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 @@
+{"senate": {"description": "\nThis release contains 1.3 million pairs of aligned text chunks (sentences or smaller fragments)\nfrom the official records (Hansards) of the 36th Canadian Parliament.\n\nThe complete Hansards of the debates in the House and Senate of the 36th Canadian Parliament,\nas far as available, were aligned. The corpus was then split into 5 sets of sentence pairs:\ntraining (80% of the sentence pairs), two sets of sentence pairs for testing (5% each), and\ntwo sets of sentence pairs for final evaluation (5% each). The current release consists of the\ntraining and testing sets. The evaluation sets are reserved for future MT evaluation purposes\nand currently not available.\n\nCaveats\n1. This release contains only sentence pairs. Even though the order of the sentences is the same\nas in the original, there may be gaps resulting from many-to-one, many-to-many, or one-to-many\nalignments that were filtered out. Therefore, this release may not be suitable for\ndiscourse-related research. \n2. Neither the sentence splitting nor the alignments are perfect. In particular, watch out for\npairs that differ considerably in length. You may want to filter these out before you do\nany statistical training.\n\nThe alignment of the Hansards was performed as part of the ReWrite project under funding\nfrom the DARPA TIDES program.\n", "citation": "\n", "homepage": "https://www.isi.edu/natural-language/download/hansard/", "license": "", "features": {"fr": {"dtype": "string", "id": null, "_type": "Value"}, "en": {"dtype": "string", "id": null, "_type": "Value"}}, "supervised_keys": null, "builder_name": "hansards", "config_name": "senate", "version": {"version_str": "1.0.0", "description": "", "datasets_version_to_prepare": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 5711686, "num_examples": 25553, "dataset_name": "hansards"}, "train": {"name": "train", "num_bytes": 40324278, "num_examples": 182135, "dataset_name": "hansards"}}, "download_checksums": {"http://www.isi.edu/natural-language/download/hansard/hansard.36.r2001-1a.senate.debates.training.tar": {"num_bytes": 13363200, "checksum": "f6e0b790133142244c3dffc8b72580bdee93958411c54b80e60febd53a22db53"}, "http://www.isi.edu/natural-language/download/hansard/hansard.36.r2001-1a.senate.debates.testing.tar": {"num_bytes": 1884160, "checksum": "86e76d003ee42458cb357b9227b147f63695195c69e89f43dcd84d7f66c819d2"}}, "download_size": 15247360, "dataset_size": 46035964, "size_in_bytes": 61283324}, "house": {"description": "\nThis release contains 1.3 million pairs of aligned text chunks (sentences or smaller fragments)\nfrom the official records (Hansards) of the 36th Canadian Parliament.\n\nThe complete Hansards of the debates in the House and Senate of the 36th Canadian Parliament,\nas far as available, were aligned. The corpus was then split into 5 sets of sentence pairs:\ntraining (80% of the sentence pairs), two sets of sentence pairs for testing (5% each), and\ntwo sets of sentence pairs for final evaluation (5% each). The current release consists of the\ntraining and testing sets. The evaluation sets are reserved for future MT evaluation purposes\nand currently not available.\n\nCaveats\n1. This release contains only sentence pairs. Even though the order of the sentences is the same\nas in the original, there may be gaps resulting from many-to-one, many-to-many, or one-to-many\nalignments that were filtered out. Therefore, this release may not be suitable for\ndiscourse-related research. \n2. Neither the sentence splitting nor the alignments are perfect. In particular, watch out for\npairs that differ considerably in length. You may want to filter these out before you do\nany statistical training.\n\nThe alignment of the Hansards was performed as part of the ReWrite project under funding\nfrom the DARPA TIDES program.\n", "citation": "\n", "homepage": "https://www.isi.edu/natural-language/download/hansard/", "license": "", "features": {"fr": {"dtype": "string", "id": null, "_type": "Value"}, "en": {"dtype": "string", "id": null, "_type": "Value"}}, "supervised_keys": null, "builder_name": "hansards", "config_name": "house", "version": {"version_str": "1.0.0", "description": "", "datasets_version_to_prepare": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 22906629, "num_examples": 122290, "dataset_name": "hansards"}, "train": {"name": "train", "num_bytes": 191459584, "num_examples": 947969, "dataset_name": "hansards"}}, "download_checksums": {"http://www.isi.edu/natural-language/download/hansard/hansard.36.r2001-1a.house.debates.training.tar": {"num_bytes": 60569600, "checksum": "8ea77a34538c7a5c942a44964dad5a56cfbac5605c0517196b7aece353ee3992"}, "http://www.isi.edu/natural-language/download/hansard/hansard.36.r2001-1a.house.debates.testing.tar": {"num_bytes": 7014400, "checksum": "c33692dfac3c727a2fb0103487638ea7f2df16ff01b856e7b8ebff68dcc5904a"}}, "download_size": 67584000, "dataset_size": 214366213, "size_in_bytes": 281950213}}

