Update files from the datasets library (from 1.0.0)

Release notes: https://github.com/huggingface/datasets/releases/tag/1.0.0

.gitattributes

@@ -0,0 +1,27 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bin.* filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zstandard filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

break_data.py

@@ -0,0 +1,264 @@
+"""TODO(break_data): Add a description here."""
+
+from __future__ import absolute_import, division, print_function
+
+import csv
+import json
+import os
+import textwrap
+
+import six
+
+import datasets
+
+
+# TODO(break): BibTeX citation
+_CITATION = """\
+@article{Wolfson2020Break,
+  title={Break It Down: A Question Understanding Benchmark},
+  author={Wolfson, Tomer and Geva, Mor and Gupta, Ankit and Gardner, Matt and Goldberg, Yoav and Deutch, Daniel and Berant, Jonathan},
+  journal={Transactions of the Association for Computational Linguistics},
+  year={2020},
+}
+"""
+
+# TODO(break):
+_DESCRIPTION = """\
+Break is a human annotated dataset of natural language questions and their Question Decomposition Meaning Representations
+(QDMRs). Break consists of 83,978 examples sampled from 10 question answering datasets over text, images and databases.
+This repository contains the Break dataset along with information on the exact data format.
+"""
+_URL = "https://github.com/allenai/Break/raw/master/break_dataset/Break-dataset.zip"
+
+
+class BreakDataConfig(datasets.BuilderConfig):
+
+    """BuilderConfig for Break"""
+
+    def __init__(self, text_features, lexicon_tokens, **kwargs):
+        """
+
+        Args:
+            text_features: `dict[string, string]`, map from the name of the feature
+        dict for each text field to the name of the column in the tsv file
+            lexicon_tokens: to define if we want to load the lexicon_tokens files or not
+            **kwargs: keyword arguments forwarded to super.
+        """
+        super(BreakDataConfig, self).__init__(version=datasets.Version("1.0.0", ""), **kwargs)
+        self.text_features = text_features
+        self.lexicon_tokens = lexicon_tokens
+
+
+class BreakData(datasets.GeneratorBasedBuilder):
+    """TODO(break_data): Short description of my dataset."""
+
+    # TODO(break_data): Set up version.
+    VERSION = datasets.Version("0.1.0")
+    BUILDER_CONFIGS = [
+        BreakDataConfig(
+            name="QDMR-high-level",
+            description=textwrap.dedent(
+                """
+             Contains questions annotated with the high-level variant of QDMR. These decomposition are exclusive to Reading
+             Comprehension tasks (Section 2). lexicon_tokens files are also provided."""
+            ),
+            text_features={
+                "question_id": "question_id",
+                "question_text": "question_text",
+                "decomposition": "decomposition",
+                "operators": "operators",
+                "split": "split",
+            },
+            lexicon_tokens=False,
+        ),
+        BreakDataConfig(
+            name="QDMR-high-level-lexicon",
+            description=textwrap.dedent(
+                """
+               Contains questions annotated with the high-level variant of QDMR. These decomposition are exclusive to Reading
+               Comprehension tasks (Section 2). lexicon_tokens files are also provided."""
+            ),
+            text_features={
+                "source": "source",
+                "allowed_tokens": "allowed_tokens",
+            },
+            lexicon_tokens=True,
+        ),
+        BreakDataConfig(
+            name="QDMR",
+            description=textwrap.dedent(
+                """
+               Contains questions over text, images and databases annotated with their Question Decomposition Meaning
+               Representation. In addition to the train, dev and (hidden) test sets we provide lexicon_tokens files. For
+               each question, the lexicon file contains the set of valid tokens that could potentially appear in its
+               decomposition """
+            ),
+            text_features={
+                "question_id": "question_id",
+                "question_text": "question_text",
+                "decomposition": "decomposition",
+                "operators": "operators",
+                "split": "split",
+            },
+            lexicon_tokens=False,
+        ),
+        BreakDataConfig(
+            name="QDMR-lexicon",
+            description=textwrap.dedent(
+                """
+                 Contains questions over text, images and databases annotated with their Question Decomposition Meaning
+               Representation. In addition to the train, dev and (hidden) test sets we provide lexicon_tokens files. For
+               each question, the lexicon file contains the set of valid tokens that could potentially appear in its
+               decomposition """
+            ),
+            text_features={
+                "source": "source",
+                "allowed_tokens": "allowed_tokens",
+            },
+            lexicon_tokens=True,
+        ),
+        BreakDataConfig(
+            name="logical-forms",
+            description=textwrap.dedent(
+                """
+               Contains questions and QDMRs annotated with full logical-forms of QDMR operators + arguments. Full logical-forms
+               were inferred by the annotation-consistency algorithm described in """
+            ),
+            lexicon_tokens=False,
+            text_features={
+                "question_id": "question_id",
+                "question_text": "question_text",
+                "decomposition": "decomposition",
+                "operators": "operators",
+                "split": "split",
+                "program": "program",
+            },
+        ),
+    ]
