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

dataset_infos.json

@@ -0,0 +1 @@
+{"mode=first,char_skip=25": {"description": "\nThe Qanta dataset is a question answering dataset based on the academic trivia game Quizbowl.\n", "citation": "\n@article{Rodriguez2019QuizbowlTC,\n  title={Quizbowl: The Case for Incremental Question Answering},\n  author={Pedro Rodriguez and Shi Feng and Mohit Iyyer and He He and Jordan L. Boyd-Graber},\n  journal={ArXiv},\n  year={2019},\n  volume={abs/1904.04792}\n}\n", "homepage": "http://www.qanta.org/", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "qanta_id": {"dtype": "int32", "id": null, "_type": "Value"}, "proto_id": {"dtype": "string", "id": null, "_type": "Value"}, "qdb_id": {"dtype": "int32", "id": null, "_type": "Value"}, "dataset": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}, "full_question": {"dtype": "string", "id": null, "_type": "Value"}, "first_sentence": {"dtype": "string", "id": null, "_type": "Value"}, "char_idx": {"dtype": "int32", "id": null, "_type": "Value"}, "sentence_idx": {"dtype": "int32", "id": null, "_type": "Value"}, "tokenizations": {"feature": {"feature": {"dtype": "int32", "id": null, "_type": "Value"}, "length": 2, "id": null, "_type": "Sequence"}, "length": -1, "id": null, "_type": "Sequence"}, "answer": {"dtype": "string", "id": null, "_type": "Value"}, "page": {"dtype": "string", "id": null, "_type": "Value"}, "raw_answer": {"dtype": "string", "id": null, "_type": "Value"}, "fold": {"dtype": "string", "id": null, "_type": "Value"}, "gameplay": {"dtype": "bool", "id": null, "_type": "Value"}, "category": {"dtype": "string", "id": null, "_type": "Value"}, "subcategory": {"dtype": "string", "id": null, "_type": "Value"}, "tournament": {"dtype": "string", "id": null, "_type": "Value"}, "difficulty": {"dtype": "string", "id": null, "_type": "Value"}, "year": {"dtype": "int32", "id": null, "_type": "Value"}}, "supervised_keys": null, "builder_name": "qanta", "config_name": "mode=first,char_skip=25", "version": {"version_str": "2018.04.18", "description": null, "datasets_version_to_prepare": null, "major": 2018, "minor": 4, "patch": 18}, "splits": {"adversarial": {"name": "adversarial", "num_bytes": 1258844, "num_examples": 1145, "dataset_name": "qanta"}, "buzzdev": {"name": "buzzdev", "num_bytes": 1553636, "num_examples": 1161, "dataset_name": "qanta"}, "buzztest": {"name": "buzztest", "num_bytes": 2653425, "num_examples": 1953, "dataset_name": "qanta"}, "buzztrain": {"name": "buzztrain", "num_bytes": 19699736, "num_examples": 16706, "dataset_name": "qanta"}, "guessdev": {"name": "guessdev", "num_bytes": 1414882, "num_examples": 1055, "dataset_name": "qanta"}, "guesstest": {"name": "guesstest", "num_bytes": 2997123, "num_examples": 2151, "dataset_name": "qanta"}, "guesstrain": {"name": "guesstrain", "num_bytes": 117599750, "num_examples": 96221, "dataset_name": "qanta"}}, "download_checksums": {"https://s3-us-west-2.amazonaws.com/pinafore-us-west-2/qanta-jmlr-datasets/qanta.mapped.2018.04.18.json": {"num_bytes": 169001564, "checksum": "5f2f429724e13df1d4b216dba5549dac597fbaf884ed0c3e01f90ee72cb2753a"}, "https://s3-us-west-2.amazonaws.com/pinafore-us-west-2/trick-tacl-datasets/qanta.tacl-trick.json": {"num_bytes": 1753354, "checksum": "73535d6493d63bad48cd61031911faf77efb584976475185d8aca5a124d1822b"}}, "download_size": 170754918, "dataset_size": 147177396, "size_in_bytes": 317932314}}

dummy/mode=first,char_skip=25/2018.4.18/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb32bbb897f87b3e6541f97990d5418e6efaca1594102d05b917ef45187e6629
+size 2117

dummy/mode=full,char_skip=25/2018.4.18/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb32bbb897f87b3e6541f97990d5418e6efaca1594102d05b917ef45187e6629
+size 2117

dummy/mode=runs,char_skip=25/2018.4.18/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb32bbb897f87b3e6541f97990d5418e6efaca1594102d05b917ef45187e6629
+size 2117

dummy/mode=sentences,char_skip=25/2018.4.18/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb32bbb897f87b3e6541f97990d5418e6efaca1594102d05b917ef45187e6629
+size 2117

