RiSAWOZ.py

# coding=utf-8
# 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.
"""RiSAWOZ: A Large-Scale Multi-Domain Wizard-of-Oz Dataset with Rich Semantic Annotations for Task-Oriented Dialogue Modeling"""


import json
import os
from typing import Dict

import datasets


_CITATION = """\
@inproceedings{quan-etal-2020-risawoz,
    title = "{R}i{SAWOZ}: A Large-Scale Multi-Domain {W}izard-of-{O}z Dataset with Rich Semantic Annotations for Task-Oriented Dialogue Modeling",
    author = "Quan, Jun  and
      Zhang, Shian  and
      Cao, Qian  and
      Li, Zizhong  and
      Xiong, Deyi",
    booktitle = "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)",
    month = nov,
    year = "2020",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://www.aclweb.org/anthology/2020.emnlp-main.67",
    pages = "930--940",
}
"""


_DESCRIPTION = """\
RiSAWOZ contains 11.2K human-to-human (H2H) multiturn semantically annotated dialogues, \
with more than 150K utterances spanning over 12 domains, \
which is larger than all previous annotated H2H conversational datasets.\
Both single- and multi-domain dialogues are constructed, accounting for 65% and 35%, respectively.
"""

_HOMEPAGE = "https://github.com/terryqj0107/RiSAWOZ"

_LICENSE = "Attribution 4.0 International (CC BY 4.0) license."

_EMPTY_BELIEF_STATE = [
    "旅游景点-名称",
    "旅游景点-区域",
    "旅游景点-景点类型",
    "旅游景点-最适合人群",
    "旅游景点-消费",
    "旅游景点-是否地铁直达",
    "旅游景点-门票价格",
    "旅游景点-电话号码",
    "旅游景点-地址",
    "旅游景点-评分",
    "旅游景点-开放时间",
    "旅游景点-特点",
    "餐厅-名称",
    "餐厅-区域",
    "餐厅-菜系",
    "餐厅-价位",
    "餐厅-是否地铁直达",
    "餐厅-人均消费",
    "餐厅-地址",
    "餐厅-电话号码",
    "餐厅-评分",
    "餐厅-营业时间",
    "餐厅-推荐菜",
    "酒店-名称",
    "酒店-区域",
    "酒店-星级",
    "酒店-价位",
    "酒店-酒店类型",
    "酒店-房型",
    "酒店-停车场",
    "酒店-房费",
    "酒店-地址",
    "酒店-电话号码",
    "酒店-评分",
    "电脑-品牌",
    "电脑-产品类别",
    "电脑-分类",
    "电脑-内存容量",
    "电脑-屏幕尺寸",
    "电脑-CPU",
    "电脑-价格区间",
    "电脑-系列",
    "电脑-商品名称",
    "电脑-系统",
    "电脑-游戏性能",
    "电脑-CPU型号",
    "电脑-裸机重量",
    "电脑-显卡类别",
    "电脑-显卡型号",
    "电脑-特性",
    "电脑-色系",
    "电脑-待机时长",
    "电脑-硬盘容量",
    "电脑-价格",
    "火车-出发地",
    "火车-目的地",
    "火车-日期",
    "火车-车型",
    "火车-坐席",
    "火车-车次信息",
    "火车-时长",
    "火车-出发时间",
    "火车-到达时间",
    "火车-票价",
    "飞机-出发地",
    "飞机-目的地",
    "飞机-日期",
    "飞机-舱位档次",
    "飞机-航班信息",
    "飞机-起飞时间",
    "飞机-到达时间",
    "飞机-票价",
    "飞机-准点率",
    "天气-城市",
    "天气-日期",
    "天气-天气",
    "天气-温度",
    "天气-风力风向",
    "天气-紫外线强度",
    "电影-制片国家/地区",
    "电影-类型",
    "电影-年代",
    "电影-主演",
    "电影-导演",
    "电影-片名",
    "电影-主演名单",
    "电影-具体上映时间",
    "电影-片长",
    "电影-豆瓣评分",
    "电视剧-制片国家/地区",
    "电视剧-类型",
    "电视剧-年代",
    "电视剧-主演",
    "电视剧-导演",
    "电视剧-片名",
    "电视剧-主演名单",
    "电视剧-首播时间",
    "电视剧-集数",
    "电视剧-单集片长",
    "电视剧-豆瓣评分",
    "辅导班-班号",
    "辅导班-难度",
    "辅导班-科目",
    "辅导班-年级",
    "辅导班-区域",
    "辅导班-校区",
    "辅导班-上课方式",
    "辅导班-开始日期",
    "辅导班-结束日期",
    "辅导班-每周",
    "辅导班-上课时间",
    "辅导班-下课时间",
    "辅导班-时段",
    "辅导班-课次",
    "辅导班-课时",
    "辅导班-教室地点",
    "辅导班-教师",
    "辅导班-价格",
    "辅导班-课程网址",
    "辅导班-教师网址",
    "汽车-名称",
    "汽车-车型",
    "汽车-级别",
    "汽车-座位数",
    "汽车-车身尺寸(mm)",
    "汽车-厂商",
    "汽车-能源类型",
    "汽车-发动机排量(L)",
    "汽车-发动机马力(Ps)",
    "汽车-驱动方式",
    "汽车-综合油耗(L/100km)",
    "汽车-环保标准",
    "汽车-驾驶辅助影像",
    "汽车-巡航系统",
    "汽车-价格(万元)",
    "汽车-车系",
    "汽车-动力水平",
    "汽车-油耗水平",
    "汽车-倒车影像",
    "汽车-定速巡航",
    "汽车-座椅加热",
    "汽车-座椅通风",
    "汽车-所属价格区间",
    "医院-名称",
    "医院-等级",
    "医院-类别",
    "医院-性质",
    "医院-区域",
    "医院-地址",
    "医院-电话",
    "医院-挂号时间",
    "医院-门诊时间",
    "医院-公交线路",
    "医院-地铁可达",
    "医院-地铁线路",
    "医院-重点科室",
    "医院-CT",
    "医院-3.0T MRI",
    "医院-DSA",
    "通用-产品类别",
    "火车-舱位档次",
    "通用-系列",
    "通用-价格区间",
    "通用-品牌"
]


