README.md

metadata

description: Dataset providing aligned data scraped from Welsh Government Websites.
license: cc
license_name: ogl
language:
  - cy
  - en
multilinguality: translation
config_names:
  - default
  - cadw
  - cdffc
  - comisynyddygymraeg
  - digitalcommunities
  - gwasanaethaucyhoeddusdigidol
  - sirgar
  - www-llyw-cymru
task_ids:
  - natural-language-inference
  - semantic-similarity-scoring
  - sentiment-classification
  - text-scoring
dataset_info:
  - config_name: default
    features:
      - name: text_en
        dtype: string
      - name: text_cy
        dtype: string
      - name: url_en
        dtype: string
      - name: url_cy
        dtype: string
    splits:
      - name: train
        num_bytes: 168665827
        num_examples: 439703
    download_size: 8113815281138152
    dataset_size: 168665827
  - config_name: cadw
    features:
      - name: text_en
        dtype: string
      - name: text_cy
        dtype: string
      - name: url_en
        dtype: string
      - name: url_cy
        dtype: string
    splits:
      - name: train
        num_bytes: 5407552
        num_examples: 15057
    download_size: 2433650
    dataset_size: 5407552
  - config_name: cdffc
    features:
      - name: text_en
        dtype: string
      - name: text_cy
        dtype: string
      - name: url_en
        dtype: string
      - name: url_cy
        dtype: string
    splits:
      - name: train
        num_bytes: 405919
        num_examples: 1076
    download_size: 135557
    dataset_size: 405919
  - config_name: comisynyddygymraeg
    features:
      - name: text_en
        dtype: string
      - name: text_cy
        dtype: string
      - name: url_en
        dtype: string
      - name: url_cy
        dtype: string
    splits:
      - name: train
        num_bytes: 1312843
        num_examples: 3072
    download_size: 430990
    dataset_size: 1312843
  - config_name: digitalcommunities
    features:
      - name: text_en
        dtype: string
      - name: text_cy
        dtype: string
      - name: url_en
        dtype: string
      - name: url_cy
        dtype: string
    splits:
      - name: train
        num_bytes: 2005659
        num_examples: 4717
    download_size: 717803
    dataset_size: 2005659
  - config_name: gwasanaethaucyhoeddusdigidol
    features:
      - name: text_en
        dtype: string
      - name: text_cy
        dtype: string
      - name: url_en
        dtype: string
      - name: url_cy
        dtype: string
    splits:
      - name: train
        num_bytes: 4769378
        num_examples: 11200
    download_size: 1631188
    dataset_size: 4769378
  - config_name: sirgar
    features:
      - name: text_en
        dtype: string
      - name: text_cy
        dtype: string
      - name: url_en
        dtype: string
      - name: url_cy
        dtype: string
    splits:
      - name: train
        num_bytes: 24326865
        num_examples: 58382
    download_size: 10116280
    dataset_size: 24326865
  - config_name: www-llyw-cymru
    features:
      - name: text_en
        dtype: string
      - name: text_cy
        dtype: string
      - name: url_en
        dtype: string
      - name: url_cy
        dtype: string
    splits:
      - name: train
        num_bytes: 130437611
        num_examples: 346199
    download_size: 65672684
    dataset_size: 130437611
configs:
  - config_name: default
    data_files:
      - split: train
        path:
          - data/cadw/0000.parquet
          - data/cdffc/0000.parquet
          - data/comisynyddygymraeg/0000.parquet
          - data/digitalcommunities/0000.parquet
          - data/gwasanaethaucyhoeddusdigidol/0000.parquet
          - data/sirgar/0000.parquet
          - data/www-llyw-cymru/0000.parquet
  - config_name: cadw
    data_files:
      - split: train
        path:
          - data/cadw/0000.parquet
  - config_name: cdffc
    data_files:
      - split: train
        path:
          - data/cdffc/0000.parquet
  - config_name: comisynyddygymraeg
    data_files:
      - split: train
        path:
          - data/comisynyddygymraeg/0000.parquet
  - config_name: digitalcommunities
    data_files:
      - split: train
        path:
          - data/digitalcommunities/0000.parquet
  - config_name: gwasanaethaucyhoeddusdigidol
    data_files:
      - split: train
        path:
          - data/gwasanaethaucyhoeddusdigidol/0000.parquet
  - config_name: www-llyw-cymru
    data_files:
      - split: train
        path:
          - data/www-llyw-cymru/0000.parquet

Llyw Cymru dataset

This is a dataset comprised of aligned sentences scraped from Welsh Government websites.

The data in this dataset is made available under the Open Government License.

Mae hon yn set ddata sy'n cynnwys brawddegau wedi'u halinio wedi'u crafu o wefannau Llywodraeth Cymru.

Mae'r data yn y set ddata hon ar gael o dan y Trwydded Llywodraeth Agored.

Example Usage

Use the datasets module from huggingface to load the dataset:

import datasets
ds = datasets.load_dataset("llyw-cymru-en-cy-ogl")

By default, all the data for each llyw.cymru website is included (the default configuration).

To load the dataset for a specific configuration (i.e: llyw.cymru website), use the following syntax:

config_name = "cadw"
datasets.load_datset("llyw-cymru-en-cy-ogl", config_name)

To see which values you can select for config_name, you can use the following API to get to all the available options:

datasets.get_dataset_config_names("llyw-cymru-en-cy-ogl")