dummy/house/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbc65e8086a55d5564f8c58340095ce485fa0a77a67a94be30ea2e18922bcdee
+size 7941

dummy/house/1.0.0/dummy_data/hansard.36.r2001-1a.house.debates.testing.tar/hansard.36/Release-2001.1a/sentence-pairs/house/debates/development/testing/hansard.36.1.house.debates.001.e.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e91ae75683715d190479a076e7e85370c2a841e9671c4d945b9e46753b4d3094
+size 320

dummy/house/1.0.0/dummy_data/hansard.36.r2001-1a.house.debates.testing.tar/hansard.36/Release-2001.1a/sentence-pairs/house/debates/development/testing/hansard.36.1.house.debates.001.f.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d08307459135f82baf800821a093fe10faa298fd36bf45e45dbdb5ab17a2dd2
+size 368

dummy/house/1.0.0/dummy_data/hansard.36.r2001-1a.house.debates.training.tar/hansard.36/Release-2001.1a/sentence-pairs/house/debates/development/training/hansard.36.1.house.debates.002.e.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c64e6c74e3956aceb9c625cfe4f866101fbf558d07dfb6141862ff2aeb7bfabe
+size 129

dummy/house/1.0.0/dummy_data/hansard.36.r2001-1a.house.debates.training.tar/hansard.36/Release-2001.1a/sentence-pairs/house/debates/development/training/hansard.36.1.house.debates.002.f.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eeaf64b79bae82651cdfc115430dc54c5d3b4b4d708585a55ac73130f4e9cca1
+size 136

dummy/senate/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13fbe9da79f4772f2442e3e94587284cca293ebb6d3fa8959284f5376c29c38c
+size 8005

dummy/senate/1.0.0/dummy_data/hansard.36.r2001-1a.senate.debates.testing.tar/hansard.36/Release-2001.1a/sentence-pairs/senate/debates/development/testing/hansard.36.1.house.debates.001.e.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e91ae75683715d190479a076e7e85370c2a841e9671c4d945b9e46753b4d3094
+size 320

dummy/senate/1.0.0/dummy_data/hansard.36.r2001-1a.senate.debates.testing.tar/hansard.36/Release-2001.1a/sentence-pairs/senate/debates/development/testing/hansard.36.1.house.debates.001.f.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d08307459135f82baf800821a093fe10faa298fd36bf45e45dbdb5ab17a2dd2
+size 368

dummy/senate/1.0.0/dummy_data/hansard.36.r2001-1a.senate.debates.training.tar/hansard.36/Release-2001.1a/sentence-pairs/senate/debates/development/training/hansard.36.1.house.debates.002.e.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c64e6c74e3956aceb9c625cfe4f866101fbf558d07dfb6141862ff2aeb7bfabe
+size 129

dummy/senate/1.0.0/dummy_data/hansard.36.r2001-1a.senate.debates.training.tar/hansard.36/Release-2001.1a/sentence-pairs/senate/debates/development/training/hansard.36.1.house.debates.002.f.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eeaf64b79bae82651cdfc115430dc54c5d3b4b4d708585a55ac73130f4e9cca1
+size 136

