#!/usr/bin/env python
# coding: utf-8

# In[1]:


#train my classifier for digits
import os
from skimage.io import imread
from skimage.color import rgb2grey

digits_images = []
targets = []
for digit in range(9):
    training_directory = 'resources/digits/'+str(digit)+'/'
    for filename in os.listdir(training_directory):
        if (filename.endswith('.png')):
            training_digit = imread(training_directory+filename)
            training_digit = rgb2grey(training_digit)#color to gray
            digits_images.append(training_digit.reshape(128*128))
            targets.append(digit)


# In[2]:


digits_images[0].shape


# In[3]:


from sklearn.model_selection import train_test_split

X_train, X_test, Y_train, Y_test = train_test_split(digits_images, targets, test_size=0.1, random_state=0)


# In[4]:


from sklearn.neighbors import KNeighborsClassifier
import numpy as np

knn = KNeighborsClassifier(n_neighbors=len(np.unique(Y_train)))
knn.fit(X_train, Y_train)


# In[5]:


score = knn.score(X_test, Y_test)
print score


# In[6]:


knn.predict([X_test[0]])


# In[7]:


from sklearn.ensemble import RandomForestClassifier

rc = RandomForestClassifier(n_estimators=9, max_depth=30)
rc.fit(X_train, Y_train)


# In[8]:


score = rc.score(X_test, Y_test)
print score


# In[9]:


rc.predict([X_test[2]])


# In[10]:


import matplotlib.pyplot as plt

plt.imshow(X_test[2].reshape(128,128), cmap="gray")


# In[ ]: