First version of the wino_x dataset.

README.md

@@ -1,3 +1,149 @@
 ---
-license: mit
----
+annotations_creators:
+- no-annotation
+language:
+- en
+- de
+- fr
+- ru
+language_creators:
+- machine-generated
+- expert-generated
+license:
+- mit
+multilinguality:
+- multilingual
+- translation
+pretty_name: Wino-X
+size_categories:
+- 1K<n<10K
+source_datasets:
+- original
+task_categories:
+- translation
+- coreference resolution
+- commonsense reasoning
+- ---
+
+# Dataset Card for Wino-X
+
+## Table of Contents
+- [Table of Contents](#table-of-contents)
+- [Dataset Description](#dataset-description)
+  - [Dataset Summary](#dataset-summary)
+  - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
+  - [Languages](#languages)
+- [Dataset Structure](#dataset-structure)
+  - [Data Instances](#data-instances)
+  - [Data Fields](#data-fields)
+  - [Data Splits](#data-splits)
+- [Dataset Creation](#dataset-creation)
+  - [Curation Rationale](#curation-rationale)
+  - [Source Data](#source-data)
+  - [Annotations](#annotations)
+  - [Personal and Sensitive Information](#personal-and-sensitive-information)
+- [Considerations for Using the Data](#considerations-for-using-the-data)
+  - [Social Impact of Dataset](#social-impact-of-dataset)
+  - [Discussion of Biases](#discussion-of-biases)
+  - [Other Known Limitations](#other-known-limitations)
+- [Additional Information](#additional-information)
+  - [Dataset Curators](#dataset-curators)
+  - [Licensing Information](#licensing-information)
+  - [Citation Information](#citation-information)
+  - [Contributions](#contributions)
+
+## Dataset Description
+
+- **Homepage:**
+- **Repository:**
+- **Paper:**
+- **Leaderboard:**
+- **Point of Contact:**
+
+### Dataset Summary
+
+[More Information Needed]
+
+### Supported Tasks and Leaderboards
+
+[More Information Needed]
+
+### Languages
+
+[More Information Needed]
+
+## Dataset Structure
+
+### Data Instances
+
+[More Information Needed]
+
+### Data Fields
+
+[More Information Needed]
+
+### Data Splits
+
+[More Information Needed]
+
+## Dataset Creation
+
+### Curation Rationale
+
+[More Information Needed]
+
+### Source Data
+
+#### Initial Data Collection and Normalization
+
+[More Information Needed]
+
+#### Who are the source language producers?
+
+[More Information Needed]
+
+### Annotations
+
+#### Annotation process
+
+[More Information Needed]
+
+#### Who are the annotators?
+
+[More Information Needed]
+
+### Personal and Sensitive Information
+
+[More Information Needed]
+
+## Considerations for Using the Data
+
+### Social Impact of Dataset
+
+[More Information Needed]
+
+### Discussion of Biases
+
+[More Information Needed]
+
+### Other Known Limitations
+
+[More Information Needed]
+
+## Additional Information
+
+### Dataset Curators
+
+[More Information Needed]
+
+### Licensing Information
+
+[More Information Needed]
+
+### Citation Information
+
+[More Information Needed]
+
+### Contributions
+
+See the [associated GitHub repository](https://github.com/demelin/Wino-X).

