File size: 6,418 Bytes
9ea5b68 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 |
# Copyright 2022 The HuggingFace Datasets Authors and the current dataset script contributor.
# Dataset license is specified in the folder data/LICENSE.txt
# 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
from unicodedata import name
import datasets
_CITATION = """\
@article{bordes2016learning,
title={Learning end-to-end goal-oriented dialog},
author={Bordes, Antoine and Boureau, Y-Lan and Weston, Jason},
journal={arXiv preprint arXiv:1605.07683},
year={2016}
}
"""
_DESCRIPTION = """\
This section presents the set of 6 tasks for testing end-to-end dialog systems in the restaurant domain described in the paper:
Antoine Bordes, Y-Lan Boureau, Jason Weston, Learning End-to-End Goal-Oriented Dialog, arxiv:1605.07683.
Each task tests a unique aspect of dialog. Tasks are designed to complement the set of 20 bAbI tasks for story understanding of the previous section.
For each task, there are 1000 dialogs for training, 1000 for development and 1000 for testing. For tasks 1-5, we also include a second test set (with suffix -OOV.txt) that contains dialogs including entities not present in training and development sets.
"""
_HOMEPAGE = "https://research.facebook.com/downloads/babi/"
_LICENSE = "data/LICENSE.txt"
DIALOG_BABI_ARCHIVE_URL = "https://www.dropbox.com/s/20rgyj8rryvos9l/dialog-bAbI-tasks-1_6.zip?dl=1"
class DialogBabiDataset(datasets.GeneratorBasedBuilder):
"""Facebook's dataset 'dialog-babi' for end-to-end dialogue learning."""
VERSION = datasets.Version("1.0.0")
BUILDER_CONFIGS = [
datasets.BuilderConfig(name="task1-API-calls", version=VERSION, description="Issuing API calls"),
datasets.BuilderConfig(name="task2-API-refine", version=VERSION, description="Updating API calls"),
datasets.BuilderConfig(name="task3-options", version=VERSION, description="Displaying options"),
datasets.BuilderConfig(name="task4-phone-address", version=VERSION, description="Providing extra information"),
datasets.BuilderConfig(name="task5-full-dialogs", version=VERSION, description="Conducting full dialogs"),
datasets.BuilderConfig(name="task6-dstc2", version=VERSION, description="2nd Dialog State Tracking Challenge (Henderson et al., 2014a)"),
]
DEFAULT_CONFIG_NAME = "task1-API-calls"
def _info(self):
if self.config.name == "task1-API-calls":
features = datasets.Features(
{
"user_turns": datasets.Sequence(datasets.Value("string")),
"system_turns": datasets.Sequence(datasets.Value("string"))
}
)
else:
features = datasets.Features(
{
"user_turns": datasets.Sequence(datasets.Value("string")),
"system_turns": datasets.Sequence(datasets.Value("string"))
}
)
return datasets.DatasetInfo(
description=_DESCRIPTION,
features=features,
homepage=_HOMEPAGE,
license=_LICENSE,
citation=_CITATION,
)
def _split_generators(self, dl_manager):
data_dir = dl_manager.download_and_extract(DIALOG_BABI_ARCHIVE_URL)
task_data_dir = os.path.join(data_dir, "dialog-bAbI-tasks-1_6", self.config.name)
return [
datasets.SplitGenerator(
name=datasets.Split.TRAIN,
# These kwargs will be passed to _generate_examples
gen_kwargs={
"filepath": os.path.join(task_data_dir, f"dialog-babi-{self.config.name}-trn.txt"),
"split": f"{self.config.name}-train",
},
),
datasets.SplitGenerator(
name=datasets.Split.VALIDATION,
# These kwargs will be passed to _generate_examples
gen_kwargs={
"filepath": os.path.join(task_data_dir, f"dialog-babi-{self.config.name}-dev.txt"),
"split": f"{self.config.name}-dev",
},
),
datasets.SplitGenerator(
name=datasets.Split.TEST,
# These kwargs will be passed to _generate_examples
gen_kwargs={
"filepath": os.path.join(task_data_dir, f"dialog-babi-{self.config.name}-tst.txt"),
"split": f"{self.config.name}-test"
},
),
datasets.SplitGenerator(
name=datasets.Split("TESTOOV"),
gen_kwargs={
"filepath": os.path.join(task_data_dir, f"dialog-babi-{self.config.name}-tst-OOV.txt"),
"split": f"{self.config.name}-test-OOV"
},
)
]
def _generate_examples(self, filepath, split):
with open(filepath, encoding="utf-8") as f:
dialogue_rows = []
dialogue_id = 1
for row in f:
turn = row.strip()
if not turn:
yield self._format_dialogue(dialogue_id,split, dialogue_rows)
dialogue_rows.clear()
dialogue_id += 1
else:
dialogue_rows.append(turn)
if dialogue_rows:
yield self._format_dialogue(dialogue_id, split, dialogue_rows)
def _format_dialogue(self, dialogue_id, split, dialogue_rows):
user_turns = []
system_turns = []
for turn in dialogue_rows:
rest_turn, sys_turn = turn.split("\t")
_, user_turn = rest_turn.split(" ", 1)
user_turns.append(user_turn)
system_turns.append(sys_turn)
example_key = f"{split}-{dialogue_id}"
return example_key, {
"user_turns": user_turns,
"system_turns": system_turns
}
|