File size: 6,799 Bytes
76d43f9 b5b8a87 76d43f9 9c5cfab 76d43f9 9c5cfab 76d43f9 9c5cfab 76d43f9 9c5cfab 76d43f9 9c5cfab 76d43f9 9c5cfab 76d43f9 9c5cfab b5b8a87 9c5cfab b5b8a87 9c5cfab |
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 |
"""TODO(blended_skill_talk): Add a description here."""
import json
import datasets
# TODO(blended_skill_talk): BibTeX citation
_CITATION = """\
@misc{smith2020evaluating,
title={Can You Put it All Together: Evaluating Conversational Agents' Ability to Blend Skills},
author={Eric Michael Smith and Mary Williamson and Kurt Shuster and Jason Weston and Y-Lan Boureau},
year={2020},
eprint={2004.08449},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
"""
# TODO(blended_skill_talk):
_DESCRIPTION = """\
A dataset of 7k conversations explicitly designed to exhibit multiple conversation modes: displaying personality, having empathy, and demonstrating knowledge.
"""
_URL = "http://parl.ai/downloads/blended_skill_talk/blended_skill_talk.tar.gz"
_TASK = ["convai2", "empathetic_dialogues", "wizard_of_wikipedia"]
class BlendedSkillTalk(datasets.GeneratorBasedBuilder):
"""TODO(blended_skill_talk): Short description of my dataset."""
# TODO(blended_skill_talk): Set up version.
VERSION = datasets.Version("1.0.0")
def _info(self):
# TODO(blended_skill_talk): Specifies the datasets.DatasetInfo object
return datasets.DatasetInfo(
# This is the description that will appear on the datasets page.
description=_DESCRIPTION,
# datasets.features.FeatureConnectors
features=datasets.Features(
{
"personas": datasets.features.Sequence(datasets.Value("string")),
"additional_context": datasets.Value("string"),
"previous_utterance": datasets.features.Sequence(datasets.Value("string")),
"context": datasets.Value("string"),
"free_messages": datasets.features.Sequence(datasets.Value("string")),
"guided_messages": datasets.features.Sequence(datasets.Value("string")),
"suggestions": datasets.features.Sequence({task: datasets.Value("string") for task in _TASK}),
"guided_chosen_suggestions": datasets.features.Sequence(datasets.Value("string")),
"label_candidates": datasets.features.Sequence(
datasets.features.Sequence(datasets.Value("string"))
),
# These are the features of your dataset like images, labels ...
}
),
# 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="https://parl.ai/projects/bst/",
citation=_CITATION,
)
def _split_generators(self, dl_manager):
"""Returns SplitGenerators."""
# TODO(blended_skill_talk): Downloads the data and defines the splits
# dl_manager is a datasets.download.DownloadManager that can be used to
# download and extract URLs
archive = dl_manager.download(_URL)
return [
datasets.SplitGenerator(
name=datasets.Split.TRAIN,
# These kwargs will be passed to _generate_examples
gen_kwargs={
"filepath": "train.json",
"files": dl_manager.iter_archive(archive),
},
),
datasets.SplitGenerator(
name=datasets.Split.VALIDATION,
# These kwargs will be passed to _generate_examples
gen_kwargs={
"filepath": "valid.json",
"files": dl_manager.iter_archive(archive),
},
),
datasets.SplitGenerator(
name=datasets.Split.TEST,
# These kwargs will be passed to _generate_examples
gen_kwargs={
"filepath": "test.json",
"files": dl_manager.iter_archive(archive),
},
),
]
def _generate_examples(self, filepath, files):
"""Yields examples."""
# TODO(blended_skill_talk): Yields (key, example) tuples from the dataset
for path, f in files:
if path == filepath:
data = json.load(f)
for id_, row in enumerate(data):
personas = [row["personas"][1][0], row["personas"][1][1]]
dialogs = [dialog[1] for dialog in row["dialog"]]
free_messages = []
guided_messages = []
for i in range(len(dialogs) // 2):
free_messages.append(dialogs[2 * i])
guided_messages.append(dialogs[2 * i + 1])
context = row["context_dataset"]
add_context = row["additional_context"] if context == "wizard_of_wikipedia" else ""
previous_utterance = [row["free_turker_utterance"], row["guided_turker_utterance"]]
suggestions = row["suggestions"]
convai_suggestions = []
empathetic_suggestions = []
wow_suggestions = []
for i in range(len(suggestions) // 2):
convai_suggestions.append(suggestions[2 * i + 1]["convai2"])
empathetic_suggestions.append(suggestions[2 * i + 1]["empathetic_dialogues"])
wow_suggestions.append(suggestions[2 * i + 1]["wizard_of_wikipedia"])
chosen_suggestions = row["chosen_suggestions"]
guided_chosen_suggestions = []
for i in range(len(chosen_suggestions) // 2):
guided_chosen_suggestions.append(chosen_suggestions[2 * i + 1])
label_candidates = row["label_candidates"] if "label_candidates" in row else []
yield id_, {
"personas": personas,
"additional_context": add_context,
"previous_utterance": previous_utterance,
"context": context,
"free_messages": free_messages,
"guided_messages": guided_messages,
"suggestions": {
"convai2": convai_suggestions,
"empathetic_dialogues": empathetic_suggestions,
"wizard_of_wikipedia": wow_suggestions,
},
"guided_chosen_suggestions": guided_chosen_suggestions,
"label_candidates": label_candidates,
}
break
|