Replace data URL in SAMSum dataset and support streaming (#4254)

* Use Hub data URL in samsum dataset

* Support streaming samsum dataset

* Clean code

* Update metadata JSON

Commit from https://github.com/huggingface/datasets/commit/19c28817d566dd1f2c0426934184c49413f97d6e

dataset_infos.json

@@ -1 +1 @@
-{"samsum": {"description": "\nSAMSum Corpus contains over 16k chat dialogues with manually annotated\nsummaries.\nThere are two features:\n  - dialogue: text of dialogue.\n  - summary: human written summary of the dialogue.\n  - id: id of a example.\n", "citation": "\n@article{gliwa2019samsum,\n  title={SAMSum Corpus: A Human-annotated Dialogue Dataset for Abstractive Summarization},\n  author={Gliwa, Bogdan and Mochol, Iwona and Biesek, Maciej and Wawer, Aleksander},\n  journal={arXiv preprint arXiv:1911.12237},\n  year={2019}\n}\n", "homepage": "https://arxiv.org/abs/1911.12237v2", "license": "CC BY-NC-ND 4.0", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "dialogue": {"dtype": "string", "id": null, "_type": "Value"}, "summary": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "samsum", "config_name": "samsum", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 9479141, "num_examples": 14732, "dataset_name": "samsum"}, "test": {"name": "test", "num_bytes": 534492, "num_examples": 819, "dataset_name": "samsum"}, "validation": {"name": "validation", "num_bytes": 516431, "num_examples": 818, "dataset_name": "samsum"}}, "download_checksums": {"https://arxiv.org/src/1911.12237v2/anc/corpus.7z": {"num_bytes": 2944100, "checksum": "a97674c66726f66b98a08ca5e8868fb8af9d4843f2b05c4f839bc5cfe91e8899"}}, "download_size": 2944100, "post_processing_size": null, "dataset_size": 10530064, "size_in_bytes": 13474164}}
+{"samsum": {"description": "\nSAMSum Corpus contains over 16k chat dialogues with manually annotated\nsummaries.\nThere are two features:\n  - dialogue: text of dialogue.\n  - summary: human written summary of the dialogue.\n  - id: id of a example.\n", "citation": "\n@article{gliwa2019samsum,\n  title={SAMSum Corpus: A Human-annotated Dialogue Dataset for Abstractive Summarization},\n  author={Gliwa, Bogdan and Mochol, Iwona and Biesek, Maciej and Wawer, Aleksander},\n  journal={arXiv preprint arXiv:1911.12237},\n  year={2019}\n}\n", "homepage": "https://arxiv.org/abs/1911.12237", "license": "CC BY-NC-ND 4.0", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "dialogue": {"dtype": "string", "id": null, "_type": "Value"}, "summary": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "samsum", "config_name": "samsum", "version": "0.0.0", "splits": {"train": {"name": "train", "num_bytes": 9479141, "num_examples": 14732, "dataset_name": "samsum"}, "test": {"name": "test", "num_bytes": 534492, "num_examples": 819, "dataset_name": "samsum"}, "validation": {"name": "validation", "num_bytes": 516431, "num_examples": 818, "dataset_name": "samsum"}}, "download_checksums": {"https://huggingface.co/datasets/datafiles/samsum/resolve/main/corpus.7z": {"num_bytes": 2944100, "checksum": "a97674c66726f66b98a08ca5e8868fb8af9d4843f2b05c4f839bc5cfe91e8899"}}, "download_size": 2944100, "post_processing_size": null, "dataset_size": 10530064, "size_in_bytes": 13474164}}

samsum.py

@@ -40,11 +40,11 @@ There are two features:
   - id: id of a example.
 """
 
-_HOMEPAGE = "https://arxiv.org/abs/1911.12237v2"
+_HOMEPAGE = "https://arxiv.org/abs/1911.12237"
 
 _LICENSE = "CC BY-NC-ND 4.0"
 
-_URLs = "https://arxiv.org/src/1911.12237v2/anc/corpus.7z"
+_URL = "https://huggingface.co/datasets/datafiles/samsum/resolve/main/corpus.7z"
 
 
 class Samsum(datasets.GeneratorBasedBuilder):
@@ -75,12 +75,10 @@ class Samsum(datasets.GeneratorBasedBuilder):
 
     def _split_generators(self, dl_manager):
         """Returns SplitGenerators."""
-        my_urls = _URLs
-        path = dl_manager.download_and_extract(my_urls)
+        path = dl_manager.download(_URL)
         return [
             datasets.SplitGenerator(
                 name=datasets.Split.TRAIN,
-                # These kwargs will be passed to _generate_examples
                 gen_kwargs={
                     "filepath": (path, "train.json"),
                     "split": "train",
@@ -88,7 +86,6 @@ class Samsum(datasets.GeneratorBasedBuilder):
             ),
             datasets.SplitGenerator(
                 name=datasets.Split.TEST,
-                # These kwargs will be passed to _generate_examples
                 gen_kwargs={
                     "filepath": (path, "test.json"),
                     "split": "test",
@@ -96,7 +93,6 @@ class Samsum(datasets.GeneratorBasedBuilder):
             ),
             datasets.SplitGenerator(
                 name=datasets.Split.VALIDATION,
-                # These kwargs will be passed to _generate_examples
                 gen_kwargs={
                     "filepath": (path, "val.json"),
                     "split": "val",
@@ -106,12 +102,11 @@ class Samsum(datasets.GeneratorBasedBuilder):
 
     def _generate_examples(self, filepath, split):
         """Yields examples."""
-
         path, fname = filepath
-
-        with py7zr.SevenZipFile(path, "r") as z:
-            for name, bio in z.readall().items():
-                if name == fname:
-                    data = json.load(bio)
+        with open(path, "rb") as f:
+            with py7zr.SevenZipFile(f, "r") as z:
+                for name, bio in z.readall().items():
+                    if name == fname:
+                        data = json.load(bio)
         for example in data:
             yield example["id"], example