dataset_infos.json

@@ -0,0 +1 @@
+{"mt_en_de": {"description": "Wino-X is a parallel dataset of German, French, and Russian Winograd schemas, aligned with their English \ncounterparts, used to examine whether neural machine translation models can perform coreference resolution that \nrequires commonsense knowledge and whether multilingual language models are capable of commonsense reasoning across \nmultiple languages.\n", "citation": "@inproceedings{Emelin2021WinoXMW,\n  title={Wino-X: Multilingual Winograd Schemas for Commonsense Reasoning and Coreference Resolution},\n  author={Denis Emelin and Rico Sennrich},\n  booktitle={EMNLP},\n  year={2021}\n}\n", "homepage": "https://github.com/demelin/Wino-X", "license": "MIT", "features": {"qID": {"dtype": "string", "id": null, "_type": "Value"}, "sentence": {"dtype": "string", "id": null, "_type": "Value"}, "translation1": {"dtype": "string", "id": null, "_type": "Value"}, "translation2": {"dtype": "string", "id": null, "_type": "Value"}, "answer": {"dtype": "int64", "id": null, "_type": "Value"}, "pronoun1": {"dtype": "string", "id": null, "_type": "Value"}, "pronoun2": {"dtype": "string", "id": null, "_type": "Value"}, "referent1_en": {"dtype": "string", "id": null, "_type": "Value"}, "referent2_en": {"dtype": "string", "id": null, "_type": "Value"}, "true_translation_referent_of_pronoun1_de": {"dtype": "string", "id": null, "_type": "Value"}, "true_translation_referent_of_pronoun2_de": {"dtype": "string", "id": null, "_type": "Value"}, "false_translation_referent_of_pronoun1_de": {"dtype": "string", "id": null, "_type": "Value"}, "false_translation_referent_of_pronoun2_de": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "wino_x", "config_name": "mt_en_de", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 1601793, "num_examples": 3774, "dataset_name": "wino_x"}}, "download_checksums": {"https://huggingface.co/datasets/demelin/wino_x/resolve/main/data/mt/en_de_test.jsonl": {"num_bytes": 2700026, "checksum": "1d18e0dcff2aae54b0ae28a43b61483e97225451d7286e4224657b2b3d93a03a"}}, "download_size": 2700026, "post_processing_size": null, "dataset_size": 1601793, "size_in_bytes": 4301819}, "mt_en_fr": {"description": "Wino-X is a parallel dataset of German, French, and Russian Winograd schemas, aligned with their English \ncounterparts, used to examine whether neural machine translation models can perform coreference resolution that \nrequires commonsense knowledge and whether multilingual language models are capable of commonsense reasoning across \nmultiple languages.\n", "citation": "@inproceedings{Emelin2021WinoXMW,\n  title={Wino-X: Multilingual Winograd Schemas for Commonsense Reasoning and Coreference Resolution},\n  author={Denis Emelin and Rico Sennrich},\n  booktitle={EMNLP},\n  year={2021}\n}\n", "homepage": "https://github.com/demelin/Wino-X", "license": "MIT", "features": {"qID": {"dtype": "string", "id": null, "_type": "Value"}, "sentence": {"dtype": "string", "id": null, "_type": "Value"}, "translation1": {"dtype": "string", "id": null, "_type": "Value"}, "translation2": {"dtype": "string", "id": null, "_type": "Value"}, "answer": {"dtype": "int64", "id": null, "_type": "Value"}, "pronoun1": {"dtype": "string", "id": null, "_type": "Value"}, "pronoun2": {"dtype": "string", "id": null, "_type": "Value"}, "referent1_en": {"dtype": "string", "id": null, "_type": "Value"}, "referent2_en": {"dtype": "string", "id": null, "_type": "Value"}, "true_translation_referent_of_pronoun1_fr": {"dtype": "string", "id": null, "_type": "Value"}, "true_translation_referent_of_pronoun2_fr": {"dtype": "string", "id": null, "_type": "Value"}, "false_translation_referent_of_pronoun1_fr": {"dtype": "string", "id": null, "_type": "Value"}, "false_translation_referent_of_pronoun2_fr": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "wino_x", "config_name": "mt_en_fr", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 1320197, "num_examples": 2988, "dataset_name": "wino_x"}}, "download_checksums": {"https://huggingface.co/datasets/demelin/wino_x/resolve/main/data/mt/en_fr_test.jsonl": {"num_bytes": 2189704, "checksum": "b5b64309b596769e1faabdf2c302c5fe9a0869d972ce39000015f0d8d76c72bb"}}, "download_size": 2189704, "post_processing_size": null, "dataset_size": 1320197, "size_in_bytes": 3509901}, "mt_en_ru": {"description": "Wino-X is a parallel dataset of German, French, and Russian Winograd schemas, aligned with their English \ncounterparts, used to examine whether neural machine translation models can perform coreference resolution that \nrequires commonsense knowledge and whether multilingual language models are capable of commonsense reasoning across \nmultiple languages.