add dataset script and documentation

README.md

@@ -1,3 +1,50 @@
 ---
+languages:
+    - ha
+    - so
+    - sw
+    - yo
+
+mutilinguality:
+    - multilingual
+
+task-categories:
+    - text-retrieval
+
 license: apache-2.0
+
+viewer: true
 ---
+
+# Dataset Summary
+
+CIRAL is a collection for cross-lingual information retrieval research across four (4) African languages. The collection comprises English queries and query-passage relevance judgements manually annotated by native speakers.
+
+This dataset stores passages which have been culled from news websites for CIRAL.
+
+## Dataset Structure
+
+This dataset is configured by language. An example of a passage data entry is
+
+```json
+{
+    'docid': 'DOCID#0#0',
+    'title': 'This is the title of a sample passage',
+    'text': 'This is the content of a sample passage',
+    'url': 'https:/\/\this-is-a-sample-url.com'
+}
+```
+
+## Load Dataset
+
+An example to load the dataset
+
+```python
+language = "hausa"
+
+dataset = load_dataset("castorini/ciral-corpus", language)
+```
+
+## Citation
+
+...

ciral.py

@@ -0,0 +1,78 @@
+import json
+from string import Template
+
+import datasets
+
+_CITATION = ""       # TODO: @theyorubayesian
+_DESCRIPTION = \
+    """
+    A collection of passages culled from news websites for Cross-Lingual Information Retrieval for African Languages.
+    """
+_HOMEPAGE = "https://github.com/castorini/ciral"
+_LICENSE = "Apache License 2.0"        
+_VERSION = "1.0.0"
+_LANGUAGES = [
+    "hausa",
+    "somali",
+    "swahili",
+    "yoruba",
+    "combined"
+]
+_DATASET_URL = Template("./passages-v1.0/${language}_passages.jsonl")
+
+
+class CiralPassages(datasets.GeneratorBasedBuilder):
+    BUILDER_CONFIGS = [
+        datasets.BuilderConfig(
+            version=datasets.Version(_VERSION),
+            name=language,
+            description=f"CIRAL passages for language: {language}"
+        ) for language in _LANGUAGES
+    ]
+    DEFAULT_CONFIG_NAME = "combined"
+
+    def _info(self):
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            citation=_CITATION,
+            features=datasets.Features({
+                "docid": datasets.Value("string"),
+                "title": datasets.Value("string"),
+                "text": datasets.Value("string"),
+                "url": datasets.Value("string")
+            }),
+            homepage=_HOMEPAGE,
+            license=_LICENSE
+        )
+
+    def _split_generators(self, dl_manager: datasets.DownloadManager):
+        language = self.config.name
+        if language == "combined":
+            language_file = dl_manager.download_and_extract({
+                _language: _DATASET_URL.substitute(language=_language)
+                for _language in _LANGUAGES[:-1]
+            })
+
+            splits = [
+                datasets.SplitGenerator(
+                    name=_language, gen_kwargs={"filepath": language_file[_language]}
+                ) for _language in _LANGUAGES[:-1]
+            ]
+            return splits
+        else:
+            language_file = dl_manager.download_and_extract(
+                _DATASET_URL.substitute(language=language))
+
+            splits = [
+                datasets.SplitGenerator(
+                    name="train",
+                    gen_kwargs={"filepath": language_file}
+                )
+            ]
+            return splits
+
+    def _generate_examples(self, filepath: str):
+        with open(filepath, encoding="utf-8") as f:
+            for line in f:
+                data = json.loads(line)
+                yield data["docid"], data