Wikisource:Wikidata

Wikidata

Information about interaction between Wikisource and Wikidata

Wikidata logo

Overview

Example Wikidata data item

Wikidata is a "knowledge base" that allows Wikimedia to keep general information in one place. This means things useful to many projects only need to be recorded and maintained once, rather than each of the hundreds of projects spending time maintaining their own versions of the same thing.

Wikidata is made up of "data items" which each relate to a single thing. For example, each author should have a data item, which will connect to every page about that author on any project via interwiki links, and a list of data statements with facts about the author (like date of birth). Each data item has a unique Q-number ID; for example, the author Douglas Addams is Q42. Similarly, properties are referenced with P-numbers; for example, date of birth is P569.

Each data item is made up of a label, a description, aliases, a section for labels and descriptions in other languages (based on templates on your talk page), statements, and interwiki links. Practically, from Wikisource's point of view, each data item holds two important things: data statements and interwiki links.

Labels and descriptions

The label is the title of the data item in bold text. The description is a short line of text underneath it. They will both be displayed in searches for this item and the label will be used wherever the data item is linked. Both are free text fields.

The label should be short and usually lowercase (unless proper noun). It does not need to be unique; the description should disambiguate between different items. Descriptions are also short and usually lowercase; they should not be sentences, just the minimum amount of information (for example, "American author" or "French physicist"; "novel by Charles Dickens" or "short story by H. P. Lovecraft").

Underneath the description will be the text "Also known as:" and a list of aliases. Aliases work as alternatives to the label. Aliases will be picked up in searches and will function as alternative labels if used as a value in a statement. There is no limit to the number of aliases a data item may have.

Different label and descriptions exist for each language; you will probably be seeing just the English set. A babel magic word on your userpage will both display your proficiency with languages and set which additional sets of labels and descriptions are available when you edit. For example, {{#babel:en|fr|de-2}}, shows fluency in English (en) and French (fr) and a basic understanding of German (de); when editing a data item, the French and German sets of label and description will be shown.

See Wikidata's Help:Label, Help:Description, and Help:Aliases for more information on these things.

Statements

Most of the data on Wikidata is held in the statements. A basic statement is made up of a property and a value (which are together called the "claim"). To this can be added sources and qualifiers. Multiple values can be assigned to each property. Each value can have multiple sources and/or qualifiers. The type of value allowed depends on the property (many are free text, date-based properties will only allow dates, others will only allow links to other data items).

Diagram of a Wikidata statement

See Wikidata's Help:Statements for more information on statements.

Interwiki links

These are called sitelinks on Wikidata. A data item can be connected to a page on Wikisource by going to the data item page on Wikidata and entering the Wikisource page name as a page link. A single data item can only be connected to a single page on each project and vice versa.

Wikisource links are held under the heading Wikisource pages linked to this item. Each page link has three parts: language, code, and linked page.

Interwiki links and some sister links are controlled via this process (both can still be manually entered on Wikisource).

A page on Wikisource can suppress interwiki links held by Wikidata. Adding the {{noexternallanglinks}} magic word to a page suppresses all interwiki links. Specific languages can be suppressed by listing the language code in the magic word; i.e. {{noexternallanglinks:es|fr|it}} to suppress links to the Spanish, French and Italian Wikisources.

Adding interwiki links

• With Javascript enabled:
  1. Find or create the data item for the page you want connected to Wikidata.
  2. Go to the section entitled "Wikisource pages linked to this item".
  3. Click the [add] link.
  4. Two fields will appear:
     1. Enter the "site" (the first field) as "enwikisource". Simply typing "en" will normally be enough as the rest will be auto-completed. Either way, there will be a short period of buffering while Wikidata processes this.
     2. Enter the "page" (the second field) as the page name, as it appears on Wikisource. There will be a short period of buffering while Wikidata processes this, after which a list of matching pages on Wikisource will appear below the field. The page link can be chosen from this list or, if complete, left as it is.
  5. Click save.
• With Javascript disabled:
  1. Find or create the data item for the page you want connected to Wikidata.
  2. Go to the section entitled "Wikisource pages linked to this item".
  3. Click the [add] link.
  4. You will be taken to a new page with the title "Set a site link" and three fields:
     1. The first field, "ID", should be preloaded with the data item ID (Q-number). If not, fill this in.
     2. In the second field, "Site id", enter "enwikisource".
     3. In the third field, "Site link", enter the page name as it appears on Wikisource.
  5. Click the "Set the site link" button.

