# 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. """MultiWOZ v2.2: Multi-domain Wizard of OZ version 2.2""" import json import datasets # TODO: Add BibTeX citation # Find for instance the citation on arxiv or on the dataset repo/website _CITATION = """\ @article{corr/abs-2007-12720, author = {Xiaoxue Zang and Abhinav Rastogi and Srinivas Sunkara and Raghav Gupta and Jianguo Zhang and Jindong Chen}, title = {MultiWOZ 2.2 : {A} Dialogue Dataset with Additional Annotation Corrections and State Tracking Baselines}, journal = {CoRR}, volume = {abs/2007.12720}, year = {2020}, url = {https://arxiv.org/abs/2007.12720}, archivePrefix = {arXiv}, eprint = {2007.12720} } """ # TODO: Add description of the dataset here # You can copy an official description _DESCRIPTION = """\ Multi-Domain Wizard-of-Oz dataset (MultiWOZ), a fully-labeled collection of human-human written conversations spanning over multiple domains and topics. MultiWOZ 2.1 (Eric et al., 2019) identified and fixed many erroneous annotations and user utterances in the original version, resulting in an improved version of the dataset. MultiWOZ 2.2 is a yet another improved version of this dataset, which identifies and fizes dialogue state annotation errors across 17.3% of the utterances on top of MultiWOZ 2.1 and redefines the ontology by disallowing vocabularies of slots with a large number of possible values (e.g., restaurant name, time of booking) and introducing standardized slot span annotations for these slots. """ _LICENSE = "Apache License 2.0" # TODO: Add link to the official dataset URLs here # The HuggingFace dataset library don't host the datasets but only point to the original files # This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method) _URL_LIST = [ ("dialogue_acts", "https://github.com/budzianowski/multiwoz/raw/master/data/MultiWOZ_2.2/dialog_acts.json") ] _URL_LIST += [ ( f"train_{i:03d}", f"https://github.com/budzianowski/multiwoz/raw/master/data/MultiWOZ_2.2/train/dialogues_{i:03d}.json", ) for i in range(1, 18) ] _URL_LIST += [ ( f"dev_{i:03d}", f"https://github.com/budzianowski/multiwoz/raw/master/data/MultiWOZ_2.2/dev/dialogues_{i:03d}.json", ) for i in range(1, 3) ] _URL_LIST += [ ( f"test_{i:03d}", f"https://github.com/budzianowski/multiwoz/raw/master/data/MultiWOZ_2.2/test/dialogues_{i:03d}.json", ) for i in range(1, 3) ] _URLs = dict(_URL_LIST) class MultiWozV22(datasets.GeneratorBasedBuilder): VERSION = datasets.Version("2.2.0") BUILDER_CONFIGS = [ datasets.BuilderConfig(name="v2.2", version=datasets.Version("2.2.0"), description="MultiWOZ v2.2"), datasets.BuilderConfig( name="v2.2_active_only", version=datasets.Version("2.2.0"), description="MultiWOZ v2.2, only keeps around frames with an active intent", ), ] DEFAULT_CONFIG_NAME = "v2.2_active_only" def _info(self): features = datasets.Features( { "dialogue_id": datasets.Value("string"), "services": datasets.Sequence(datasets.Value("string")), "turns": datasets.Sequence( { "turn_id": datasets.Value("string"), "speaker": datasets.ClassLabel(names=["USER", "SYSTEM"]), "utterance": datasets.Value("string"), "frames": datasets.Sequence( { "service": datasets.Value("string"), "state": { "active_intent": datasets.Value("string"), "requested_slots": datasets.Sequence(datasets.Value("string")), "slots_values": datasets.Sequence( { "slots_values_name": datasets.Value("string"), "slots_values_list": datasets.Sequence(datasets.Value("string")), } ), }, "slots": datasets.Sequence( { "slot": datasets.Value("string"), "value": datasets.Value("string"), "start": datasets.Value("int32"), "exclusive_end": datasets.Value("int32"), "copy_from": datasets.Value("string"), "copy_from_value": datasets.Sequence(datasets.Value("string")), } ), } ), "dialogue_acts": datasets.Features( { "dialog_act": datasets.Sequence( { "act_type": datasets.Value("string"), "act_slots": datasets.Sequence( datasets.Features( { "slot_name": datasets.Value("string"), "slot_value": datasets.Value("string"), } ), ), } ), "span_info": datasets.Sequence( { "act_type": datasets.Value("string"), "act_slot_name": datasets.Value("string"), "act_slot_value": datasets.Value("string"), "span_start": datasets.Value("int32"), "span_end": datasets.Value("int32"), } ), } ), } ), } ) return datasets.DatasetInfo( description=_DESCRIPTION, features=features, # Here we define them above because they are different between the two configurations supervised_keys=None, homepage="https://github.com/budzianowski/multiwoz/tree/master/data/MultiWOZ_2.2", license=_LICENSE, citation=_CITATION, ) def _split_generators(self, dl_manager): data_files = dl_manager.download_and_extract(_URLs) self.stored_dialogue_acts = json.load(open(data_files["dialogue_acts"])) return [ datasets.SplitGenerator( name=spl_enum, gen_kwargs={ "filepaths": data_files, "split": spl, }, ) for spl, spl_enum in [ ("train", datasets.Split.TRAIN), ("dev", datasets.Split.VALIDATION), ("test", datasets.Split.TEST), ] ] def _generate_examples(self, filepaths, split): id_ = -1 file_list = [fpath for fname, fpath in filepaths.items() if fname.startswith(split)] for filepath in file_list: dialogues = json.load(open(filepath)) for dialogue in dialogues: id_ += 1 mapped_acts = self.stored_dialogue_acts.get(dialogue["dialogue_id"], {}) res = { "dialogue_id": dialogue["dialogue_id"], "services": dialogue["services"], "turns": [ { "turn_id": turn["turn_id"], "speaker": turn["speaker"], "utterance": turn["utterance"], "frames": [ { "service": frame["service"], "state": { "active_intent": frame["state"]["active_intent"] if "state" in frame else "", "requested_slots": frame["state"]["requested_slots"] if "state" in frame else [], "slots_values": { "slots_values_name": [ sv_name for sv_name, sv_list in frame["state"]["slot_values"].items() ] if "state" in frame else [], "slots_values_list": [ sv_list for sv_name, sv_list in frame["state"]["slot_values"].items() ] if "state" in frame else [], }, }, "slots": [ { "slot": slot["slot"], "value": "" if "copy_from" in slot else slot["value"], "start": slot.get("start", -1), "exclusive_end": slot.get("exclusive_end", -1), "copy_from": slot.get("copy_from", ""), "copy_from_value": slot["value"] if "copy_from" in slot else [], } for slot in frame["slots"] ], } for frame in turn["frames"] if ( "active_only" not in self.config.name or frame.get("state", {}).get("active_intent", "NONE") != "NONE" ) ], "dialogue_acts": { "dialog_act": [ { "act_type": act_type, "act_slots": { "slot_name": [sl_name for sl_name, sl_val in dialog_act], "slot_value": [sl_val for sl_name, sl_val in dialog_act], }, } for act_type, dialog_act in mapped_acts.get(turn["turn_id"], {}) .get("dialog_act", {}) .items() ], "span_info": [ { "act_type": span_info[0], "act_slot_name": span_info[1], "act_slot_value": span_info[2], "span_start": span_info[3], "span_end": span_info[4], } for span_info in mapped_acts.get(turn["turn_id"], {}).get("span_info", []) ], }, } for turn in dialogue["turns"] ], } yield id_, res