Distribution-based dataloaders

Dataloaders that return data by sampling from some pre-defined distributions.

BaseDistributionDataLoader

 BaseDistributionDataLoader ()

Base class for data loaders. The idea of the data loader is to provide all external information to the environment (including lagged data, demand etc.). Internal data influenced by past decisions (like inventory levels) is to be added from within the environment

source

NormalDistributionDataLoader

 NormalDistributionDataLoader (mean:float, std:float, num_units:int,
                               truncated_low:int=0,
                               truncated_high:int=None)

A dataloader that generates a dataset of normally distributed values.

	Type	Default
mean	float
std	float
num_units	int
truncated_low	int	0
truncated_high	int	None

dataloader = NormalDistributionDataLoader(mean=3, std=4, num_units=2)

sample_X, sample_Y = dataloader[0]
print("sample:", sample_X, sample_Y)
print("sample shape Y:", sample_Y.shape)

## The next print should give an error:
#print("length:", len(dataloader))

sample: None [0. 0.]
sample shape Y: (2,)

dataloader.train()
dataloader.val()
dataloader.test()