import numpy as np 
import seaborn as sns
import matplotlib.pyplot as plt

from sklearn.linear_model import LogisticRegression  
from sklearn.model_selection import train_test_split 
from sklearn.neighbors import KNeighborsClassifier 
from sklearn import svm  
from sklearn import metrics 
from sklearn.tree import DecisionTreeClassifier

df['color_int'] = df['color'].astype('category').cat.codes

transform = {'Ideal':0, 'Premium':1, 'Good':2, 'Very Good':3, 'Fair':4}
df['cut_int']=df['cut'].apply(lambda x: transform[x])

transform_clarity ={'SI2':6, 'SI1':5, 'VS1':3, 'VS2':4, 'VVS2':2, 'VVS1':1, 'I1':7, 'IF':0}
df['clarity_int']=df['clarity'].apply(lambda x: transform_clarity[x])

features = ['carat','color_int','cut_int','clarity_int','x','y','z','depth','table']
X = df[features]
y = df['price']

train_X, test_X, train_y, test_y = train_test_split(X, y, test_size = 0.25, random_state=0)

from sklearn.ensemble import RandomForestRegressor
from sklearn.tree import DecisionTreeRegressor
from sklearn.metrics import mean_absolute_error

init_features = ['carat','color_int','cut_int','clarity_int','x','y','z','depth','table']
from random import random, randint, choice

def diff(li1, li2):
    return (list(list(set(li1)-set(li2)) + list(set(li2)-set(li1))))

def step(features):

    if random()<0.5: 
      indx = randint(0,len(features)-1)
      print("removing",features[indx])
      features.remove(features[indx])
      return(features)
    else:
      d = diff(features,init_features)
      if not d:
        print("do nothing")
        return(features)
      else:
        f = choice(d)
        print("adding",f)
        features.append(f)
        return(features)

def evaluate(features):

    X = df[features]
    y = df['price']
    train_X, test_X, train_y, test_y = train_test_split(X, y, random_state=1)
    rf.fit(train_X,train_y)
    predictions = rf.predict(test_X)
    val_mae = mean_absolute_error(predictions, test_y)
    return(val_mae)

def optmize():

    print("###")
    start_features = ['carat','color_int','cut_int','clarity_int','x','y','z','depth','table']
    ff=start_features
    old_mae = 10000000

    for i in range(10):
      new_features = step(ff)
      mae = evaluate(new_features)
      print(i,mae,features)
      if mae < old_mae:
        ff = new_features
        old_mae = mae
        print("accepting result")
      else:
        print("rejecting result")