\n", "citation": "@inproceedings{Emelin2021WinoXMW,\n  title={Wino-X: Multilingual Winograd Schemas for Commonsense Reasoning and Coreference Resolution},\n  author={Denis Emelin and Rico Sennrich},\n  booktitle={EMNLP},\n  year={2021}\n}\n", "homepage": "https://github.com/demelin/Wino-X", "license": "MIT", "features": {"qID": {"dtype": "string", "id": null, "_type": "Value"}, "sentence": {"dtype": "string", "id": null, "_type": "Value"}, "translation1": {"dtype": "string", "id": null, "_type": "Value"}, "translation2": {"dtype": "string", "id": null, "_type": "Value"}, "answer": {"dtype": "int64", "id": null, "_type": "Value"}, "pronoun1": {"dtype": "string", "id": null, "_type": "Value"}, "pronoun2": {"dtype": "string", "id": null, "_type": "Value"}, "referent1_en": {"dtype": "string", "id": null, "_type": "Value"}, "referent2_en": {"dtype": "string", "id": null, "_type": "Value"}, "true_translation_referent_of_pronoun1_ru": {"dtype": "string", "id": null, "_type": "Value"}, "true_translation_referent_of_pronoun2_ru": {"dtype": "string", "id": null, "_type": "Value"}, "false_translation_referent_of_pronoun1_ru": {"dtype": "string", "id": null, "_type": "Value"}, "false_translation_referent_of_pronoun2_ru": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "wino_x", "config_name": "mt_en_ru", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 1314290, "num_examples": 2238, "dataset_name": "wino_x"}}, "download_checksums": {"https://huggingface.co/datasets/demelin/wino_x/resolve/main/data/mt/en_ru_test.jsonl": {"num_bytes": 1965547, "checksum": "94f232e299b1f27cb97cdae15e25f2e6157b974abfda0754fca91a6d5bea0d7d"}}, "download_size": 1965547, "post_processing_size": null, "dataset_size": 1314290, "size_in_bytes": 3279837}, "lm_en_de": {"description": "Wino-X is a parallel dataset of German, French, and Russian Winograd schemas, aligned with their English \ncounterparts, used to examine whether neural machine translation models can perform coreference resolution that \nrequires commonsense knowledge and whether multilingual language models are capable of commonsense reasoning across \nmultiple languages.\n", "citation": "@inproceedings{Emelin2021WinoXMW,\n  title={Wino-X: Multilingual Winograd Schemas for Commonsense Reasoning and Coreference Resolution},\n  author={Denis Emelin and Rico Sennrich},\n  booktitle={EMNLP},\n  year={2021}\n}\n", "homepage": "https://github.com/demelin/Wino-X", "license": "MIT", "features": {"qID": {"dtype": "string", "id": null, "_type": "Value"}, "sentence": {"dtype": "string", "id": null, "_type": "Value"}, "context_en": {"dtype": "string", "id": null, "_type": "Value"}, "context_de": {"dtype": "string", "id": null, "_type": "Value"}, "answer": {"dtype": "int64", "id": null, "_type": "Value"}, "option1_en": {"dtype": "string", "id": null, "_type": "Value"}, "option2_en": {"dtype": "string", "id": null, "_type": "Value"}, "option1_de": {"dtype": "string", "id": null, "_type": "Value"}, "option2_de": {"dtype": "string", "id": null, "_type": "Value"}, "context_referent_of_option1_de": {"dtype": "string", "id": null, "_type": "Value"}, "context_referent_of_option2_de": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "wino_x", "config_name": "lm_en_de", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 2432303, "num_examples": 5835, "dataset_name": "wino_x"}}, "download_checksums": {"https://huggingface.co/datasets/demelin/wino_x/resolve/main/data/lm/en_de_test.jsonl": {"num_bytes": 3465097, "checksum": "aa91c114832c351c7c4a836041120efc9b5cca2c1620ffcdc5f7f4fa1a7af613"}}, "download_size": 3465097, "post_processing_size": null, "dataset_size": 2432303, "size_in_bytes": 5897400}, "lm_en_fr": {"description": "Wino-X is a parallel dataset of German, French, and Russian Winograd schemas, aligned with their English \ncounterparts, used to examine whether neural machine translation models can perform coreference resolution that \nrequires commonsense knowledge and whether multilingual language models are capable of commonsense reasoning across \nmultiple languages.