Converte a base de código para python 3x

src/main.py

@@ -2,7 +2,7 @@
 from transform import Transform
 
 print('1 - Train\n2 - Test')
-x = int(raw_input('Option:'))
+x = int(input('Option:'))
 
 if x is 1:
     NeuralNetwork()

src/nn/__init__.py

@@ -1,11 +1,11 @@
+'''base da classe neuralnetwork'''
 import tensorflow as tf
 import numpy as np
 import time
 #
-from dataset.train import *
-import convert
-import neural
-import test
+from .dataset.train import *
+from . import neural
+from . import test
 
 class NeuralNetwork:
     def __init__(self):
@@ -15,14 +15,17 @@ def __init__(self):
 
         print('desfrdgthygju', len(data_input))
 
-        [to_train, to_test, res_to_train, res_to_test] = convert.dataset_matrix_to_lists(data_input, data_output, 75, 30)
+        [to_train,
+         to_test,
+         res_to_train,
+         res_to_test] = convert.dataset_matrix_to_lists(data_input, data_output, 75, 30)
 
         for x in range(0, 1):
             [epoch, err] = neural.train(to_train, res_to_train)
             print('epoch:', epoch, 'mse:', err)
             if err < target:
                 test_success = test_success + 1
-            test.test_neural(epoch, to_test, res_to_test);
+            test.test_neural(epoch, to_test, res_to_test)
 
         print('time', time.time() - start)
         print('success ', test_success)

src/nn/confusion.py

@@ -1,10 +1,12 @@
+''' '''
 import seaborn as sn
 import pandas as pd
 import matplotlib.pyplot as plt
 #
-import convert
+from . import convert
 
 def build_matrix_array(matrix, desired_results, obtained_results):
+    ''' D '''
     for position in range(0, len(desired_results)):
         #
         desired = int(convert.convert_binary_to_int(desired_results[position]))
@@ -16,9 +18,11 @@ def build_matrix_array(matrix, desired_results, obtained_results):
     return matrix
 
 def confusion_matrix_graphic(array):
-    df_cm = pd.DataFrame(array, index = [i for i in '0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ'],
-                      columns = [i for i in '0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ'])
-    plt.figure(figsize = (10,7))
+    ''' creates confusion matrix graphic '''
+    df_cm = pd.DataFrame(array,
+                         index=[i for i in '0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ'],
+                         columns=[i for i in '0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ'])
+    plt.figure(figsize=(10, 7))
     sn.heatmap(df_cm, annot=True)
     # Display matrix
     plt.matshow(array)

src/nn/convert.py

@@ -1,10 +1,17 @@
-import numpy as np
+''' # convert the shit ou of here! '''
 
-# convert the shit ou of here!
 # only 6bits
+import numpy as np
+
+
 def convert_binary_to_int(binary_value):
 
-    result = binary_value[5] * 1 + binary_value[4] * 2 + binary_value[3] * 4 + binary_value[2] * 8 + binary_value[1] * 16 + binary_value[0] * 32
+    result = (binary_value[5] * 1
+              + binary_value[4] * 2
+              + binary_value[3] * 4
+              + binary_value[2] * 8
+              + binary_value[1] * 16
+              + binary_value[0] * 32)
 
     if result < 36 and result > -1:
         return result

src/nn/dataset/train.py

@@ -1,16 +1,17 @@
-from numbers.zero import *
-from numbers.one import *
-from numbers.two import *
-from numbers.three import *
-from numbers.four import *
-from numbers.five import *
-from numbers.six import *
-from numbers.seven import *
-from numbers.eight import *
-from numbers.nine import *
+''' Train data into a input and a output '''
+from .numbers.zero import *
+from .numbers.one import *
+from .numbers.two import *
+from .numbers.three import *
+from .numbers.four import *
+from .numbers.five import *
+from .numbers.six import *
+from .numbers.seven import *
+from .numbers.eight import *
+from .numbers.nine import *
 #
-from chars.ca import *
-from chars.cb import *
+from .chars.ca import *
+from .chars.cb import *
 
 data_input = [
     input_zero,

src/nn/test.py

@@ -1,7 +1,7 @@
 import tensorflow as tf
 import numpy as np
 #
-import confusion
+from . import confusion
 
 meta_graph_src = './nn/tmp/my_test_model.meta'
 checkpoint_src = './nn/tmp'
@@ -21,7 +21,7 @@ def test_neural(epoch, data_input, data_output):
 
     with tf.Session() as sess:
         saver = tf.train.import_meta_graph(meta_graph_src)
-        saver.restore(sess,tf.train.latest_checkpoint(checkpoint_src))
+        saver.restore(sess, tf.train.latest_checkpoint(checkpoint_src))
         graph = tf.get_default_graph()
         w1 = sess.run(graph.get_tensor_by_name("w1:0"))
         b1 = sess.run(graph.get_tensor_by_name("b1:0"))

src/transform/__init__.py

@@ -10,21 +10,30 @@ def __init__(self):
 
         # make it black and white
         gray = col.convert('L')
-        bw = gray.point(lambda x: 0 if x<128 else 255, '1')
+        bw = gray.point(lambda x: 0 if x < 128 else 255, '1')
 
         # get transformed data
         pixels = list(bw.getdata())
         width, height = bw.size
 
         # make a matrix
-        pixels = [pixels[i * width:(i + 1) * width] for i in xrange(height)]
+        pixels = [pixels[i * width:(i + 1) * width] for i in range(height)]
 
         # update values
-        pixels = [[1. if pixels[i][g] == 0 else 0. for g in xrange(width)] for i in xrange(height)]
+        pixels = [[1. if pixels[i][g] == 0 else 0. for g in range(width)] for i in range(height)]
+        pixels2 = [[' ' if pixels[i][g] == 0 else '@' for g in range(width)] for i in range(height)]
+
 
+        print('Matriz Resultado:')
+
         for h in range(0, len(pixels)):
         	print(pixels[h])
 
+        print('Matriz visual:')
+
+        for h in range(0, len(pixels2)):
+            print(pixels2[h])
+
         # test with trained neural network
 
     def __del__(self):