glue-suite-v2.py

from datasets.features import Features, Sequence, Value
from evaluate.module import EvaluationModuleInfo
# from evaluate.evaluation_suite import SubTask
import evaluate

from typing import Optional, Union, Callable
from dataclasses import dataclass
from datasets import Dataset

@dataclass
class SubTask:
    data: Optional[Union[str, Dataset]] = None
    subset: Optional[str] = None
    split: Optional[str] = None
    data_preprocessor: Optional[Callable] = None
    args_for_task: Optional[dict] = None


class Suite(evaluate.EvaluationSuite):

    def __init__(self, name):
        super().__init__(name)

    def _info(self):
        return EvaluationModuleInfo(
            module_type="evaluation_suite",
            description="dummy metric for tests",
            citation="insert citation here",
            features=Features({"predictions": Value("int64"), "references": Value("int64")}))

    def setup(self):
        self.preprocessor = None #lambda x: x["text"].lower()
        self.suite = [
            SubTask(
                data="imdb",
                split="test[:10]",
                data_preprocessor=lambda x: x["text"].lower(),#self.preprocessor,
                args_for_task={
                    "metric": "accuracy",
                    "input_column": "text",
                    "label_column": "label",
                    "label_mapping": {
                        "LABEL_0": 0.0,
                        "LABEL_1": 1.0
                    }
                }
            ),
            SubTask(
                data="sst2",
                split="test[:10]",
                data_preprocessor=self.preprocessor,
                args_for_task={
                    "metric": "accuracy",
                    "input_column": "sentence",
                    "label_column": "label",
                    "label_mapping": {
                        "LABEL_0": 0.0,
                        "LABEL_1": 1.0
                    }
                }
            )
        ]