Datasets:
Delete loading script
Browse files- cmu_hinglish_dog.py +0 -190
cmu_hinglish_dog.py
DELETED
|
@@ -1,190 +0,0 @@
|
|
| 1 |
-
# coding=utf-8
|
| 2 |
-
# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
|
| 3 |
-
#
|
| 4 |
-
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 5 |
-
# you may not use this file except in compliance with the License.
|
| 6 |
-
# You may obtain a copy of the License at
|
| 7 |
-
#
|
| 8 |
-
# http://www.apache.org/licenses/LICENSE-2.0
|
| 9 |
-
#
|
| 10 |
-
# Unless required by applicable law or agreed to in writing, software
|
| 11 |
-
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 12 |
-
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 13 |
-
# See the License for the specific language governing permissions and
|
| 14 |
-
# limitations under the License.
|
| 15 |
-
|
| 16 |
-
import json
|
| 17 |
-
import os
|
| 18 |
-
import re
|
| 19 |
-
|
| 20 |
-
import datasets
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
_CITATION = """\
|
| 24 |
-
@inproceedings{cmu_dog_emnlp18,
|
| 25 |
-
title={A Dataset for Document Grounded Conversations},
|
| 26 |
-
author={Zhou, Kangyan and Prabhumoye, Shrimai and Black, Alan W},
|
| 27 |
-
year={2018},
|
| 28 |
-
booktitle={Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing}
|
| 29 |
-
}
|
| 30 |
-
|
| 31 |
-
@inproceedings{khanuja-etal-2020-gluecos,
|
| 32 |
-
title = "{GLUEC}o{S}: An Evaluation Benchmark for Code-Switched {NLP}",
|
| 33 |
-
author = "Khanuja, Simran and
|
| 34 |
-
Dandapat, Sandipan and
|
| 35 |
-
Srinivasan, Anirudh and
|
| 36 |
-
Sitaram, Sunayana and
|
| 37 |
-
Choudhury, Monojit",
|
| 38 |
-
booktitle = "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics",
|
| 39 |
-
month = jul,
|
| 40 |
-
year = "2020",
|
| 41 |
-
address = "Online",
|
| 42 |
-
publisher = "Association for Computational Linguistics",
|
| 43 |
-
url = "https://www.aclweb.org/anthology/2020.acl-main.329",
|
| 44 |
-
pages = "3575--3585"
|
| 45 |
-
}
|
| 46 |
-
"""
|
| 47 |
-
|
| 48 |
-
_DESCRIPTION = """\
|
| 49 |
-
This is a collection of text conversations in Hinglish (code mixing between Hindi-English) and their corresponding English only versions. Can be used for Translating between the two.
|
| 50 |
-
"""
|
| 51 |
-
|
| 52 |
-
_HOMEPAGE = "http://festvox.org/cedar/data/notyet/"
|
| 53 |
-
_URL_HINGLISH = "http://festvox.org/cedar/data/notyet/CMUHinglishDoG.zip"
|
| 54 |
-
# From: https://github.com/festvox/datasets-CMU_DoG/archive/master/Conversations.zip
|
| 55 |
-
_URL_ENGLISH = "data-english.zip"
|
| 56 |
-
|
| 57 |
-
|
| 58 |
-
class CMUHinglishDoG(datasets.GeneratorBasedBuilder):
|
| 59 |
-
"""Load the CMU Hinglish DoG Data for MT"""
|
| 60 |
-
|
| 61 |
-
def _info(self):
|
| 62 |
-
features = datasets.Features(
|
| 63 |
-
{
|
| 64 |
-
"date": datasets.Value("string"),
|
| 65 |
-