+
+    def _info(self):
+        # TODO(break_data): Specifies the datasets.DatasetInfo object
+        features = {text_feature: datasets.Value("string") for text_feature in six.iterkeys(self.config.text_features)}
+        return datasets.DatasetInfo(
+            # This is the description that will appear on the datasets page.
+            description=_DESCRIPTION,
+            # datasets.features.FeatureConnectors
+            features=datasets.Features(
+                features
+                # 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://github.com/allenai/Break",
+            citation=_CITATION,
+        )
+        # if
+
+    def _split_generators(self, dl_manager):
+        """Returns SplitGenerators."""
+        # TODO(break_data): Downloads the data and defines the splits
+        # dl_manager is a datasets.download.DownloadManager that can be used to
+        # download and extract URLs
+        dl_dir = dl_manager.download_and_extract(_URL)
+        data_dir = os.path.join(dl_dir, "Break-dataset")
+        qdmr_high_level = os.path.join(data_dir, "QDMR-high-level")
+        qdmr = os.path.join(data_dir, "QDMR")
+        logical = os.path.join(data_dir, "logical-forms")
+        if self.config.name == "QDMR" or self.config.name == "QDMR-lexicon":
+            return [
+                datasets.SplitGenerator(
+                    name=datasets.Split.TRAIN,
+                    # These kwargs will be passed to _generate_examples
+                    gen_kwargs={
+                        "filepath": os.path.join(qdmr, "train.csv")
+                        if not self.config.lexicon_tokens
+                        else os.path.join(qdmr, "train_lexicon_tokens.json")
+                    },
+                ),
+                datasets.SplitGenerator(
+                    name=datasets.Split.VALIDATION,
+                    # These kwargs will be passed to _generate_examples
+                    gen_kwargs={
+                        "filepath": os.path.join(qdmr, "dev.csv")
+                        if not self.config.lexicon_tokens
+                        else os.path.join(qdmr, "dev_lexicon_tokens.json")
+                    },
+                ),
+                datasets.SplitGenerator(
+                    name=datasets.Split.TEST,
+                    # These kwargs will be passed to _generate_examples
+                    gen_kwargs={
+                        "filepath": os.path.join(qdmr, "test.csv")
+                        if not self.config.lexicon_tokens
+                        else os.path.join(qdmr, "test_lexicon_tokens.json")
+                    },
+                ),
+            ]
+        elif self.config.name == "QDMR-high-level" or self.config.name == "QDMR-high-level-lexicon":
+            return [
+                datasets.SplitGenerator(
+                    name=datasets.Split.TRAIN,
+                    # These kwargs will be passed to _generate_examples
+                    gen_kwargs={
+                        "filepath": os.path.join(qdmr_high_level, "train.csv")
+                        if not self.config.lexicon_tokens
+                        else os.path.join(qdmr_high_level, "train_lexicon_tokens.json")
+                    },
+                ),
+                datasets.SplitGenerator(
+                    name=datasets.Split.VALIDATION,
+                    # These kwargs will be passed to _generate_examples
+                    gen_kwargs={
+                        "filepath": os.path.join(qdmr_high_level, "dev.csv")
+                        if not self.config.lexicon_tokens
+                        else os.path.join(qdmr_high_level, "dev_lexicon_tokens.json")
+                    },
+                ),
+                datasets.SplitGenerator(
+                    name=datasets.Split.TEST,
+                    # These kwargs will be passed to _generate_examples
+                    gen_kwargs={
+                        "filepath": os.path.join(qdmr_high_level, "test.csv")
+                        if not self.config.lexicon_tokens
+                        else os.path.join(qdmr_high_level, "test_lexicon_tokens.json")
+                    },
+                ),
+            ]
+        elif self.config.name == "logical-forms":
+            return [
+                datasets.SplitGenerator(
+                    name=datasets.Split.TRAIN,
+                    # These kwargs will be passed to _generate_examples
+                    gen_kwargs={"filepath": os.path.join(logical, "train.csv")},
+                ),
+                datasets.SplitGenerator(
+                    name=datasets.Split.VALIDATION,
+                    # These kwargs will be passed to _generate_examples
+                    gen_kwargs={"filepath": os.path.join(logical, "dev.csv")},
+                ),
+                datasets.SplitGenerator(
+                    name=datasets.Split.TEST,
+                    # These kwargs will be passed to _generate_examples
+                    gen_kwargs={"filepath": os.path.join(logical, "test.csv")},
+                ),
+            ]
+
+    def _generate_examples(self, filepath):
+        """Yields examples."""
+        # TODO(break_data): Yields (key, example) tuples from the dataset
+        with open(filepath, encoding="utf-8") as f:
+            if (
+                self.config.name == "QDMR-high-level"
+                or self.config.name == "QDMR"
+                or self.config.name == "logical-forms"
+            ):
+                data = csv.DictReader(f)
+                for id_, row in enumerate(data):
+                    yield id_, row
+            elif self.config.name == "QDMR-high-level-lexicon" or self.config.name == "QDMR-lexicon":
+                for id_, row in enumerate(f):
+                    data = json.loads(row)
+                    yield id_, data

