chapter-4/3-reduce-feature-length-with-pca.ipynb Wrong data used in annoy section

[gitgithan]

2 occurences of feature_list should be dataset instead, also 2048 should be 100. These are defined earlier in

num_items = 100000
num_dimensions = 100
dataset = np.random.randn(num_items, num_dimensions)
dataset /= np.linalg.norm(dataset, axis=1).reshape(-1, 1)

annoy_training_time = []
annoy_test_time = []
annoy_trees = [
    1, 2, 3, 4, 5, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 200, 300
]
for num_trees in annoy_trees:
    t = AnnoyIndex(2048)  # Length of item vector that will be indexed
    for i in range(num_images):
        feature = feature_list[i]
        t.add_item(i, feature)
    start_time = time.time()
    t.build(num_trees)  #50 trees
    end_time = time.time()
    annoy_training_time.append(end_time - start_time)
    start_time = time.time()
    indices = t.get_nns_by_vector(feature_list[random_image_index],
                                  5,
                                  include_distances=True)
    end_time = time.time()
    annoy_test_time.append(end_time - start_time)
    print("For number of trees = ", num_trees, ",\tTime to train = ",
          annoy_training_time[-1], ",\tTime to test = ", annoy_test_time[-1])