%matplotlib inline
from fastai.gen_doc.nbdoc import *
from fastai.vision import * 

path = untar_data(URLs.MNIST_SAMPLE); path

tfms = get_transforms(do_flip=False)
data = ImageDataBunch.from_folder(path, ds_tfms=tfms, size=24)

data.show_batch(rows=3, figsize=(5,5))

pd.read_csv(path/'labels.csv').head()

data = ImageDataBunch.from_csv(path, ds_tfms=tfms, size=28)

data.show_batch(rows=3, figsize=(5,5))

planet = untar_data(URLs.PLANET_SAMPLE)

df =pd.read_csv(planet/'labels.csv')
df.head()

data = ImageDataBunch.from_csv(planet, folder='train', size=128, suffix='.jpg', label_delim=' ',
    ds_tfms=get_transforms(flip_vert=True, max_lighting=0.1, max_zoom=1.05, max_warp=0.))

data.show_batch(rows=3, figsize=(10,8), ds_type=DatasetType.Valid)

show_doc(ImageDataBunch)

show_doc(ImageDataBunch.from_folder)

data = ImageDataBunch.from_folder(path, ds_tfms=tfms, size=24)

show_doc(ImageDataBunch.from_csv)

data = ImageDataBunch.from_csv(path, ds_tfms=tfms, size=24);

show_doc(ImageDataBunch.from_df)

df = pd.read_csv(path/'labels.csv', header='infer')
df.head()

data = ImageDataBunch.from_df(path, df, ds_tfms=tfms, size=24)

show_doc(ImageDataBunch.from_name_re)

fn_paths = [path/name for name in df['name']]; fn_paths[:2]

pat = r"/(\d)/\d+\.png$"
data = ImageDataBunch.from_name_re(path, fn_paths, pat=pat, ds_tfms=tfms, size=24)

data.classes

show_doc(ImageDataBunch.from_name_func)

def get_labels(file_path): return '3' if '/3/' in str(file_path) else '7'
data = ImageDataBunch.from_name_func(path, fn_paths, label_func=get_labels, ds_tfms=tfms, size=24)
data.classes

show_doc(ImageDataBunch.from_lists)

labels_ls = list(map(get_labels, fn_paths))
data = ImageDataBunch.from_lists(path, fn_paths, labels=labels_ls, ds_tfms=tfms, size=24)
data.classes

show_doc(ImageDataBunch.create_from_ll)

show_doc(ImageDataBunch.single_from_classes)

jekyll_note('This method is deprecated, you should use DataBunch.load_empty now.')

path = untar_data(URLs.CIFAR); path

show_doc(channel_view)

data = ImageDataBunch.from_folder(path, ds_tfms=tfms, valid='test', size=24)

def channel_view(x:Tensor)->Tensor:
    "Make channel the first axis of `x` and flatten remaining axes"
    return x.transpose(0,1).contiguous().view(x.shape[1],-1) 

t = torch.Tensor(128, 3, 24, 24)

t.size()

tensor = channel_view(t)

tensor.size()

show_doc(ImageDataBunch.batch_stats)

data.batch_stats()

show_doc(ImageDataBunch.normalize)

data.normalize(cifar_stats)

data.batch_stats()

show_doc(normalize)

show_doc(denormalize)

show_doc(normalize_funcs)

data = ImageDataBunch.from_folder(untar_data(URLs.MNIST_SAMPLE),
                                  ds_tfms=tfms, size=24)
data.normalize()
data.show_batch(rows=3, figsize=(6,6))

show_doc(get_annotations)

show_doc(bb_pad_collate)

show_doc(ImageItemList, title_level=3)

show_doc(ImageItemList.from_folder)

show_doc(ImageItemList.from_df)

show_doc(get_image_files)

show_doc(ImageItemList.open)

show_doc(ImageItemList.show_xys)

show_doc(ImageItemList.show_xyzs)

show_doc(ObjectCategoryList, title_level=3)

show_doc(ObjectItemList, title_level=3)

show_doc(SegmentationItemList, title_level=3)

show_doc(SegmentationLabelList, title_level=3)

show_doc(PointsLabelList, title_level=3)

show_doc(PointsItemList, title_level=3)

show_doc(ImageImageList, title_level=3)

show_doc(download_images)

show_doc(verify_images)

show_doc(PointsItemList.get)

show_doc(SegmentationLabelList.new)

show_doc(ImageItemList.from_csv)

show_doc(ObjectCategoryList.get)

show_doc(ImageItemList.get)

show_doc(SegmentationLabelList.reconstruct)

show_doc(ImageImageList.show_xys)

show_doc(ImageImageList.show_xyzs)

show_doc(ImageItemList.open)

show_doc(PointsItemList.analyze_pred)

show_doc(SegmentationLabelList.analyze_pred)

show_doc(PointsItemList.reconstruct)

show_doc(SegmentationLabelList.open)

show_doc(ImageItemList.reconstruct)

show_doc(resize_to)

show_doc(ObjectCategoryList.reconstruct)

show_doc(PointsLabelList.reconstruct)

show_doc(PointsLabelList.analyze_pred)

show_doc(PointsLabelList.get)

show_doc(ObjectCategoryList.analyze_pred)