument('-phys','-physiochemical_descriptors', default = False, help='Print out data about descriptors. Default = false')

parser.add_argument('-smile', default = 'CCO', help ='Smiles for which skin permeability prediction is requested')

parser.add_argument('-plot', action ='store_true', help='Plot experimental versus predicted')

parser.add_argument('-pca','-PCA', action ='store_true', help='Create a PCA')

args = parser.parse_args()

# -----------------------------------------------------------------------------

#Plot version numbers for Python, matplotlib, pandas and numpy

if args.verbose:

print(' ')

print(' ################### Versions ################### ')

print('Python version:\n{}'.format(sys.version))

print('matplotlib version: {}'.format(matplotlib.__version__))

print('pandas version: {}'.format(pandas.__version__))

print('sklearn version: {}'.format(sklearn.__version__))

print('numpy version: {}'.format(np.__version__))

print('RDkit version: {}'.format(rdkit.__version__))

print('nonconformist version: {}'.format(nonconformist.__version__))

print('dill version: {}'.format(dill.__version__))

# -----------------------------------------------------------------------------

def read_data(infile):

return file_data

def calculate_descriptors(smiles):

return descriptors_df

def create_indices_test_training_calibration(data):

return train, calibrate, test

def create_train_test_calibrate_sets(data, descriptors_df, train_i, calibrate_i, test_i):

return Xtrain, Xtest, Xcalibrate, ytrain, ytest, ycalibrate

def write_csv_with_data(data, descriptors_df, newfilename):

return True

def randomize_new_indices(train_i, calibrate_i, test_i, data, i):

return list(train), list(calibrate), list(test)

def calculate_prediction_y_and_error(median_values):

return Y_pred_median, error_median

def save_models(ICPs):

return True

def load_models():

return icps

def predict_from_smiles_conformal_median(regressor_list, smiles):

return C

def create_conformal_model():

#print('explained variance ratio (first two components): %s' % str(pca.explained_variance_ratio_))

# -----------------------------------------------------------------------------

if args.verbose:

print(" Verbose output turned on")

print(' Model Type is: '+args.m) # m = model_type

#print(parser)

if args.c:

#Create conformal models

if args.verbose:

print(' Input data file: ')

print(' '+args.i) # i = input_file

create_conformal_model()

if args.p:

if args.verbose:

print('Predict using earlier models')

try:

classifier_list = load_models()

except:

print('No conformal model found')

smile_code = [args.smile] #args.smiles

predict_from_smiles_conformal_median(classifier_list, smile_code)