Wikidata dump extension (enwiki section links)

The dataset contains mappings between Wikidata entities and Wikipedia sections. The mappings come in addition to the existing Wikidata sitelinks referencing Wikipedia pages. The creation of the present dataset stems from the observation that only a fraction of Wikidata entities has a corresponding Wikipedia article in any language (we refer to the remaining entities, without an article, as orphans). However, a substantial number of orphan entities are indeed available in Wikipedia, but not at the page level; orphan entities can be described within existing Wikipedia articles in the form of sections, subsections, and paragraphs of a more generic concept or fact. The dataset provides a fine-grained mapping between Wikidata orphan entities and Wikipedia (sub)-sections. Mappings are provided for English language. The dataset is available in JSON and RDF formats and complies with the Wikibase data model. In the JSON representation, an entity contains two fields: id (the unique identifier of an entity) and sectionlinks (links to Wikipedia sections). Each sectionlink record comprises a list of records1 with three fields: site, title, and url. A section title is appended to the page title separated with # symbol. Such a compound title is then URL-encoded and added to the URL path. Following the Wikidata guidelines, each entity is encoded as a single line. Example: {
"id": "Q715509",
"sectionlinks": {
"enwiki": [
{
"site": "enwiki",
"title": "Places in Harry Potter#Azkaban",
"url": "https://en.wikipedia.org/wiki/Places_in_Harry_Potter#Azkaban"
}
],
}
} The RDF dump is serialized using the Turtle format and stores nodes describing Wikipedia links. Section titles are added in the same manner as described above. Example: <https://en.wikipedia.org/wiki/Places_in_Harry_Potter#Azkaban> a schema:Article ;
schema:about wd:Q715509 ;
schema:inLanguage "en" ;
schema:isPartOf <https://en.wikipedia.org/> ;
schema:name "Places in Harry Potter#Azkaban"@en .

<https://en.wikipedia.org/> wikibase:wikiGroup "wikipedia" .
1 As opposed to sitelinks, where each entity can be mapped with a unique Wikipedia page (one-to-one mapping), in sectionlinks we allow a one-to-many mapping, i.e., an entity can be mapped to multiple sections. For example, Tennis racket concept can be mapped to Tennis#Rackets and Racket (sports equipment)#Tennis sections.

本数据集包含维基数据（Wikidata）实体与维基百科（Wikipedia）章节间的映射关系。此类映射是对现有引用维基百科页面的维基数据站点链接（sitelinks）的补充。

本数据集的构建源于一项观察：仅一小部分维基数据实体拥有对应语言的维基百科条目，我们将这类无对应条目的剩余实体称为孤立实体（orphans）。然而，大量孤立实体其实可在维基百科中找到相关内容，只是并非以独立条目形式存在——孤立实体可作为通用概念或事实相关的现有维基百科条目的章节、子章节与段落被描述。

本数据集提供了维基数据孤立实体与维基百科（子）章节间的细粒度映射，仅涵盖英语语言的映射结果。

数据集以JSON与RDF格式发布，且符合维基数据仓库（Wikibase）数据模型。在JSON表示形式中，每个实体包含两个字段：id（实体的唯一标识符）与章节链接（sectionlinks，指向维基百科章节的链接）。每条章节链接记录由包含三个字段的记录列表组成：site、title与url。章节标题会以#符号分隔附加在页面标题之后，该复合标题经URL编码后加入URL路径。遵循维基数据规范，每个实体均以单行形式编码。示例如下：
{
"id": "Q715509",
"sectionlinks": {
"enwiki": [
{
"site": "enwiki",
"title": "Places in Harry Potter#Azkaban",
"url": "https://en.wikipedia.org/wiki/Places_in_Harry_Potter#Azkaban"
}
]
}
}
RDF转储采用Turtle格式进行序列化，存储描述维基百科链接的节点。章节标题的添加方式与前文所述一致。示例如下：
<https://en.wikipedia.org/wiki/Places_in_Harry_Potter#Azkaban> a schema:Article ;
schema:about wd:Q715509 ;
schema:inLanguage "en" ;
schema:isPartOf <https://en.wikipedia.org/> ;
schema:name "Places in Harry Potter#Azkaban"@en .

<https://en.wikipedia.org/> wikibase:wikiGroup "wikipedia" .

1 与站点链接（sitelinks，每个实体可映射至唯一维基百科页面，即一对一映射）不同，章节链接（sectionlinks）支持一对多映射，即一个实体可对应多个章节。例如，“网球拍”（Tennis racket）概念可映射至《Tennis》#Rackets与《Racket (sports equipment)》#Tennis两个章节。

搜集汇总

背景与挑战

背景概述

该数据集是维基数据的一个扩展，专门提供维基数据实体与英文维基百科章节之间的细粒度映射关系，特别关注那些没有独立维基百科文章的'孤儿'实体。它支持一个实体映射到多个章节，并以JSON和RDF格式提供，遵循维基基地数据模型。