dataset_infos.json

@@ -0,0 +1 @@
+{"QDMR-high-level": {"description": "Break is a human annotated dataset of natural language questions and their Question Decomposition Meaning Representations\n(QDMRs). Break consists of 83,978 examples sampled from 10 question answering datasets over text, images and databases. \nThis repository contains the Break dataset along with information on the exact data format.\n", "citation": "@article{Wolfson2020Break,\n  title={Break It Down: A Question Understanding Benchmark},\n  author={Wolfson, Tomer and Geva, Mor and Gupta, Ankit and Gardner, Matt and Goldberg, Yoav and Deutch, Daniel and Berant, Jonathan},\n  journal={Transactions of the Association for Computational Linguistics},\n  year={2020},\n}\n", "homepage": "https://github.com/allenai/Break", "license": "", "features": {"question_id": {"dtype": "string", "id": null, "_type": "Value"}, "question_text": {"dtype": "string", "id": null, "_type": "Value"}, "decomposition": {"dtype": "string", "id": null, "_type": "Value"}, "operators": {"dtype": "string", "id": null, "_type": "Value"}, "split": {"dtype": "string", "id": null, "_type": "Value"}}, "supervised_keys": null, "builder_name": "break_data", "config_name": "QDMR-high-level", "version": {"version_str": "1.0.0", "description": "", "datasets_version_to_prepare": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 482339, "num_examples": 3195, "dataset_name": "break_data"}, "train": {"name": "train", "num_bytes": 5148086, "num_examples": 17503, "dataset_name": "break_data"}, "validation": {"name": "validation", "num_bytes": 914780, "num_examples": 3130, "dataset_name": "break_data"}}, "download_checksums": {"https://github.com/allenai/Break/raw/master/break_dataset/Break-dataset.zip": {"num_bytes": 15971078, "checksum": "37efea4fa1b7774d077ff0452e5e199cecba8216c12da76781010f189d1cf259"}}, "download_size": 15971078, "dataset_size": 6545205, "size_in_bytes": 22516283}, "QDMR-high-level-lexicon": {"description": "Break is a human annotated dataset of natural language questions and their Question Decomposition Meaning Representations\n(QDMRs). Break consists of 83,978 examples sampled from 10 question answering datasets over text, images and databases. \nThis repository contains the Break dataset along with information on the exact data format.\n", "citation": "@article{Wolfson2020Break,\n  title={Break It Down: A Question Understanding Benchmark},\n  author={Wolfson, Tomer and Geva, Mor and Gupta, Ankit and Gardner, Matt and Goldberg, Yoav and Deutch, Daniel and Berant, Jonathan},\n  journal={Transactions of the Association for Computational Linguistics},\n  year={2020},\n}\n", "homepage": "https://github.com/allenai/Break", "license": "", "features": {"source": {"dtype": "string", "id": null, "_type": "Value"}, "allowed_tokens": {"dtype": "string", "id": null, "_type": "Value"}}, "supervised_keys": null, "builder_name": "break_data", "config_name": "QDMR-high-level-lexicon", "version": {"version_str": "1.0.0", "description": "", "datasets_version_to_prepare": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 4240755, "num_examples": 3195, "dataset_name": "break_data"}, "train": {"name": "train", "num_bytes": 23234518, "num_examples": 17503, "dataset_name": "break_data"}, "validation": {"name": "validation", "num_bytes": 