path = "/content/drive/My Drive/Colab Notebooks/pillbox/Pillbox.csv"
images_path = "/content/drive/My Drive/Colab Notebooks/pillbox/pillbox_production_images_full_201812"
data = pd.read_csv(path, usecols=['has_image','splshape_text','medicine_name', 'splimprint', 'splimage', 'splcolor_text','source'],
                            dtype={'has_image': 'str','splshape_text': 'str', 'medicine_name' : 'str',  'splimprint' : 'str' , 'splimage' : 'str', 'splcolor_text' : 'str', 'source' : 'str'})
data = data.replace(np.nan, '', regex=True)


data_has_image = data.loc[data['has_image'] == 'True']
unique_shapes = data_has_image.splshape_text.unique()
print(unique_shapes)
unique_shape_count = len(data_has_image.splshape_text.unique())
shape_to_index = {}
for i in range(unique_shape_count):
  shape_to_index[unique_shapes[i]] = i
shape_to_index
print("unique shape: ", len(data_has_image.splshape_text.unique()))
data_has_image = data_has_image [:1300] #120

print("has image len: ", len(data_has_image))
images = []
labels = []
for row_i in range(data_has_image.shape[0]):
    row = data_has_image.iloc[row_i]
    if row.loc['splimage'] == "":
        continue
    image_path = images_path+"/" + row.loc['splimage'] + ".jpg"
    #image = cv2.imread(image_path, 0)
    try:
        image = Image.open(image_path).convert('LA')
    except:
        continue

    desired_size = 32
    im_pth = image_path
    im = Image.open(im_pth)
    old_size = im.size  # old_size[0] is in (width, height) format
    ratio = float(desired_size) / max(old_size)
    new_size = tuple([int(x * ratio) for x in old_size])
    # use thumbnail() or resize() method to resize the input image
    # thumbnail is a in-place operation
    # im.thumbnail(new_size, Image.ANTIALIAS)
    im = im.resize(new_size, Image.ANTIALIAS)
    # create a new image and paste the resized on it

    new_im = Image.new("RGB", (desired_size, desired_size))
    new_im.paste(im, ((desired_size - new_size[0]) // 2,
                      (desired_size - new_size[1]) // 2))

    #new_im.show()
    image = new_im

    #splimprint = row.loc['splimprint']
    #splimprint_vector = splimprint_to_vector(splimprint)
    splshape = row.loc['splshape_text']
    splshape_vector = np.zeros(unique_shape_count)
    splshape_vector[shape_to_index[splshape]] = 1
    #print("shape and vec: ",splshape, splshape_vector)
    images.append(np.array(image))
    labels.append(splshape_vector)

    if row_i % 200 == 0:
      print("loading images: ", row_i, "/", data_has_image.shape[0])