"docIdx": datasets.Value("int64"),
|
| 66 |
-
"translation": datasets.Translation(languages=["en", "hi_en"]),
|
| 67 |
-
"uid": datasets.Value("string"),
|
| 68 |
-
"utcTimestamp": datasets.Value("string"),
|
| 69 |
-
"rating": datasets.Value("int64"),
|
| 70 |
-
"status": datasets.Value("int64"),
|
| 71 |
-
"uid1LogInTime": datasets.Value("string"),
|
| 72 |
-
"uid1LogOutTime": datasets.Value("string"),
|
| 73 |
-
"uid1response": {
|
| 74 |
-
"response": datasets.Sequence(datasets.Value("int64")),
|
| 75 |
-
"type": datasets.Value("string"),
|
| 76 |
-
},
|
| 77 |
-
"uid2response": {
|
| 78 |
-
"response": datasets.Sequence(datasets.Value("int64")),
|
| 79 |
-
"type": datasets.Value("string"),
|
| 80 |
-
},
|
| 81 |
-
"user2_id": datasets.Value("string"),
|
| 82 |
-
"whoSawDoc": datasets.Sequence(datasets.Value("string")),
|
| 83 |
-
"wikiDocumentIdx": datasets.Value("int64"),
|
| 84 |
-
}
|
| 85 |
-
)
|
| 86 |
-
return datasets.DatasetInfo(
|
| 87 |
-
description=_DESCRIPTION,
|
| 88 |
-
features=features,
|
| 89 |
-
supervised_keys=None,
|
| 90 |
-
homepage=_HOMEPAGE,
|
| 91 |
-
citation=_CITATION,
|
| 92 |
-
)
|
| 93 |
-
|
| 94 |
-
def _split_generators(self, dl_manager):
|
| 95 |
-
"""The linking part between Hinglish data and English data is inspired from the implementation in GLUECoS.
|
| 96 |
-
Refer here for the original script https://github.com/microsoft/GLUECoS/blob/7fdc51653e37a32aee17505c47b7d1da364fa77e/Data/Preprocess_Scripts/preprocess_mt_en_hi.py"""
|
| 97 |
-
|
| 98 |
-
eng_path = dl_manager.download_and_extract(_URL_ENGLISH)
|
| 99 |
-
data_dir_en = os.path.join(eng_path, "Conversations")
|
| 100 |
-
|
| 101 |
-
hi_en_path = dl_manager.download_and_extract(_URL_HINGLISH)
|
| 102 |
-
data_dir_hi_en = os.path.join(hi_en_path, "CMUHinglishDoG", "Conversations_Hinglish")
|
| 103 |
-
|
| 104 |
-
hi_en_dirs = {
|
| 105 |
-
"train": os.path.join(data_dir_hi_en, "train"),
|
| 106 |
-
"valid": os.path.join(data_dir_hi_en, "valid"),
|
| 107 |
-
"test": os.path.join(data_dir_hi_en, "test"),
|
| 108 |
-
}
|
| 109 |
-
|
| 110 |
-
return [
|
| 111 |
-
datasets.SplitGenerator(
|
| 112 |
-
name=datasets.Split.TRAIN,
|
| 113 |
-
gen_kwargs={
|
| 114 |
-
"hi_en_dir": hi_en_dirs["train"],
|
| 115 |
-
"data_dir_en": data_dir_en,
|
| 116 |
-
},
|
| 117 |
-
),
|
| 118 |
-
datasets.SplitGenerator(
|
| 119 |
-
name=datasets.Split.TEST,
|
| 120 |
-
gen_kwargs={
|
| 121 |
-
"hi_en_dir": hi_en_dirs["test"],
|
| 122 |
-
"data_dir_en": data_dir_en,
|
| 123 |
-
},
|
| 124 |
-
),
|
| 125 |
-
datasets.SplitGenerator(
|
| 126 |
-
name=datasets.Split.VALIDATION,
|
| 127 |
-
gen_kwargs={
|
| 128 |
-
"hi_en_dir": hi_en_dirs["valid"],
|
| 129 |
-
"data_dir_en": data_dir_en,
|
| 130 |
-
},
|
| 131 |
-
),
|
| 132 |
-
]
|
| 133 |
-
|
| 134 |
-
def _generate_examples(self, hi_en_dir, data_dir_en):
|
| 135 |
-
"""Yields examples."""