\n", "citation": "@inproceedings{Emelin2021WinoXMW,\n  title={Wino-X: Multilingual Winograd Schemas for Commonsense Reasoning and Coreference Resolution},\n  author={Denis Emelin and Rico Sennrich},\n  booktitle={EMNLP},\n  year={2021}\n}\n", "homepage": "https://github.com/demelin/Wino-X", "license": "MIT", "features": {"qID": {"dtype": "string", "id": null, "_type": "Value"}, "sentence": {"dtype": "string", "id": null, "_type": "Value"}, "context_en": {"dtype": "string", "id": null, "_type": "Value"}, "context_fr": {"dtype": "string", "id": null, "_type": "Value"}, "answer": {"dtype": "int64", "id": null, "_type": "Value"}, "option1_en": {"dtype": "string", "id": null, "_type": "Value"}, "option2_en": {"dtype": "string", "id": null, "_type": "Value"}, "option1_fr": {"dtype": "string", "id": null, "_type": "Value"}, "option2_fr": {"dtype": "string", "id": null, "_type": "Value"}, "context_referent_of_option1_fr": {"dtype": "string", "id": null, "_type": "Value"}, "context_referent_of_option2_fr": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "wino_x", "config_name": "lm_en_fr", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 1184932, "num_examples": 2793, "dataset_name": "wino_x"}}, "download_checksums": {"https://huggingface.co/datasets/demelin/wino_x/resolve/main/data/lm/en_fr_test.jsonl": {"num_bytes": 1679293, "checksum": "5eb5baeba31db9c20aafb5379e80d14b0a32d69d4d3a9925c4a1ee7e77eb11d1"}}, "download_size": 1679293, "post_processing_size": null, "dataset_size": 1184932, "size_in_bytes": 2864225}, "lm_en_ru": {"description": "Wino-X is a parallel dataset of German, French, and Russian Winograd schemas, aligned with their English \ncounterparts, used to examine whether neural machine translation models can perform coreference resolution that \nrequires commonsense knowledge and whether multilingual language models are capable of commonsense reasoning across \nmultiple languages.\n", "citation": "@inproceedings{Emelin2021WinoXMW,\n  title={Wino-X: Multilingual Winograd Schemas for Commonsense Reasoning and Coreference Resolution},\n  author={Denis Emelin and Rico Sennrich},\n  booktitle={EMNLP},\n  year={2021}\n}\n", "homepage": "https://github.com/demelin/Wino-X", "license": "MIT", "features": {"qID": {"dtype": "string", "id": null, "_type": "Value"}, "sentence": {"dtype": "string", "id": null, "_type": "Value"}, "context_en": {"dtype": "string", "id": null, "_type": "Value"}, "context_ru": {"dtype": "string", "id": null, "_type": "Value"}, "answer": {"dtype": "int64", "id": null, "_type": "Value"}, "option1_en": {"dtype": "string", "id": null, "_type": "Value"}, "option2_en": {"dtype": "string", "id": null, "_type": "Value"}, "option1_ru": {"dtype": "string", "id": null, "_type": "Value"}, "option2_ru": {"dtype": "string", "id": null, "_type": "Value"}, "context_referent_of_option1_ru": {"dtype": "string", "id": null, "_type": "Value"}, "context_referent_of_option2_ru": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "wino_x", "config_name": "lm_en_ru", "version": {"version_str": "1.1.0", "description": null, "major": 1, "minor": 1, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 753838, "num_examples": 1487, "dataset_name": "wino_x"}}, "download_checksums": {"https://huggingface.co/datasets/demelin/wino_x/resolve/main/data/lm/en_ru_test.jsonl": {"num_bytes": 1017037, "checksum": "1bd110d3bc11e5f3034f7040fca56a818607182a52df42dcfa16d606528c2a25"}}, "download_size": 1017037, "post_processing_size": null, "dataset_size": 753838, "size_in_bytes": 1770875}}

wino_x.py

@@ -0,0 +1,168 @@
+# 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.