hansards.py

@@ -0,0 +1,157 @@
+"""TODO(hansards): Add a description here."""
+
+from __future__ import absolute_import, division, print_function
+
+import glob
+import os
+
+import datasets
+
+
+# TODO(hansards): BibTeX citation
+_CITATION = """
+"""
+
+# TODO(hansards):
+_DESCRIPTION = """
+This release contains 1.3 million pairs of aligned text chunks (sentences or smaller fragments)
+from the official records (Hansards) of the 36th Canadian Parliament.
+
+The complete Hansards of the debates in the House and Senate of the 36th Canadian Parliament,
+as far as available, were aligned. The corpus was then split into 5 sets of sentence pairs:
+training (80% of the sentence pairs), two sets of sentence pairs for testing (5% each), and
+two sets of sentence pairs for final evaluation (5% each). The current release consists of the
+training and testing sets. The evaluation sets are reserved for future MT evaluation purposes
+and currently not available.
+
+Caveats
+1. This release contains only sentence pairs. Even though the order of the sentences is the same
+as in the original, there may be gaps resulting from many-to-one, many-to-many, or one-to-many
+alignments that were filtered out. Therefore, this release may not be suitable for
+discourse-related research.
+2. Neither the sentence splitting nor the alignments are perfect. In particular, watch out for
+pairs that differ considerably in length. You may want to filter these out before you do
+any statistical training.
+
+The alignment of the Hansards was performed as part of the ReWrite project under funding
+from the DARPA TIDES program.
+"""
+
+_URL = "https://www.isi.edu/natural-language/download/hansard/"
+_DATA_URL = "http://www.isi.edu/natural-language/download/hansard/"
+_HOUSE_DEBATES_TRAIN_SET_FILE = "hansard.36.r2001-1a.house.debates.training.tar"
+_HOUSE_DEBATES_TEST_SET_FILE = "hansard.36.r2001-1a.house.debates.testing.tar"
+_SENATE_DEBATES_TRAIN_SET_FILE = "hansard.36.r2001-1a.senate.debates.training.tar"
+_SENATE_DEBATES_TEST_SET_FILE = "hansard.36.r2001-1a.senate.debates.testing.tar"
+
+
+class HansardsConfig(datasets.BuilderConfig):
+    """BuilderConfig for Hansards."""
+
+    def __init__(self, **kwargs):
+        """BuilderConfig for Hansards.
+        Args:
+          **kwargs: keyword arguments forwarded to super.
+        """
+        super(HansardsConfig, self).__init__(version=datasets.Version("1.0.0", ""), **kwargs)
+
+
+class Hansards(datasets.GeneratorBasedBuilder):
+    """TODO(hansards): Short description of my dataset."""
+
+    # TODO(hansards): Set up version.
+    VERSION = datasets.Version("0.1.0")
+    BUILDER_CONFIGS = [
+        HansardsConfig(
+            name="house",
+            description="""\
+          Alignment of debates in the House of the 36th Canadian Parliament: 1,070K sentence pairs.
+          """,
+        ),
+        HansardsConfig(
+            name="senate",
+            description="""\
+          Alignment of debates in the Senate of the 36th Canadian Parliament: 208K sentence pairs.
+          """,
+        ),
+    ]
+
+    def _info(self):
+        # TODO(hansards): 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(
+                {
+                    "fr": datasets.Value("string"),
+                    "en": 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=_URL,
+            citation=_CITATION,
+        )
+
+    def _split_generators(self, dl_manager):
+        """Returns SplitGenerators."""
+        # TODO(hansards): Downloads the data and defines the splits
+        # dl_manager is a datasets.download.DownloadManager that can be used to
+        # download and extract URLs
+        name = self.config.name
+        if name == "house":
+            urls_to_download = {
+                "train": os.path.join(_DATA_URL, _HOUSE_DEBATES_TRAIN_SET_FILE),
+                "test": os.path.join(_DATA_URL, _HOUSE_DEBATES_TEST_SET_FILE),
+            }
+        elif name == "senate":
+            urls_to_download = {
+                "train": os.path.join(_DATA_URL, _SENATE_DEBATES_TRAIN_SET_FILE),
+                "test": os.path.join(_DATA_URL, _SENATE_DEBATES_TEST_SET_FILE),
+            }
+        else:
+            raise ValueError("Wrong builder config name '{}', it has to be either 'house' or 'senate'.".format(name))
+        downloaded_files = dl_manager.download_and_extract(urls_to_download)
+        if type(downloaded_files) == str:
+            downloaded_files = {k: downloaded_files for k in urls_to_download.keys()}
+        fr_files = {}
+        en_files = {}
+        for split_name in downloaded_files.keys():
+            archive_dir = "hansard.36/Release-2001.1a/sentence-pairs/{}/debates/development/{}".format(
+                name, split_name + "ing"
+            )
+            data_dir = os.path.join(downloaded_files[split_name], archive_dir)
+            split_compress_files = list(sorted(glob.glob(os.path.join(data_dir, "*.gz"))))
+            split_compress_files += list(sorted(glob.glob(os.path.join(data_dir, "**/*.gz"))))
+            fr_split_compress_files = sorted([f for f in split_compress_files if f.endswith(".f.gz")])
+            en_split_compress_files = sorted([f for f in split_compress_files if f.endswith(".e.gz")])
+            fr_files[split_name] = dl_manager.extract(fr_split_compress_files)
+            en_files[split_name] = dl_manager.extract(en_split_compress_files)
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                # These kwargs will be passed to _generate_examples
+                gen_kwargs={"fr_files": fr_files["train"], "en_files": en_files["train"]},
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split.TEST,
+                # These kwargs will be passed to _generate_examples
+                gen_kwargs={"fr_files": fr_files["test"], "en_files": en_files["test"]},
+            ),
+        ]
+
+    def _generate_examples(self, fr_files, en_files):
+        """Yields examples."""
+        # TODO(hansards): Yields (key, example) tuples from the dataset
+        for fr_file, en_file in zip(fr_files, en_files):
+            with open(fr_file, "rb") as fr:
+                with open(en_file, "rb") as en:
+                    for j, (fr_line, en_line) in enumerate(zip(fr, en)):
+                        line_id = "{}:{}".format(fr_file, j)
+                        rec = {"fr": fr_line.decode("ISO-8859-1").strip(), "en": en_line.decode("ISO-8859-1").strip()}
+                        yield line_id, rec