feat: Add data, script and readme

.gitattributes

@@ -1,54 +1,6 @@
-*.7z filter=lfs diff=lfs merge=lfs -text
-*.arrow filter=lfs diff=lfs merge=lfs -text
-*.bin filter=lfs diff=lfs merge=lfs -text
-*.bz2 filter=lfs diff=lfs merge=lfs -text
-*.ckpt 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
-*.lz4 filter=lfs diff=lfs merge=lfs -text
-*.mlmodel filter=lfs diff=lfs merge=lfs -text
-*.model filter=lfs diff=lfs merge=lfs -text
-*.msgpack filter=lfs diff=lfs merge=lfs -text
-*.npy filter=lfs diff=lfs merge=lfs -text
-*.npz 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
-*.pickle filter=lfs diff=lfs merge=lfs -text
-*.pkl 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
-*.safetensors 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
-*.wasm filter=lfs diff=lfs merge=lfs -text
-*.xz filter=lfs diff=lfs merge=lfs -text
-*.zip filter=lfs diff=lfs merge=lfs -text
-*.zst filter=lfs diff=lfs merge=lfs -text
-*tfevents* filter=lfs diff=lfs merge=lfs -text
-# Audio files - uncompressed
-*.pcm filter=lfs diff=lfs merge=lfs -text
-*.sam filter=lfs diff=lfs merge=lfs -text
-*.raw filter=lfs diff=lfs merge=lfs -text
-# Audio files - compressed
-*.aac filter=lfs diff=lfs merge=lfs -text
-*.flac filter=lfs diff=lfs merge=lfs -text
-*.mp3 filter=lfs diff=lfs merge=lfs -text
-*.ogg filter=lfs diff=lfs merge=lfs -text
-*.wav filter=lfs diff=lfs merge=lfs -text
-# Image files - uncompressed
-*.bmp filter=lfs diff=lfs merge=lfs -text
-*.gif filter=lfs diff=lfs merge=lfs -text
-*.png filter=lfs diff=lfs merge=lfs -text
-*.tiff filter=lfs diff=lfs merge=lfs -text
-# Image files - compressed
-*.jpg filter=lfs diff=lfs merge=lfs -text
-*.jpeg filter=lfs diff=lfs merge=lfs -text
-*.webp filter=lfs diff=lfs merge=lfs -text
+data/nn.jsonl filter=lfs diff=lfs merge=lfs -text
+data/sv.jsonl filter=lfs diff=lfs merge=lfs -text
+data/da.jsonl filter=lfs diff=lfs merge=lfs -text
+data/fo.jsonl filter=lfs diff=lfs merge=lfs -text
+data/is.jsonl filter=lfs diff=lfs merge=lfs -text
+data/nb.jsonl filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1 @@
+.venv

