Datasets:
Tasks:
Text Classification
Sub-tasks:
acceptability-classification
Languages:
Italian
ArXiv:
License:
import csv | |
import sys | |
import datasets | |
from typing import List | |
csv.field_size_limit(sys.maxsize) | |
_CITATION = """\ | |
@inproceedings{trotta-etal-2021-monolingual, | |
author = {Trotta, Daniela and Guarasci, Raffaele and Leonardelli, Elisa and Tonelli, Sara}, | |
title = {Monolingual and Cross-Lingual Acceptability Judgments with the Italian {CoLA} corpus}, | |
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2021", | |
month = nov, | |
year = {2021}, | |
address = "Punta Cana, Dominican Republic and Online", | |
publisher = "Association for Computational Linguistics", | |
url = "https://arxiv.org/abs/2109.12053", | |
} | |
""" | |
_DESCRIPTION = """\ | |
The Italian Corpus of Linguistic Acceptability includes almost 10k sentences taken from | |
linguistic literature with a binary annotation made by the original authors themselves. | |
The work is inspired by the English Corpus of Linguistic Acceptability (CoLA) by Warstadt et al. | |
Part of the dataset has been manually annotated to highlight 9 linguistic phenomena. | |
""" | |
_HOMEPAGE = "https://github.com/dhfbk/ItaCoLA-dataset" | |
_LICENSE = "None" | |
_SPLITS = ["train", "test"] | |
class ItaColaConfig(datasets.BuilderConfig): | |
"""BuilderConfig for ItaCoLA.""" | |
def __init__( | |
self, | |
features, | |
data_url, | |
**kwargs, | |
): | |
""" | |
Args: | |
features: `list[string]`, list of the features that will appear in the | |
feature dict. Should not include "label". | |
data_url: `string`, url to download the zip file from. | |
**kwargs: keyword arguments forwarded to super. | |
""" | |
super().__init__(version=datasets.Version("1.0.0"), **kwargs) | |
self.data_url = data_url | |
self.features = features | |
class ItaCola(datasets.GeneratorBasedBuilder): | |
VERSION = datasets.Version("1.0.0") | |
BUILDER_CONFIGS = [ | |
ItaColaConfig( | |
name="scores", | |
features=["unique_id", "source", "acceptability", "sentence"], | |
data_url="https://raw.githubusercontent.com/dhfbk/ItaCoLA-dataset/main/ItaCoLA_dataset.tsv" | |
), | |
ItaColaConfig( | |
name="phenomena", | |
features=[ | |
"unique_id", | |
"source", | |
"acceptability", | |
"sentence", | |
"cleft_construction", | |
"copular_construction", | |
"subject_verb_agreement", | |
"wh_islands_violations", | |
"simple", | |
"question", | |
"auxiliary", | |
"bind", | |
"indefinite_pronouns", | |
], | |
data_url="https://github.com/dhfbk/ItaCoLA-dataset/raw/main/ItaCoLA_dataset_phenomenon.tsv" | |
), | |
] | |
DEFAULT_CONFIG_NAME = "scores" | |
def _info(self): | |
features = {feature: datasets.Value("int32") for feature in self.config.features} | |
features["source"] = datasets.Value("string") | |
features["sentence"] = datasets.Value("string") | |
return datasets.DatasetInfo( | |
description=_DESCRIPTION, | |
features=datasets.Features(features), | |
homepage=_HOMEPAGE, | |
license=_LICENSE, | |
citation=_CITATION, | |
) | |
def _split_generators(self, dl_manager): | |
"""Returns SplitGenerators.""" | |
data_file = dl_manager.download_and_extract(self.config.data_url) | |
if self.config.name == "scores": | |
return [ | |
datasets.SplitGenerator( | |
name=datasets.Split.TRAIN, | |
gen_kwargs={ | |
"filepath": data_file, | |
"split": "train", | |
"features": self.config.features, | |
}, | |
), | |
datasets.SplitGenerator( | |
name=datasets.Split.TEST, | |
gen_kwargs={ | |
"filepath": data_file, | |
"split": "test", | |
"features": self.config.features, | |
}, | |
), | |
] | |
else: | |
return [ | |
datasets.SplitGenerator( | |
name=datasets.Split.TRAIN, | |
gen_kwargs={ | |
"filepath": data_file, | |
"split": "train", | |
"features": self.config.features, | |
}, | |
), | |
] | |
def _generate_examples(self, filepath: str, split: str, features: List[str]): | |
"""Yields examples as (key, example) tuples.""" | |
with open(filepath, encoding="utf8") as f: | |
for id_, row in enumerate(f): | |
if id_ == 0: | |
continue | |
ex_split = None | |
fields = row.strip().split("\t") | |
if len(fields) < 6: | |
ex_split = fields[-1] | |
fields = fields[:-1] | |
if ex_split is None or ex_split.strip() == split: | |
yield id_, { | |
k:v.strip() for k,v in zip(features, fields) | |
} | |