class RiSAWOZ(datasets.GeneratorBasedBuilder):
    """RiSAWOZ: A Large-Scale Multi-Domain Wizard-of-Oz Dataset with Rich Semantic Annotations for Task-Oriented Dialogue Modeling"""

    VERSION = datasets.Version("1.1.0")

    def _info(self):
        features = datasets.Features(
            {
                "dialogue_id": datasets.Value("string"),
                "goal": datasets.Value("string"),
                "domains": datasets.Sequence(datasets.Value("string")),
                "dialogue": datasets.Sequence(
                    {
                        "turn_id": datasets.Value("int32"),
                        "turn_domain": datasets.Sequence(datasets.Value("string")),
                        "user_utterance": datasets.Value("string"),
                        "system_utterance": datasets.Value("string"),
                        "belief_state": {
                            "inform slot-values": {
                                d: datasets.Value("string") for d in _EMPTY_BELIEF_STATE
                            },
                            # "inform slot-values": datasets.Value("string"),
                            "turn_inform": {
                                d: datasets.Value("string") for d in _EMPTY_BELIEF_STATE
                            },
                            "turn request": datasets.Sequence(datasets.Value("string")),
                        },
                        "user_actions": datasets.Sequence(
                            datasets.Sequence(datasets.Value("string"))
                        ),
                        "system_actions": datasets.Sequence(
                            datasets.Sequence(datasets.Value("string"))
                        ),
                        "db_results": datasets.Sequence(datasets.Value("string")),
                        "segmented_user_utterance": datasets.Value("string"),
                        "segmented_system_utterance": datasets.Value("string"),
                    }
                ),
            }
        )

        return datasets.DatasetInfo(
            # This is the description that will appear on the datasets page.
            description=_DESCRIPTION,
            # This defines the different columns of the dataset and their types
            features=features,  # Here we define them above because they are different between the two configurations
            # 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,
            # License for the dataset if available
            license=_LICENSE,
            # Citation for the dataset
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        """Returns SplitGenerators."""

        # dl_manager is a datasets.download.DownloadManager that can be used to download and extract URLs
        # It can accept any type or nested list/dict and will give back the same structure with the url replaced with path to local files.
        # By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
        _URL = {"train": "train.json", "test": "test.json", "dev": "dev.json"}

        data_dir = dl_manager.download_and_extract(_URL)

        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                # These kwargs will be passed to _generate_examples
                gen_kwargs={
                    "filepath": data_dir["train"],
                    "split": "train",
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.TEST,
                # These kwargs will be passed to _generate_examples
                gen_kwargs={"filepath": data_dir["test"], "split": "test"},
            ),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION,
                # These kwargs will be passed to _generate_examples
                gen_kwargs={
                    "filepath": data_dir["dev"],
                    "split": "dev",
                },
            ),
        ]

    def _generate_examples(
        self,
        filepath,
        split,  # method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
    ):
        """Yields examples as (key, example) tuples."""
        # This method handles input defined in _split_generators to yield (key, example) tuples from the dataset.
        # The `key` is here for legacy reason (tfds) and is not important in itself.

        with open(filepath, encoding="utf-8") as f:
            all_data = json.load(f)
            id_ = 0
            for data in all_data:
                for slot in _EMPTY_BELIEF_STATE:
                    for dia in data["dialogue"]:
                        if slot not in dia["belief_state"]["inform slot-values"]:
                            dia["belief_state"]["inform slot-values"][slot] = ""
                        if slot not in dia["belief_state"]["turn_inform"]:
                            dia["belief_state"]["turn_inform"][slot] = ""

                yield id_, {
                    "dialogue_id": data["dialogue_id"],
                    "goal": data["goal"],
                    "domains": data["domains"],
                    "dialogue": data["dialogue"],
                }
                id_ += 1