4158679, "num_examples": 3130, "dataset_name": "break_data"}}, "download_checksums": {"https://github.com/allenai/Break/raw/master/break_dataset/Break-dataset.zip": {"num_bytes": 15971078, "checksum": "37efea4fa1b7774d077ff0452e5e199cecba8216c12da76781010f189d1cf259"}}, "download_size": 15971078, "dataset_size": 31633952, "size_in_bytes": 47605030}, "QDMR": {"description": "Break is a human annotated dataset of natural language questions and their Question Decomposition Meaning Representations\n(QDMRs). Break consists of 83,978 examples sampled from 10 question answering datasets over text, images and databases. \nThis repository contains the Break dataset along with information on the exact data format.\n", "citation": "@article{Wolfson2020Break,\n  title={Break It Down: A Question Understanding Benchmark},\n  author={Wolfson, Tomer and Geva, Mor and Gupta, Ankit and Gardner, Matt and Goldberg, Yoav and Deutch, Daniel and Berant, Jonathan},\n  journal={Transactions of the Association for Computational Linguistics},\n  year={2020},\n}\n", "homepage": "https://github.com/allenai/Break", "license": "", "features": {"question_id": {"dtype": "string", "id": null, "_type": "Value"}, "question_text": {"dtype": "string", "id": null, "_type": "Value"}, "decomposition": {"dtype": "string", "id": null, "_type": "Value"}, "operators": {"dtype": "string", "id": null, "_type": "Value"}, "split": {"dtype": "string", "id": null, "_type": "Value"}}, "supervised_keys": null, "builder_name": "break_data", "config_name": "QDMR", "version": {"version_str": "1.0.0", "description": "", "datasets_version_to_prepare": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 900632, "num_examples": 8069, "dataset_name": "break_data"}, "train": {"name": "train", "num_bytes": 12790466, "num_examples": 44321, "dataset_name": "break_data"}, "validation": {"name": "validation", "num_bytes": 2237472, "num_examples": 7760, "dataset_name": "break_data"}}, "download_checksums": {"https://github.com/allenai/Break/raw/master/break_dataset/Break-dataset.zip": {"num_bytes": 15971078, "checksum": "37efea4fa1b7774d077ff0452e5e199cecba8216c12da76781010f189d1cf259"}}, "download_size": 15971078, "dataset_size": 15928570, "size_in_bytes": 31899648}, "QDMR-lexicon": {"description": "Break is a human annotated dataset of natural language questions and their Question Decomposition Meaning Representations\n(QDMRs). Break consists of 83,978 examples sampled from 10 question answering datasets over text, images and databases. \nThis repository contains the Break dataset along with information on the exact data format.\n", "citation": "@article{Wolfson2020Break,\n  title={Break It Down: A Question Understanding Benchmark},\n  author={Wolfson, Tomer and Geva, Mor and Gupta, Ankit and Gardner, Matt and Goldberg, Yoav and Deutch, Daniel and Berant, Jonathan},\n  journal={Transactions of the Association for Computational Linguistics},\n  year={2020},\n}\n", "homepage": "https://github.com/allenai/Break", "license": "", "features": {"source": {"dtype": "string", "id": null, "_type": "Value"}, "allowed_tokens": {"dtype": "string", "id": null, "_type": "Value"}}, "supervised_keys": null, "builder_name": "break_data", "config_name": "QDMR-lexicon", "version": {"version_str": "1.0.0", "description": "", "datasets_version_to_prepare": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 