logitorch.data_collators.prover_collator

Classes

Node
PRoverProofWriterCollator	A collator class for processing data in the PRoverProofWriter format.

Module Contents

class logitorch.data_collators.prover_collator.Node(head: str)

head

class logitorch.data_collators.prover_collator.PRoverProofWriterCollator(pretrained_roberta_tokenizer: str)

A collator class for processing data in the PRoverProofWriter format.

Args:

pretrained_roberta_tokenizer (str): The name or path of the pretrained RoBERTa tokenizer.

Attributes:

tokenizer (RobertaTokenizer): The pretrained RoBERTa tokenizer.

Methods:

get_proof_graph_with_fail(proof_str: str) -> Tuple[List[str], List[str]]: Extracts the proof graph and edges from a proof string with a “FAIL” node.

get_proof_graph(proof_str: str) -> Tuple[List[str], List[Tuple[str, str]]]: Extracts the proof graph and edges from a proof string.

get_node_edge_label_constrained(x: str) -> Tuple[List[int], List[np.ndarray]]: Generates node and edge labels for a given input.

__call__(batch) -> Tuple[Dict[str, torch.Tensor], torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor]: Collates a batch of data into tokenized tensors.

get_node_edge_label_constrained(x: str) → Tuple[List[int], List[numpy.ndarray]]

Generates node and edge labels for a given input.

Args:

x (str): The input.

Returns:

Tuple[List[int], List[np.ndarray]]: A tuple containing the list of node labels and the list of edge labels.

get_proof_graph(proof_str: str) → Tuple[List[str], List[Tuple[str, str]]]

Extracts the proof graph and edges from a proof string.

Args:

proof_str (str): The proof string.

Returns:

Tuple[List[str], List[Tuple[str, str]]]: A tuple containing the list of nodes and the list of edges.

get_proof_graph_with_fail(proof_str: str) → Tuple[List[str], List[str]]

Extracts the proof graph and edges from a proof string with a “FAIL” node.

Args:

proof_str (str): The proof string.

Returns:

Tuple[List[str], List[str]]: A tuple containing the list of nodes and the list of edges.

tokenizer