webnlgqa.py

import os
import zipfile
import json
import base64
import sys
import traceback

import datasets

_CITATION = """\
    @inproceedings{lecorve2022sparql2text,
        title={SPARQL-to-Text Question Generation for Knowledge-Based Conversational Applications},
        author={Lecorv\'e, Gw\'enol\'e and Veyret, Morgan and Brabant, Quentin and Rojas-Barahona, Lina M.},
        journal={Proceedings of the Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the International Joint Conference on Natural Language Processing (AACL-IJCNLP)},
        year={2022}
    }
"""

_HOMEPAGE = ""

_URLS = {
    "train": "train.json",
    "dev": "dev.json",
    "test": "test.json",
    "challenge": "challenge.json"
}

_DESCRIPTION = """\
Augmented version of WebNLG v3.0 English with follow-up SPARQL queries with their associated answer(s). A small portion of it also contains natural language questions associated with the queries.
"""

class WebNLGQA(datasets.GeneratorBasedBuilder):
    """
    WebNLG-QA: Augmented version of WebNLG v3.0 English with follow-up SPARQL queries with their associated answer(s). A small portion of it also contains natural language questions associated with the queries.
    """

    VERSION = datasets.Version("1.0.0")

    def _info(self):
        return datasets.DatasetInfo(
            # This is the description that will appear on the datasets page.
            description=_DESCRIPTION,
            # datasets.features.FeatureConnectors
            features=datasets.Features(
                {
                    "category": datasets.Value("string"),
                    "size": datasets.Value("int32"),
                    "id": datasets.Value("string"),
                    "eid": datasets.Value("string"),
                    "original_triple_sets": [
                        {"subject": datasets.Value("string"),
                         "property": datasets.Value("string"),
                         "object": datasets.Value("string")}
                    ],
                    "modified_triple_sets": [
                        {"subject": datasets.Value("string"),
                         "property": datasets.Value("string"),
                         "object": datasets.Value("string")}
                    ],
                    "shape": datasets.Value("string"),
                    "shape_type": datasets.Value("string"),
                    "lex": datasets.Sequence(
                        {
                            "comment": datasets.Value("string"),
                            "lid": datasets.Value("string"),
                            "text": datasets.Value("string"),
                            "lang": datasets.Value("string"),
                        }
                    ),
                    "test_category": datasets.Value("string"),
                    "dbpedia_links": datasets.Sequence(datasets.Value("string")),
                    "links": datasets.Sequence(datasets.Value("string")),
                    "graph": [
                        [datasets.Value("string")]
                    ],
                    "main_entity": datasets.Value("string"),
                    "mappings": [
                                    {
                                        "modified": datasets.Value("string"),
                                        "readable": datasets.Value("string"),
                                        "graph": datasets.Value("string")
                                    }
                                ],
                    "dialogue": [
                            {
                                "question": [ {
                                  "source": datasets.Value("string"),
                                  "text": datasets.Value("string")
                                }],
                                "graph_query": datasets.Value("string"),
                                "readable_query": datasets.Value("string"),
                                "graph_answer": [
                                    datasets.Value("string")
                                ],
                                "readable_answer": [
                                    datasets.Value("string")
                                ],
                                "type": [ datasets.Value("string") ]
                            }
                        ]
                }
            ),
            # If there's a common (input, target) tuple from the features,
            # specify them here. They'll be used if as_supervised=True in
            # builder.as_dataset
            supervised_keys=None,
            # Homepage of the dataset for documentation
            homepage=_HOMEPAGE,
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        """Returns SplitGenerators."""
        # Downloads the data and defines the splits
        # dl_manager is a datasets.download.DownloadManager that can be used to
        # download and extract URLs
        paths = dl_manager.download_and_extract(_URLS)
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={"filepath": paths['train'],
                            "split": "train"}
            ),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION,
                gen_kwargs={"filepath": paths['dev'],
                            "split": "dev"}
            ),
            datasets.SplitGenerator(
                name=datasets.Split.TEST,
                gen_kwargs={"filepath": paths['test'],
                            "split": "test"}
            ),
            datasets.SplitGenerator(
                name="challenge",
                gen_kwargs={"filepath": paths['challenge'],
                            "split": "challenge"}
            )
        ]


    def _generate_examples(self, filepath, split):
        """Yields examples."""

        def transform_sample(original_sample):
            transformed_sample = {
                    "category": "",
                    "size": -1,
                    "id": "",
                    "eid": "",
                    "original_triple_sets": [],
                    "modified_triple_sets": [],
                    "shape": "",
                    "shape_type": "",
                    "lex": [],
                    "test_category": "",
                    "dbpedia_links": [],
                    "links": [],
                    "graph": [],
                    "main_entity": "",
                    "mappings": [],
                    "dialogue": []
                }

            for (old_key, new_key) in [("modifiedtripleset", "modified_triple_sets"), ("originaltriplesets", "original_triple_sets"), ("dbpedialinks", "dbpedia_links"), ("lexicalisations", "lex"), ("xml_id", "eid")]:
                original_sample[new_key] = original_sample[old_key]
                del original_sample[old_key]

            original_sample["original_triple_sets"] = original_sample["original_triple_sets"]["originaltripleset"][0]

            for l in original_sample["lex"]:
                l["lid"] = l["xml_id"]
                del l["xml_id"]
                l["text"] = l["lex"]
                del l["lex"]

            for turn in original_sample["dialogue"]:
                if "question" in turn:
                    old_format = turn["question"]
                    new_format = []
                    for source, text in old_format.items():
                        new_format.append({"source": source, "text": text})
                    turn["question"] = new_format


            for k in transformed_sample:
                if k in original_sample:
                    transformed_sample[k] = original_sample[k]
            # transformed_sample.update(original_sample)

            return transformed_sample

        # Yields (key, example) tuples from the dataset
        with open(filepath,'r') as f:
            data = json.load(f)
            key = 0
            for it in data:
                yield key, transform_sample(it)
                key += 1