10331822, "num_examples": 8069, "dataset_name": "break_data"}, "train": {"name": "train", "num_bytes": 56913064, "num_examples": 44321, "dataset_name": "break_data"}, "validation": {"name": "validation", "num_bytes": 9936933, "num_examples": 7760, "dataset_name": "break_data"}}, "download_checksums": {"https://github.com/allenai/Break/raw/master/break_dataset/Break-dataset.zip": {"num_bytes": 15971078, "checksum": "37efea4fa1b7774d077ff0452e5e199cecba8216c12da76781010f189d1cf259"}}, "download_size": 15971078, "dataset_size": 77181819, "size_in_bytes": 93152897}, "logical-forms": {"description": "Break is a human annotated dataset of natural language questions and their Question Decomposition Meaning Representations\n(QDMRs). Break consists of 83,978 examples sampled from 10 question answering datasets over text, images and databases. \nThis repository contains the Break dataset along with information on the exact data format.\n", "citation": "@article{Wolfson2020Break,\n  title={Break It Down: A Question Understanding Benchmark},\n  author={Wolfson, Tomer and Geva, Mor and Gupta, Ankit and Gardner, Matt and Goldberg, Yoav and Deutch, Daniel and Berant, Jonathan},\n  journal={Transactions of the Association for Computational Linguistics},\n  year={2020},\n}\n", "homepage": "https://github.com/allenai/Break", "license": "", "features": {"question_id": {"dtype": "string", "id": null, "_type": "Value"}, "question_text": {"dtype": "string", "id": null, "_type": "Value"}, "decomposition": {"dtype": "string", "id": null, "_type": "Value"}, "operators": {"dtype": "string", "id": null, "_type": "Value"}, "split": {"dtype": "string", "id": null, "_type": "Value"}, "program": {"dtype": "string", "id": null, "_type": "Value"}}, "supervised_keys": null, "builder_name": "break_data", "config_name": "logical-forms", "version": {"version_str": "1.0.0", "description": "", "datasets_version_to_prepare": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 927038, "num_examples": 8006, "dataset_name": "break_data"}, "train": {"name": "train", "num_bytes": 19821676, "num_examples": 44098, "dataset_name": "break_data"}, "validation": {"name": "validation", "num_bytes": 3504893, "num_examples": 7719, "dataset_name": "break_data"}}, "download_checksums": {"https://github.com/allenai/Break/raw/master/break_dataset/Break-dataset.zip": {"num_bytes": 15971078, "checksum": "37efea4fa1b7774d077ff0452e5e199cecba8216c12da76781010f189d1cf259"}}, "download_size": 15971078, "dataset_size": 24253607, "size_in_bytes": 40224685}}

dummy/QDMR-high-level-lexicon/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e76ecbc0b26764bb8fd3672fc1a9b1b29b5f1f2288b5fdc8944b7c6b6d01af93
+size 3333

dummy/QDMR-high-level/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:292bd95a179b813dcc77d8ca8a3ed47b18d927f8e0a570aabb53c98416bdd5f6
+size 1873

dummy/QDMR-lexicon/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f1977230ec1d2511e843b24e97d920118a1d8df1166265a9534a014422fedfa1
+size 3245

dummy/QDMR/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b2c808130f43b920356afbf6871d306ce5c8d96adce9f09e2fc8446b767b8751
+size 4492

dummy/logical-forms/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21c0e6ffcf822bbabf13335a2bb267ce7feca3bd8e7a3de9e8eab85c2e88893f
+size 1884