dna

Class to represent DNA sequences.

Attributes

NUCLEOBASES
PAD_TOKEN

Classes

DNA

Module Contents

dna.NUCLEOBASES = ('A', 'C', 'T', 'G')

dna.PAD_TOKEN = '0'

class dna.DNA(proxy_fmt='onehot-np', **kwargs)

Bases: gflownet.envs.sequences.base.SequenceBase

Parameters:

proxy_fmt (str) –

Specifies the proxy format. Options:

• onehot: One-hot encoding

• letters: The nucleobases as a list of strings

• np or numpy: numpy, for the onehot case

• torch or tensor: torch tensor, for the onehot case

states2proxy_onehot(states)

Prepares a batch of states in “environment format” for a proxy model: states are one-hot encoded. If numpy is True (default), the output is converted into a numpy array, otherwise it remains a torch tensor.

Example, with max_length = 5:

• Sequence (tokens): ACGC

• state: [1, 2, 4, 2, 0]

• policy format:

  [0, 1, 0, 0, 0, (A)

  0, 0, 1, 0, 0, (C) 0, 0, 0, 0, 1, (G) 0, 0, 1, 0, 0, (C) 1, 0, 0, 0, 0] (PAD)

Parameters:

states (tensor) – A batch of states in environment format, either as a list of states or as a single tensor.

Returns:

A numpy array containing the one-hot encoding of all the states in the batch.

Return type:

Union[torchtyping.TensorType[batch, policy_input_dim], numpy.typing.NDArray]