|
| 136 |
-
english_files_train = os.listdir(os.path.join(data_dir_en, "train"))
|
| 137 |
-
english_files_val = os.listdir(os.path.join(data_dir_en, "valid"))
|
| 138 |
-
english_files_test = os.listdir(os.path.join(data_dir_en, "test"))
|
| 139 |
-
|
| 140 |
-
hinglish_files = os.listdir(hi_en_dir)
|
| 141 |
-
key = 0
|
| 142 |
-
for f in hinglish_files:
|
| 143 |
-
en_file_path = f.split(".json")[0] + ".json"
|
| 144 |
-
found = True
|
| 145 |
-
# Looks for the corresponding english file in all 3 splits
|
| 146 |
-
if en_file_path in english_files_train:
|
| 147 |
-
en = json.load(open(os.path.join(os.path.join(data_dir_en, "train"), en_file_path)))
|
| 148 |
-
elif en_file_path in english_files_val:
|
| 149 |
-
en = json.load(open(os.path.join(os.path.join(data_dir_en, "valid"), en_file_path)))
|
| 150 |
-
elif en_file_path in english_files_test:
|
| 151 |
-
en = json.load(open(os.path.join(os.path.join(data_dir_en, "test"), en_file_path)))
|
| 152 |
-
else:
|
| 153 |
-
found = False
|
| 154 |
-
if found:
|
| 155 |
-
hi_en = json.load(open(os.path.join(hi_en_dir, f)))
|
| 156 |
-
|
| 157 |
-
assert len(en["history"]) == len(hi_en["history"])
|
| 158 |
-
|
| 159 |
-
for x, y in zip(en["history"], hi_en["history"]):
|
| 160 |
-
assert x["docIdx"] == y["docIdx"]
|
| 161 |
-
assert x["uid"] == y["uid"]
|
| 162 |
-
assert x["utcTimestamp"] == y["utcTimestamp"]
|
| 163 |
-
|
| 164 |
-
x["text"] = re.sub("\t|\n", " ", x["text"])
|
| 165 |
-
y["text"] = re.sub("\t|\n", " ", y["text"])
|
| 166 |
-
line = {
|
| 167 |
-
"date": hi_en["date"],
|
| 168 |
-
"uid": x["uid"],
|
| 169 |
-
"docIdx": x["docIdx"],
|
| 170 |
-
"utcTimestamp": x["utcTimestamp"],
|
| 171 |
-
"translation": {"hi_en": y["text"], "en": x["text"]},
|
| 172 |
-
"rating": hi_en["rating"],
|
| 173 |
-
"status": hi_en["status"],
|
| 174 |
-
"uid1LogOutTime": hi_en.get("uid1LogOutTime"),
|
| 175 |
-
"uid1LogInTime": hi_en["uid1LogInTime"],
|
| 176 |
-
"uid1response": {
|
| 177 |
-
"response": hi_en["uid1response"]["response"] if "uid1response" in hi_en else [],
|
| 178 |
-
"type": hi_en["uid1response"]["type"] if "uid1response" in hi_en else None,
|
| 179 |
-
},
|
| 180 |
-
"uid2response": {
|
| 181 |
-
"response": hi_en["uid2response"]["response"] if "uid2response" in hi_en else [],
|
| 182 |
-
"type": hi_en["uid2response"]["type"] if "uid2response" in hi_en else None,
|
| 183 |
-
},
|
| 184 |
-
"user2_id": hi_en["user2_id"],
|
| 185 |
-
"whoSawDoc": hi_en["whoSawDoc"],
|
| 186 |
-
"wikiDocumentIdx": hi_en["wikiDocumentIdx"],
|
| 187 |
-
}
|
| 188 |
-
|
| 189 |
-
yield key, line
|
| 190 |
-
key += 1
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|