README.md

@@ -1,3 +1,149 @@
 ---
-license: cc-by-sa-4.0
+pretty_name: ScandiQA
+language:
+- da
+- sv
+- no
+license:
+- cc-by-sa-4.0
+multilinguality:
+- multilingual
+size_categories:
+- 1K<n<10K
+source_datasets:
+- mkqa
+- natural_questions
+task_categories:
+- question-answering
+task_ids:
+- extractive-qa
 ---
+
+# Dataset Card for ScandiQA
+
+## Dataset Description
+
+- **Repository:** <https://github.com/alexandrainst/scandi-qa>
+- **Point of Contact:** [Dan Saattrup Nielsen](mailto:dan.nielsen@alexandra.dk)
+- **Size of downloaded dataset files:** 69 MB
+- **Size of the generated dataset:** 67 MB
+- **Total amount of disk used:** 136 MB
+
+### Dataset Summary
+
+ScandiQA is a dataset of questions and answers in the Danish, Norwegian, and Swedish
+languages. All samples come from the Natural Questions (NQ) dataset, which is a large
+question answering dataset from Google searches. The Scandinavian questions and answers
+come from the MKQA dataset, where 10,000 NQ samples were manually translated into,
+among others, Danish, Norwegian, and Swedish. However, this did not include a
+translated context, hindering the training of extractive question answering models.
+
+We merged the NQ dataset with the MKQA dataset, and extracted contexts as either "long
+answers" from the NQ dataset, being the paragraph in which the answer was found, or
+otherwise we extract the context by locating the paragraphs which have the largest
+cosine similarity to the question, and which contains the desired answer.
+
+Further, many answers in the MKQA dataset were "language normalised": for instance, all
+date answers were converted to the format "YYYY-MM-DD", meaning that in most cases
+these answers are not appearing in any paragraphs. We solve this by extending the MKQA
+answers with plausible "answer candidates", being slight perturbations or translations
+of the answer.
+
+With the contexts extracted, we translated these to Danish, Swedish and Norwegian using
+the [DeepL translation service](https://www.deepl.com/pro-api?cta=header-pro-api) for
+Danish and Swedish, and the [Google Translation
+service](https://cloud.google.com/translate/docs/reference/rest/) for Norwegian. After
+translation we ensured that the Scandinavian answers do indeed occur in the translated
+contexts.
+
+As we are filtering the MKQA samples at both the "merging stage" and the "translation
+stage", we are not able to fully convert the 10,000 samples to the Scandinavian
+languages, and instead get roughly 8,000 samples per language. These have further been
+split into a training, validation and test split, with the latter two containing
+roughly 750 samples. The splits have been created in such a way that the proportion of
+samples without an answer is roughly the same in each split.
+
+
+### Supported Tasks and Leaderboards
+
+Training machine learning models for extractive question answering is the intended task
+for this dataset. No leaderboard is active at this point.
+
+
+### Languages
+
+The dataset is available in Danish (`da`), Swedish (`sv`) and Norwegian (`no`).
+
+
+## Dataset Structure
+
+### Data Instances
+
+- **Size of downloaded dataset files:** 69 MB
+- **Size of the generated dataset:** 67 MB
+- **Total amount of disk used:** 136 MB
+
+An example from the `train` split of the `da` subset looks as follows.
+```
+{
+    'example_id': 123,
+    'question': 'Er dette en test?',
+    'answer': 'Dette er en test',
+    'answer_start': 0,
+    'context': 'Dette er en testkontekst.',
+    'answer_en': 'This is a test',
+    'answer_start_en': 0,
+    'context_en': "This is a test",
+    'title_en': 'Train test'
+}
+```
+
+### Data Fields
+
+The data fields are the same among all splits.
+
+- `example_id`: an `int64` feature.
+- `question`: a `string` feature.
+- `answer`: a `string` feature.
+- `answer_start`: an `int64` feature.
+- `context`: a `string` feature.
+- `answer_en`:  a `string` feature.
+- `answer_start_en`: an `int64` feature.
+- `context_en`: a `string` feature.
+- `title_en`: a `string` feature.
+
+### Data Splits
+
+|   name   | train | validation | test |
+|----------|------:|-----------:|-----:|
+| da       | 6311  | 749        | 750  |
+| sv       | 6299  | 750        | 749  |
+| no       | 6314  | 749        | 750  |
+
+
+## Dataset Creation
+
+### Curation Rationale
+
+The Scandinavian languages does not have any gold standard question answering dataset.
+This is not quite gold standard, but the fact both the questions and answers are all
+manually translated, it is a solid silver standard dataset.
+
+### Source Data
+
+The original data was collected from the [MKQA](https://github.com/apple/ml-mkqa/) and
+[Natural Questions](https://ai.google.com/research/NaturalQuestions) datasets from
+Apple and Google, respectively.
+
+
+## Additional Information
+
+### Dataset Curators
+
+[Dan Saattrup Nielsen](https://saattrupdan.github.io/) from the [The Alexandra
+Institute](https://alexandra.dk/) curated this dataset.
+
+### Licensing Information
+
+The dataset is licensed under the [CC BY-SA 4.0
+license](https://creativecommons.org/licenses/by-sa/4.0/).

