Loader script

swahili-safi.py

@@ -0,0 +1,71 @@
+"""Cleaned dataset for Swahili Language Modeling"""
+
+
+import datasets
+
+
+_CITATION = """\
+@InProceedings{huggingface:flax-community,
+title = Cleaned dataset for Swahili Language Modeling,
+authors={Fitsum, Alok, Patrick},
+year={2021},
+link = https://huggingface.co/datasets/flax-community/swahili-safi
+}
+"""
+
+_DESCRIPTION = """Cleaned dataset for Swahili Language Modeling"""
+_HOMEPAGE = "https://huggingface.co/datasets/flax-community/swahili-safi"
+_LICENSE = "Attribution 4.0 International"
+_REPO_URL = "https://huggingface.co/datasets/flax-community/swahili-safi/resolve/main/"
+_TRAIN= [_REPO_URL + file_name for file_name in [
+    "data/train.txt",
+]]
+
+
+class SwahiliSafi(datasets.GeneratorBasedBuilder):
+    """The Swahili dataset for language modeling"""
+
+    VERSION = datasets.Version("1.0.0")
+    BUILDER_CONFIGS = [
+        datasets.BuilderConfig(
+            name="swahili-safi",
+            version=VERSION,
+            description="Language modeling dataset for Swahili"
+        ),
+    ]
+
+    def _info(self):
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            features=datasets.Features(
+                {
+                    "text": datasets.Value("string"),
+                }
+            ),
+            supervised_keys=None,
+            homepage=_HOMEPAGE,
+            license=_LICENSE,
+            citation=_CITATION,
+        )
+
+    def _split_generators(self, dl_manager):
+        """Returns SplitGenerators."""
+        train_files = dl_manager.download(_TRAIN)
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={
+                    "data_files": train_files,
+                    "split": "train",
+                },
+            )
+        ]
+
+    def _generate_examples(self, data_files):
+        """Yields examples."""
+        _id = 0
+        for filepath in data_files:
+            with open(filepath, mode="r", encoding="utf-8") as f:
+                for line in f:
+                    yield _id, {"text": line.strip()},
+                    _id += 1