qanta.py

@@ -0,0 +1,290 @@
+"""qanta dataset."""
+
+from __future__ import absolute_import, division, print_function
+
+import json
+from typing import List, Tuple
+
+import datasets
+
+
+_CITATION = """
+@article{Rodriguez2019QuizbowlTC,
+  title={Quizbowl: The Case for Incremental Question Answering},
+  author={Pedro Rodriguez and Shi Feng and Mohit Iyyer and He He and Jordan L. Boyd-Graber},
+  journal={ArXiv},
+  year={2019},
+  volume={abs/1904.04792}
+}
+"""
+
+_DESCRIPTION = """
+The Qanta dataset is a question answering dataset based on the academic trivia game Quizbowl.
+"""
+
+
+_QANTA_URL = "https://s3-us-west-2.amazonaws.com/pinafore-us-west-2/qanta-jmlr-datasets/qanta.mapped.2018.04.18.json"
+_TRICK_URL = "https://s3-us-west-2.amazonaws.com/pinafore-us-west-2/trick-tacl-datasets/qanta.tacl-trick.json"
+_VERSION = datasets.Version("2018.04.18")
+_FIRST = "first"
+_FULL = "full"
+_SENTENCES = "sentences"
+_RUNS = "runs"
+# Order matters, the first one is default
+_MODES = [_FULL, _FIRST, _SENTENCES, _RUNS]
+_DEFAULT_CHAR_SKIP = 25
+
+
+class QantaConfig(datasets.BuilderConfig):
+    """BuilderConfig for Qanta."""
+
+    def __init__(self, mode: str, char_skip: int, **kwargs):
+        super(QantaConfig, self).__init__(version=_VERSION, **kwargs)
+        self.mode = mode
+        self.char_skip = char_skip
+
+
+def create_char_runs(text: str, char_skip: int) -> List[Tuple[str, int]]:
+    """
+    Returns runs of the question based on skipping char_skip characters at a time. Also returns the indices used
+    q: name this first united states president.
+    runs with char_skip=10:
+    ['name this ',
+        'name this first unit',
+        'name this first united state p',
+        'name this first united state president.']
+    :param char_skip: Number of characters to skip each time
+    """
+    char_indices = list(range(char_skip, len(text) + char_skip, char_skip))
+    return [(text[:idx], idx) for idx in char_indices]
+
+
+def with_default(key, lookup, default):
+    if key in lookup:
+        value = lookup[key]
+        if value is None:
+            return default
+        else:
+            return value
+    else:
+        return default
+
+
+def question_to_examples(question, mode: str, char_skip: int):
+    features = {
+        "qanta_id": question["qanta_id"],
+        "proto_id": with_default("proto_id", question, ""),
+        "qdb_id": with_default("qdb_id", question, -1),
+        # We refer to the actual answer as page, but this
+        # may be misleading externally, so rename here to
+        # be clearer
+        "page": question["page"],
+        "answer": question["page"],
+        "raw_answer": question["answer"],
+        "dataset": with_default("dataset", question, ""),
+        "full_question": question["text"],
+        "first_sentence": question["first_sentence"],
+        "tokenizations": question["tokenizations"],
+        "fold": question["fold"],
+        "gameplay": question["gameplay"],
+        "category": with_default("category", question, ""),
+        "subcategory": with_default("subcategory", question, ""),
+        "tournament": question["tournament"],
+        "difficulty": with_default("difficulty", question, ""),
+        "year": question["year"],
+        "char_idx": -1,
+        "sentence_idx": -1,
+    }
+    if mode == _FULL:
+        yield {
+            "text": question["text"],
+            "id": str(question["qanta_id"]) + "-full",
+            **features,
+        }
+    elif mode == _FIRST:
+        yield {
+            "text": question["first_sentence"],
+            "id": str(question["qanta_id"]) + "-first",
+            **features,
+        }
+    elif mode == _RUNS:
+        text = question["text"]
+        for text_run, char_idx in create_char_runs(text, char_skip):
+            yield {
+                "text": text_run,
+                "char_idx": char_idx,
+                "id": str(question["qanta_id"]) + "-char-" + str(char_idx),
+                **features,
+            }
+    elif mode == _SENTENCES:
+        for sentence_idx, (start, end) in enumerate(question["tokenizations"]):
+            sentence = question["text"][start:end]
+            yield {
+                "text": sentence,
+                "sentence_idx": sentence_idx,
+                "id": str(question["qanta_id"]) + "-sentence-" + str(sentence_idx),
+                **features,
+            }
+    else:
+        raise ValueError(f"Invalid mode: {mode}")
+
+
+_FEATURES = {
+    # Generated ID based modes set, unique
+    "id": datasets.Value("string"),
+    # Dataset defined IDs
+    "qanta_id": datasets.Value("int32"),
+    "proto_id": datasets.Value("string"),
+    "qdb_id": datasets.Value("int32"),
+    "dataset": datasets.Value("string"),
+    # Inputs
+    "text": datasets.Value("string"),
+    "full_question": datasets.Value("string"),
+    "first_sentence": datasets.Value("string"),
+    "char_idx": datasets.Value("int32"),
+    "sentence_idx": datasets.Value("int32"),
+    # Character indices of sentences: List[Tuple[int, int]]
+    "tokenizations": datasets.features.Sequence(datasets.features.Sequence(datasets.Value("int32"), length=2)),
+    # Labels: Number is equal to number of unique pages across all folds