The "language" will be derived from the site, which will appear under "code". The page title will be shown as a link under "linked page".

Adding status badge

There is the ability to add a status badge to a link to indicate the transcription status of the linked work

• not proofread Q20748091
• proofread Q20748092
• validated Q20748093
• problematic Q20748094

This status is based on the status of the index: page for the work.

To add the status, it can be added at the time of adding the link, or later by editing the link.

1. When editing the interwiki links, click on the badge icon to the right of the link.
2. Hover over the ellipsis symbol will produce a drop down list of status, the four Wikisource badges are at the bottom of the list
3. Select the appropriate status for the work
4. Save the interwiki link.

Editing interwiki links

• With Javascript enabled:
  1. Find or create the data item you want to edit.
  2. Go to the section entitled "Wikisource pages linked to this item".
  3. Find the specific page link to be editing.
  4. Click the [edit] link at the end of this row.
  5. Make changes:
     1. The page link can be edited but the site cannot.
     2. The entire link can be deleted by clicking the "remove" link.
     3. The page link can be moved to a different data link by clicking the "move" link. Wikidata will ask for the item id (Q-number) of the page the link is to be moved to.
  6. Add a status badge for the work
  7. Save the changes or click cancel to abandon them.

Data statements

When used on a page that is connected to Wikidata, the #property parser function can access data from that item.

Properties can be accessed by the name of the property or by the property number. For example:

• {{#property:genre}} or {{#property:P136}} will return the "genre" value.
• {{#property:date of birth}} or {{#property:P569}} will return the date of birth value.

Wikidata has a list of its properties, to show what can be returned (if the data item contains the property as one of its statements).

Modules

Data can also be accessed with Lua modules, which are much more flexible.

Current Wikidata related modules and their uses

Return	Module	Function
ID number	Wikibase	{{#invoke:Wikibase|id}}
Label (page title) on Wikidata	Wikibase	{{#invoke:Wikibase|label}}
Page title on Wikisource	Wikibase	{{#invoke:Wikibase|page}}
Wikilinked value	Wikidata	{{#invoke:Wikidata|getValue}}
Non-wikilinked value	Wikidata	{{#invoke:Wikidata|getRawValue}}
Non-wikilinked date	Wikidata	{{#invoke:Wikidata|getDateValue}}

Queries

Wikidata queries can be formatted using {{SPARQL}}.

See also Category:Wikidata queries

Managing data items

Books Taskforce logo.

Authors

Important statements to use for all authors are:

• instance of (P31) → human (Q5)
• sex or gender (P21) → as appropriate for the author—the available options are male (Q6581097), female (Q6581072), intersex (Q1097630), transgender female (Q1052281), transgender male (Q15145783), or genderqueer (Q48270)

Other common statements to add are:

• date of birth (P569)
• date of death (P570)
• place of birth (P19)
• place of death (P20)
• country of citizenship (P27)
• occupation (P106) → options include author Q482980, as well as specific professions
• field of work (P101) → compliments occupation, options include physics, chemistry (or specific sub-fields), etc
• position held (P39)
• image (P18) → filename on Commons, usually a portrait of the author
• native language (P103)
• Any authority control statement, such as VIAF identifier (P214)

Consider adding instance of→author (or politician, etc) by clicking "[add]" in the bottom right of the existing instance of statement

Works

Works on Wikidata are split into two data items:

• Work items: For the work in general.
• Edition items: For each specific version or edition.

