Lichess/chess-openings-translations

Parallel corpus of chess opening names, pulled from Wikidata (https://query.wikidata.org/) on May 12th, 2026.

SARQL queries:

SELECT ?item ?lang ?label WHERE {
  ?item wdt:P31 wd:Q103632 .
  ?item rdfs:label ?label .
  BIND(LANG(?label) AS ?lang)
}

3994 results in 650 ms (labels.csv)

SELECT ?item ?lang ?title WHERE {
  ?item wdt:P31 wd:Q103632 .
  ?sitelink schema:about ?item ;
            schema:isPartOf [ wikibase:wikiGroup "wikipedia" ] ;
            schema:inLanguage ?lang ;
            schema:name ?title .
}

2983 results in 749 ms (sitelinks.csv)

import pandas as pd
from datasets import Dataset
labels = pd.read_csv("labels.csv")
sitelinks = pd.read_csv("sitelinks.csv")
en_labels = labels[labels["lang"] == "en"][["item", "label"]].rename(columns={"label": "label_en"})
en_titles = sitelinks[sitelinks["lang"] == "en"][["item", "title"]].rename(columns={"title": "title_en"})
df = labels.merge(sitelinks, on=["item", "lang"], how="outer").merge(en_labels, on="item", how="left").merge(en_titles, on="item", how="left")
df["item"] = df["item"].str.replace("http://www.wikidata.org/entity/", "")
Dataset.from_pandas(df).select_columns(['item', 'lang', 'label_en', 'label', 'title_en', 'title']).push_to_hub("Lichess/chess-openings-translations")

Licensing

Structured data from Wikidata is cc0: https://www.wikidata.org/wiki/Wikidata:Licensing