+    "answer": datasets.Value("string"),
+    "page": datasets.Value("string"),
+    "raw_answer": datasets.Value("string"),
+    # Meta Information
+    "fold": datasets.Value("string"),
+    "gameplay": datasets.Value("bool"),
+    "category": datasets.Value("string"),
+    "subcategory": datasets.Value("string"),
+    "tournament": datasets.Value("string"),
+    "difficulty": datasets.Value("string"),
+    "year": datasets.Value("int32"),
+}
+
+
+class Qanta(datasets.GeneratorBasedBuilder):
+    """The Qanta dataset is a question answering dataset based on the academic trivia game Quizbowl."""
+
+    VERSION = _VERSION
+    BUILDER_CONFIGS = [
+        QantaConfig(
+            name=f"mode={mode},char_skip={_DEFAULT_CHAR_SKIP}",
+            description=f"Question format: {mode}, char_skip: {_DEFAULT_CHAR_SKIP}",
+            mode=mode,
+            char_skip=_DEFAULT_CHAR_SKIP,
+        )
+        for mode in _MODES
+    ]
+
+    def _info(self):
+        return datasets.DatasetInfo(
+            # This is the description that will appear on the datasets page.
+            description=_DESCRIPTION,
+            # datasets.features.FeatureConnectors
+            features=datasets.Features(_FEATURES),
+            # Number of classes is a function of the dataset, ClassLabel doesn't support dynamic
+            # definition, so have to defer conversion to classes to later, so can't define
+            # supervied keys
+            supervised_keys=None,
+            # Homepage of the dataset for documentation
+            homepage="http://www.qanta.org/",
+            citation=_CITATION,
+        )
+
+    def _split_generators(self, dl_manager):
+        """Returns SplitGenerators."""
+        qanta_path = dl_manager.download_and_extract(_QANTA_URL)
+        trick_path = dl_manager.download_and_extract(_TRICK_URL)
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split("guesstrain"),
+                gen_kwargs={
+                    "qanta_filepath": qanta_path,
+                    "trick_filepath": trick_path,
+                    "fold": "guesstrain",
+                    "mode": self.config.mode,
+                    "char_skip": self.config.char_skip,
+                },
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split("buzztrain"),
+                gen_kwargs={
+                    "qanta_filepath": qanta_path,
+                    "trick_filepath": trick_path,
+                    "fold": "buzztrain",
+                    "mode": self.config.mode,
+                    "char_skip": self.config.char_skip,
+                },
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split("guessdev"),
+                gen_kwargs={
+                    "qanta_filepath": qanta_path,
+                    "trick_filepath": trick_path,
+                    "fold": "guessdev",
+                    "mode": self.config.mode,
+                    "char_skip": self.config.char_skip,
+                },
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split("buzzdev"),
+                gen_kwargs={
+                    "qanta_filepath": qanta_path,
+                    "trick_filepath": trick_path,
+                    "fold": "buzzdev",
+                    "mode": self.config.mode,
+                    "char_skip": self.config.char_skip,
+                },
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split("guesstest"),
+                gen_kwargs={
+                    "qanta_filepath": qanta_path,
+                    "trick_filepath": trick_path,
+                    "fold": "guesstest",
+                    "mode": self.config.mode,
+                    "char_skip": self.config.char_skip,
+                },
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split("buzztest"),
+                gen_kwargs={
+                    "qanta_filepath": qanta_path,
+                    "trick_filepath": trick_path,
+                    "fold": "buzztest",
+                    "mode": self.config.mode,
+                    "char_skip": self.config.char_skip,
+                },
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split("adversarial"),
+                gen_kwargs={
+                    "qanta_filepath": qanta_path,
+                    "trick_filepath": trick_path,
+                    "fold": "adversarial",
+                    "mode": self.config.mode,
+                    "char_skip": self.config.char_skip,
+                },
+            ),
+        ]
+
+    def _generate_examples(
+        self,
+        qanta_filepath: str,
+        trick_filepath: str,
+        fold: str,
+        mode: str,
+        char_skip: int,
+    ):
+        """Yields examples."""
+        if mode not in _MODES:
+            raise ValueError(f"Invalid mode: {mode}")
+
+        if fold == "adversarial":
+            path = trick_filepath
+        else:
+            path = qanta_filepath
+        with open(path, encoding="utf-8") as f:
+            questions = json.load(f)["questions"]
+            for q in questions:
+                if q["page"] is not None and q["fold"] == fold:
+                    for example in question_to_examples(q, mode, char_skip):
+                        yield example["id"], example