old_bailey_proceedings.py

# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import os
import datasets
import glob
import xml.etree.ElementTree as ET

_CITATION = """
@misc{Old Bailey Proceedings,
  author    = {Mariona Coll Ardanuy and
               Federico Nanni and
               Kaspar Beelen and
               Kasra Hosseini and
               Ruth Ahnert and
               Jon Lawrence and
               Katherine McDonough and
               Giorgia Tolfo and
               Daniel C. S. Wilson and
               Barbara McGillivray},
  title     = {Living Machines: {A} study of atypical animacy},
  journal   = {CoRR},
  volume    = {abs/2005.11140},
  year      = {2020},
  url       = {https://arxiv.org/abs/2005.11140},
  eprinttype = {arXiv},
  eprint    = {2005.11140},
  timestamp = {Sat, 23 Jan 2021 01:12:25 +0100},
  biburl    = {https://dblp.org/rec/journals/corr/abs-2005-11140.bib},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}
"""


_DESCRIPTION = """The dataset consists of 2,163 transcriptions of the Proceedings and 475 Ordinary's Accounts marked up in TEI-XML, 
and contains some documentation covering the data structure and variables. Each Proceedings file represents one session of the court (1674-1913), 
and each Ordinary's Account file represents a single pamphlet (1676-1772)
"""

_HOMEPAGE = "https://www.dhi.ac.uk/projects/old-bailey/"

_DATASETNAME = "old_bailey_proceedings"

_LICENSE = "Creative Commons Attribution 4.0 International"

_URLS = {
    _DATASETNAME: "https://www.dhi.ac.uk/san/data/oldbailey/oldbailey.zip",
}

logger = datasets.utils.logging.get_logger(__name__)


class OldBaileyProceedings(datasets.GeneratorBasedBuilder):
    """The dataset consists of 2,163 transcriptions of the Proceedings and 475 Ordinary's Accounts marked up in TEI-XML,
    and contains some documentation covering the data structure and variables. Each Proceedings file represents one session of the court (1674-1913),
     and each Ordinary's Account file represents a single pamphlet (1676-1772)"""

    VERSION = datasets.Version("7.2.0")

    def _info(self):
        features = datasets.Features(
            {
                "id": datasets.Value("string"),
                "text": datasets.Value("string"),
                "places": datasets.Sequence(datasets.Value("string")),
                "type": datasets.Value("string"),
                "persons": datasets.Sequence(datasets.Value("string")),
                "date": datasets.Value("string"),
            }
        )
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=features,
            homepage=_HOMEPAGE,
            license=_LICENSE,
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        urls = _URLS[_DATASETNAME]
        data_dir = dl_manager.download_and_extract(urls)
        oa_dir = "ordinarysAccounts"
        obp_dir = "sessionsPapers"
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={
                    "data_dirs": {
                        "OA": os.path.join(data_dir, oa_dir),
                        "OBP": os.path.join(data_dir, obp_dir),
                    },
                },
            ),
        ]

    def convert_text_to_features(self, file, key):
        if key == "OA":
            root_tag = "p"
        else:
            root_tag = "div1/p"
        try:
            xml_data = ET.parse(file)
            root = xml_data.getroot()
            start = root.find("./text/body/div0")
            id = start.attrib["id"]
            date = start.find("interp[@type='date']").attrib["value"]
            text_parts = []
            places, persons = [], []
            for content in start.findall(root_tag):
                for place in content.findall("placeName"):
                    if place.text:
                        place_name = place.text.replace("\n", "").strip()
                    if place_name:
                        places.append(place.text)
                for person in content.findall("persName"):
                    full_name = []
                    for name_part in person.itertext():
                        name_part = (
                            name_part.replace("\n", "").replace("\t", "").strip()
                        )
                        if name_part:
                            full_name.append(name_part)
                    if full_name:
                        persons.append(" ".join(full_name))
                for text_snippet in content.itertext():
                    text_snippet = (
                        text_snippet.replace("\n", "").replace("\t", "").strip()
                    )
                    if text_snippet:
                        text_parts.append(text_snippet)
            full_text = " ".join(text_parts)
            return 0, {
                "id": id,
                "date": date,
                "type": key,
                "places": places,
                "persons": persons,
                "text": full_text,
            }
        except Exception as e:
            return -1, repr(e)

    def _generate_examples(self, data_dirs):
        for key, data_dir in data_dirs.items():
            for file in glob.glob(os.path.join(data_dir, "*.xml")):
                status_code, ret_val = self.convert_text_to_features(file, key)
                if status_code:
                    logger.warn(f"{file}:{ret_val}")
                    input()
                    continue
                else:
                    yield ret_val["id"], ret_val