+""" Wino-X is a parallel dataset of German, French, and Russian Winograd schemas, aligned with their English
+counterparts, used to examine whether neural machine translation models can perform coreference resolution that
+requires commonsense knowledge and whether multilingual language models are capable of commonsense reasoning across
+multiple languages. """
+
+import csv
+import json
+import os
+
+import datasets
+
+_CITATION = """\
+@inproceedings{Emelin2021WinoXMW,
+  title={Wino-X: Multilingual Winograd Schemas for Commonsense Reasoning and Coreference Resolution},
+  author={Denis Emelin and Rico Sennrich},
+  booktitle={EMNLP},
+  year={2021}
+}
+"""
+
+# You can copy an official description
+_DESCRIPTION = """\
+Wino-X is a parallel dataset of German, French, and Russian Winograd schemas, aligned with their English 
+counterparts, used to examine whether neural machine translation models can perform coreference resolution that 
+requires commonsense knowledge and whether multilingual language models are capable of commonsense reasoning across 
+multiple languages.
+"""
+
+_HOMEPAGE = "https://github.com/demelin/Wino-X"
+
+_LICENSE = "MIT"
+
+# The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
+# This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
+_URLS = {
+    "mt_en_de": "https://huggingface.co/datasets/demelin/wino_x/resolve/main/data/mt/en_de_test.jsonl",
+    "mt_en_fr": "https://huggingface.co/datasets/demelin/wino_x/resolve/main/data/mt/en_fr_test.jsonl",
+    "mt_en_ru": "https://huggingface.co/datasets/demelin/wino_x/resolve/main/data/mt/en_ru_test.jsonl",
+    "lm_en_de": "https://huggingface.co/datasets/demelin/wino_x/resolve/main/data/lm/en_de_test.jsonl",
+    "lm_en_fr": "https://huggingface.co/datasets/demelin/wino_x/resolve/main/data/lm/en_fr_test.jsonl",
+    "lm_en_ru": "https://huggingface.co/datasets/demelin/wino_x/resolve/main/data/lm/en_ru_test.jsonl"
+}
+
+
+class WinoX(datasets.GeneratorBasedBuilder):
+    """ Wino-X is a dataset of German, French, and Russian Winograd schemas, aligned with their English counterparts """
+
+    VERSION = datasets.Version("1.1.0")
+    BUILDER_CONFIGS = [
+        datasets.BuilderConfig(name="mt_en_de", version=VERSION,
+                               description="This is the EN-DE part of the Wino-X translation data."),
+        datasets.BuilderConfig(name="mt_en_fr", version=VERSION,
+                               description="This is the EN-FR part of the Wino-X translation data."),
+        datasets.BuilderConfig(name="mt_en_ru", version=VERSION,
+                               description="This is the EN-RU part of the Wino-X translation data."),
+        datasets.BuilderConfig(name="lm_en_de", version=VERSION,
+                               description="This is the EN-DE part of the Wino-X language modeling data."),
+        datasets.BuilderConfig(name="lm_en_fr", version=VERSION,
+                               description="This is the EN-FR part of the Wino-X language modeling data."),
+        datasets.BuilderConfig(name="lm_en_ru", version=VERSION,
+                               description="This is the EN-RU part of the Wino-X language modeling data."),
+    ]
+
+    def _info(self):
+
+        # MT example:
+        # {"qID": "3UDTAB6HH8D37OQL3O6F3GXEEOF09Z-1",
+        # "sentence": "The woman looked for a different vase for the bouquet because it was too small.",
+        # "translation1": "Die Frau suchte nach einer anderen Vase für den Blumenstrauß, weil sie zu klein war.",
+        # "translation2": "Die Frau suchte nach einer anderen Vase für den Blumenstrauß, weil er zu klein war.",
+        # "answer": 1,
+        # "pronoun1": "sie",
+        # "pronoun2": "er",
+        # "referent1_en": "vase",
