minicheck_web/minicheck.py

# Adapt code from https://github.com/yuh-zha/AlignScore/tree/main

import sys
sys.path.append("..")

from minicheck_web.inference import Inferencer
from typing import List, Dict
import numpy as np


class MiniCheck:
    def __init__(self, path, chunk_size=None, max_input_length=None, batch_size=16) -> None:

        self.model = Inferencer(
            path=path,
            batch_size=batch_size, 
            chunk_size=chunk_size,
            max_input_length=max_input_length,
        )

    def score(self, data: Dict) -> List[float]:
        '''
        pred_labels: 0 / 1 (0: unsupported, 1: supported)
        max_support_probs: the probability of "supported" for the chunk that determin the final pred_label
        used_chunks: divided chunks of the input document
        support_prob_per_chunk: the probability of "supported" for each chunk
        '''

        inputs = data['inputs']
        docs = inputs['docs']
        claims = inputs['claims']

        assert isinstance(docs, list) or isinstance(docs, np.ndarray), f"docs must be a list or np.ndarray"
        assert isinstance(claims, list) or isinstance(claims, np.ndarray), f"claims must be a list or np.ndarray"  

        max_support_prob, used_chunk, support_prob_per_chunk = self.model.fact_check(docs, claims)
        pred_label = [1 if prob > 0.5 else 0 for prob in max_support_prob]

        return pred_label, max_support_prob, used_chunk, support_prob_per_chunk