data/da.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e524f30d6b2528b2fad64c8cb5800fbc4860c93731288893a645c8731784b4a5
+size 586791170

data/fo.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d199629c089a9a3b5df9cc64427737c4496897b6717bde7b3eea7b588ee0c8ed
+size 18919558

data/is.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:693254d4d0248648f6c985781e98f3a4d3f3cd0a1f513987a309cc65eccc6b50
+size 114716852

data/nb.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9b73ea2f7f06682c6b2a631f0d3b2f3dc609ebe8b60352bd4e935f5ef15312c
+size 1094829731

data/nn.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:65040a3d175e5717bae504441fbad6fe1870119c5cc93a1ab12727b07f5eb02b
+size 255058352

data/sv.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a57244428c760f9b07d5ba2560d4d2f0d65f8c3a517d879e3f6c81c3550ba5b
+size 2424831491

scandi_wiki.py

@@ -0,0 +1,126 @@
+# Copyright 2022 The HuggingFace Datasets Authors and Dan Saattrup Nielsen.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""Python build script for the ScandiWiki dataset."""
+
+
+import json
+from pathlib import Path
+from typing import List
+
+import datasets
+from datasets import Version
+from datasets.builder import BuilderConfig, GeneratorBasedBuilder
+from datasets.download import DownloadManager
+from datasets.features import Features, Value
+from datasets.info import DatasetInfo
+from datasets.splits import SplitGenerator
+
+_DESCRIPTION = """
+ScandiWiki is a parsed and deduplicated version of the Danish, Norwegian Bokmål,
+Norwegian Nynorsk, Swedish, Icelandic and Faroese Wikipedia corpora, as of January
+2023.
+"""
+
+_LICENSE = "CC BY-SA 4.0"
+_URLS = {
+    "da": "https://huggingface.co/datasets/alexandrainst/scandi-wiki/resolve/main/data/da.jsonl",
+    "sv": "https://huggingface.co/datasets/alexandrainst/scandi-wiki/resolve/main/data/sv.jsonl",
+    "nb": "https://huggingface.co/datasets/alexandrainst/scandi-wiki/resolve/main/data/nb.jsonl",
+    "nn": "https://huggingface.co/datasets/alexandrainst/scandi-wiki/resolve/main/data/nn.jsonl",
+    "is": "https://huggingface.co/datasets/alexandrainst/scandi-wiki/resolve/main/data/is.jsonl",
+    "fo": "https://huggingface.co/datasets/alexandrainst/scandi-wiki/resolve/main/data/fo.jsonl",
+}
+
+# _CITATION = """
+# @InProceedings{huggingface:dataset,
+# title = {ScandiWiki: A Scandinavian Wikipedia Dump},
+# author={Dan Saattrup Nielsen},
+# year={2022}
+# }
+# """
+
+
+class ScandiWiki(GeneratorBasedBuilder):
+    """Scandinavian part of Wikipedia."""
+
+    VERSION = Version("1.0.0")
+
+    BUILDER_CONFIGS = [
+        BuilderConfig(
+            name="da",
+            version=VERSION,
+            description="The deduplicated Danish part of Wikipedia.",
+        ),
+        BuilderConfig(
+            name="sv",
+            version=VERSION,
+            description="The deduplicated Swedish part of Wikipedia.",
+        ),
+        BuilderConfig(
+            name="nb",
+            version=VERSION,
+            description="The deduplicated Norwegian Bokmål part of Wikipedia.",
+        ),
+        BuilderConfig(
+            name="nn",
+            version=VERSION,
+            description="The deduplicated Norwegian Nynorsk part of Wikipedia.",
+        ),
+        BuilderConfig(
+            name="is",
+            version=VERSION,
+            description="The deduplicated Icelandic part of Wikipedia.",
+        ),
+        BuilderConfig(
+            name="fo",
+            version=VERSION,
+            description="The deduplicated Faroese part of Wikipedia.",
+        ),
+    ]
+
+    def _info(self) -> DatasetInfo:
+        features = Features(
+            {
+                "id": Value("string"),
+                "url": Value("string"),
+                "title": Value("string"),
+                "text": Value("string"),
+            }
+        )
+        return DatasetInfo(
+            description=_DESCRIPTION,
+            features=features,
+            # homepage=_HOMEPAGE,
+            license=_LICENSE,
+            # citation=_CITATION,
+        )
+
+    def _split_generators(self, dl_manager: DownloadManager) -> List[SplitGenerator]:
+        url = _URLS[self.config.name]
+        downloaded_file = dl_manager.download(url)
+        return [
+            SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs=dict(
+                    filepath=downloaded_file,
+                    split="train",
+                ),
+            ),
+        ]
+
+    def _generate_examples(self, filepath: str, split):
+        with Path(filepath).open(encoding="utf-8") as f:
+            for key, row in enumerate(f):
+                data = json.loads(row)
+                yield key, data