+        # "referent2_en": "bouquet",
+        # "true_translation_referent_of_pronoun1_de": "Vase",
+        # "true_translation_referent_of_pronoun2_de": "Blumenstrauß",
+        # "false_translation_referent_of_pronoun1_de": "Vase",
+        # "false_translation_referent_of_pronoun2_de": "Blumenstrauß"}
+
+        tgt_lang = self.config.name.split('_')[-1]
+        if self.config.name.startswith('mt_'):
+            features = datasets.Features(
+                {
+                    "qID": datasets.Value("string"),
+                    "sentence": datasets.Value("string"),
+                    "translation1": datasets.Value("string"),
+                    "translation2": datasets.Value("string"),
+                    "answer": datasets.Value("int64"),
+                    "pronoun1": datasets.Value("string"),
+                    "pronoun2": datasets.Value("string"),
+                    "referent1_en": datasets.Value("string"),
+                    "referent2_en": datasets.Value("string"),
+                    "true_translation_referent_of_pronoun1_{}".format(tgt_lang): datasets.Value("string"),
+                    "true_translation_referent_of_pronoun2_{}".format(tgt_lang): datasets.Value("string"),
+                    "false_translation_referent_of_pronoun1_{}".format(tgt_lang): datasets.Value("string"),
+                    "false_translation_referent_of_pronoun2_{}".format(tgt_lang): datasets.Value("string")
+                }
+            )
+
+        # LM example:
+        # {"qID": "3UDTAB6HH8D37OQL3O6F3GXEEOF09Z-1",
+        # "sentence": "The woman looked for a different vase for the bouquet because it was too small.",
+        # "context_en": "The woman looked for a different vase for the bouquet because _ was too small.",
+        # "context_de": "Die Frau suchte nach einer anderen Vase für den Blumenstrauß, weil _ zu klein war.",
+        # "option1_en": "the vase",
+        # "option2_en": "the bouquet",
+        # "option1_de": "die Vase",
+        # "option2_de": "der Blumenstrauß",
+        # "answer": 1,
+        # "context_referent_of_option1_de": "Vase",
+        # "context_referent_of_option2_de": "Blumenstrauß"}
+
+        else:
+            features = datasets.Features(
+                {
+                    "qID": datasets.Value("string"),
+                    "sentence": datasets.Value("string"),
+                    "context_en": datasets.Value("string"),
+                    "context_{}".format(tgt_lang): datasets.Value("string"),
+                    "answer": datasets.Value("int64"),
+                    "option1_en": datasets.Value("string"),
+                    "option2_en": datasets.Value("string"),
+                    "option1_{}".format(tgt_lang): datasets.Value("string"),
+                    "option2_{}".format(tgt_lang): datasets.Value("string"),
+                    "context_referent_of_option1_{}".format(tgt_lang): datasets.Value("string"),
+                    "context_referent_of_option2_{}".format(tgt_lang): 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,
+            # 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):
+        downloaded_files = dl_manager.download_and_extract(_URLS[self.config.name])
+        return [datasets.SplitGenerator(name=datasets.Split.TEST,
+                                        gen_kwargs={'filepath': downloaded_files, 'split': 'test'})]
+
+    # method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
+    def _generate_examples(self, filepath, split):
+        # The `key` is for legacy reasons (tfds) and is not important in itself, but must be unique for each example.
+        with open(filepath, encoding="utf-8") as f:
+            for key, row in enumerate(f):
+                data = json.loads(row)
+                yield key, data