import numpy as np
import cv2

ID_APP_MAP_FILENAME = 'miguel/train.csv'
IMAGE_PATH_FORMAT = 'miguel/%04d_%s.png'

def load_id_app_map(filename):
    result = {}
    with open(filename) as f:
        lines = f.read().splitlines()[1:]
        for line in lines:
            parts = map(int, line.split(','))
            if len(parts) == 2:
                result[parts[0]] = parts[1]
            elif len(parts) == 1: # no label
                result[parts[0]] = -1

    return result

def gen_label_array(label, label_count):
    label_flags = np.zeros(label_count, np.uint8)
    if label >= 0: # -1 signifies no label
        label_flags[label] = 1
    return label_flags

def load_flat_image(filename):
    image = cv2.imread(filename, cv2.IMREAD_GRAYSCALE)
    if image is None:
        raise RuntimeError('Failed to load image "%s".' % filename)
    return image.flatten()

id_app_map = load_id_app_map(ID_APP_MAP_FILENAME)

label_count = np.max(id_app_map.values()) + 1

rows = []
for id,appliance in sorted(id_app_map.items()):
    labels = gen_label_array(appliance, label_count)
    flattened_c = load_flat_image(IMAGE_PATH_FORMAT % (id, 'c'))
    flattened_v = load_flat_image(IMAGE_PATH_FORMAT % (id, 'v'))
    rows.append(np.hstack([labels, flattened_c]))
    rows.append(np.hstack([labels, flattened_v]))

data = np.vstack(rows)

print data.shape