Each work should have a work item and at least one edition. There may be multiple editions on one Wikisource. Each language subdomain probably uses a different edition, which are all linked to the same work item. Wikipedia and most other projects will usually link directly to the work item, unless there is something of note about a given edition.

Work items should hold information true to all editions, versions, manifestations, or implementations of the work. This includes the title, author and genre, as well as any generally-important information about the first edition such as the original language, country of origin and original publication date.

Edition items should hold information the applies just to that version of the work. This includes some generic information, such as title and author. The more edition-specific data should be the publisher (along with location and date of publication), any additional work that applies to that edition (such as illustrations, translation or editing credits), and so on. If proofread, the scan on Commons can be linked as well.

Example

Work item	Edition items
Treasure Island (Q185118)	Treasure Island (1883), published in the UK by Cassell & Company (Q14944007) Treasure Island (1911), published in the US by Charles Scriber's Sons (Q14944010)

Which could be represented as:-

diagram of the above table

See Wikidata's Books task force for more information. See Wikidata:List of properties for a list of work- and edition-specific properties.

Work items

Important statements to use for all works are:

• instance of (P31) → as appropriate for the work, options include: book (Q571), creative work (Q386724), novel (Q8261), short story (Q49084), etc (both "book" and "creative work" combined are quite common)
• edition (P747) → the corresponding edition item(s)
• author (P50)
• title (P357)
• subtitle (P392)
• language of work or name (P407)

Other common statements to add are:

• image (P18) → filename on Commons, often the most popular or widely-recognised cover or title page
• country of origin (P495)
• genre (P136)
• subject heading (P921) → data item of the topic or subject of the work (reverse property subject of (P805) can link back)
• narrative set in (P840)
• characters (P110674)
• preceded by (P155) → data item of the previous item in the series
• succeeded by (P156) → data item of the sequel or next in series
• series (P179) → data item of the series in which this work is an instalment
• dedicated to (P825)

Edition items

Important statements to use for all editions are:

• instance of (P31) → as appropriate for the work, options include: edition (Q3331189)
• edition or translation of (P629) → the corresponding work item
• title (P357)
• subtitle (P392)
• author (P50)
• date of publication (P577)
• place of publication (P291)
• publisher (P123)
• scan file (P996)
• language (P407)
• full text available at (P953) → URL for the text on Wikisource (although this is redundant to the Wikisource site link, it may be of use to other, non-Wikimedia, users of Wikidata's data)
• license (P275) → matches the licence template, eg. public domain (Q19652), CC-BY-SA (Q6905942), etc (most PD-x licences should just used public domain on Wikidata)

Other common statements to add are:

• image (P18) → filename on Commons, often the cover or title page of the edition
• Any authority control statement, such as:
  • LCOC LCCN (bibliographic) (P1144)
  • OCLC control number (P243)
  • Open Library identifier (P648)
• translator (P955)
• editor (P98)
• illustrator (P110)
• country of origin (P495) → country data item, useful for work published in countries other than the original
• language (P407) → language of this edition, useful for translations
• edition number (P393) → number of the edition (e.g. first, second, 1, 2, etc)
• number of pages (P1104)
• Wikisource index page (Property:P1957)

Everything else

Portals should use instance of→Wikimedia portal and portal's main topic (P1204). The topic data item should use topic's main portal (P1151) to point back to the portal.

Maintenance

Some maintenance resources for Wikidata.

Tracking categories:

• Category:Author pages connected to Wikidata, lists all pages in the Author namespace currently linked to Wikidata
• Category:Author pages not connected to Wikidata, lists all pages in the Author namespace that have not yet been linked to Wikidata.

Wikidata queries:

• all books before 1923 in English, without wikisource page, ordered by decreasing number of sitelinks : this list shows the most well known books without wikisource page, either because the wikisource page was not added to wikidata, or because the book has not been transcribed yet.

Useful pages:

• Wikidata maintenance subproject
  • Intersection of authors linked to Wikipedia but not Wikidata

See also

• Wikidata:Wikisource