from fastai.gen_doc.nbdoc import *
from fastai.basics import * 

show_doc(DataBunch)

show_doc(DataBunch.create)

jekyll_warn("You can pass regular pytorch Dataset here, but they'll require more attributes than the basic ones to work with the library. See below for more details.")

show_doc(DataBunch.show_batch)

show_doc(DataBunch.dl)

show_doc(DataBunch.one_batch)

show_doc(DataBunch.one_item)

show_doc(DataBunch.sanity_check)

show_doc(DataBunch.save)

show_doc(load_data)

jekyll_important("The arguments you passed when you created your first `DataBunch` aren't saved, so you should pass them here if you don't want the default.")

jekyll_note("Data cannot be serialized on Windows and then loaded on Linux or vice versa because `Path` object doesn't support this. We will find a workaround for that in v2.")

show_doc(DataBunch.export)

show_doc(DataBunch.load_empty, full_name='load_empty')

show_doc(DataBunch.add_test)

show_doc(DataBunch.add_tfm)

class ArrayDataset(Dataset):
    "Sample numpy array dataset"
    def __init__(self, x, y):
        self.x, self.y = x, y
        self.c = 2 # binary label
    
    def __len__(self):
        return len(self.x)
    
    def __getitem__(self, i):
        return self.x[i], self.y[i]

train_x = np.random.rand(10, 3, 3) # 10 images (3x3)
train_y = np.random.rand(10, 1).round() # binary label

valid_x = np.random.rand(10, 3, 3)
valid_y = np.random.rand(10, 1).round()

train_ds, valid_ds = ArrayDataset(train_x, train_y), ArrayDataset(valid_x, valid_y)
data = DataBunch.create(train_ds, valid_ds, bs=2, num_workers=1)
data.one_batch()

show_doc(DeviceDataLoader)

show_doc(DeviceDataLoader.create)

show_doc(DeviceDataLoader.add_tfm)

show_doc(DeviceDataLoader.remove_tfm)

show_doc(DeviceDataLoader.new)

show_doc(DeviceDataLoader.proc_batch)

show_doc(DatasetType, doc_string=False)

show_doc(DeviceDataLoader.collate_fn)