Help:Contributing/import: Difference between revisions

From Wikibase
Jump to navigation Jump to search
← Older edit

SeedBot (talk | contribs)

4 user groups
5,323 edits
Restore full page + round-2 edits (previous update replaced the whole page by mistake) (via update-page on MediaWiki MCP Server)
Add "When the classic page already exists" section (Sep 22 2026) (via update-page on MediaWiki MCP Server)
 
(22 intermediate revisions by the same user not shown)
Line 1: Line 1:
Instead of hand-creating an item per author or source, use the external-authority pages in the '''semantic tools''' sidebar.
Instead of hand-creating an item per author or source, use the external-authority pages in the '''semantic tools''' sidebar.


Line 15: Line 16:
|}
|}


Every import flow offers a '''manual fallback''' — ''Create the item manually instead'' — on the search page and on the candidate-selection page. The manual form is '''prefilled with what you typed in the search boxes''' (a person name is split into given/family, e.g. "Marie Curie" → given ''Marie'', family ''Curie''), so you correct instead of retyping.
Every import flow offers a '''manual fallback''' — ''Create the item manually instead'' — on the search page and on the candidate-selection page. The manual form is '''prefilled with what you typed in the search boxes''' (a person name is split into given/family, e.g. "Marie Curie" → given ''Marie'', family ''Curie''), so you correct instead of retyping. When the form is prefilled from '''fetched''' data (the website/web page URL entry), the form is titled '''''Item details''''' — you are verifying fetched content, not entering an item from blank.
 
== People: lifecycle fields ==


The person review step pre-fills these from the authority when available:


* '''Given name''' / '''Family name''' — the item's '''label is the full name, auto-generated''' from these two fields (there is no separate label field). The <code>Person:</code> page title follows the same rule; a harvested record that has only a label keeps it.
* '''Place of birth''' / '''Place of death''' — '''OpenStreetMap search boxes''': type a place name and pick the place from the suggestions (searched on OpenStreetMap's Nominatim — villages, cities and regions alike, no local place items needed). Picking a suggestion stores the place's '''OpenStreetMap id''' (<syntaxhighlight lang="text" inline>node/123456</syntaxhighlight>, <syntaxhighlight lang="text" inline>way/…</syntaxhighlight> or <syntaxhighlight lang="text" inline>relation/…</syntaxhighlight>) as an <code>place of birth (OSM)</code> / <code>place of death (OSM)</code> statement; the <code>Person:</code> page infobox renders it as a link to openstreetmap.org. A place '''fetched from the authority''' is automatically looked up on OpenStreetMap and the best match is proposed: ''"{field} fetched from source: {place}, we think this corresponds to {name} ({id})."'' — '''Yes, that's right''' keeps it, '''No, let me correct''' clears the box for you to search. No match shows ''"External record: {place} — search OpenStreetMap to confirm the place, then pick it from the suggestions."'' — the field must hold a picked OpenStreetMap id to save (a raw place name alone is rejected).
* '''Description''' — a one-line summary of the person (below the name fields; pre-filled from the authority when available). Descriptions can be up to '''2000 characters''' on this instance.
* '''Date of birth''' / '''Date of death''' — day-precision date fields (<code>YYYY-MM-DD</code>).
* '''Place of birth''' / '''Place of death''' — entity search boxes: type a name to pick an existing place item (or type a Q-id directly). Create missing places first (see [[Help:Contributing/entities]]).
* '''This person is deceased''' — toggle, off by default; tick it to reveal the date/place of death fields.
* '''OpenAlex ID''' — the person's OpenAlex author identifier, fetched from the authority when available (otherwise type it, or leave empty).
* '''Portrait''' — optional: upload a local image file or paste an image URL (toggle), plus the '''portrait license''' (required when a portrait is provided). The file is saved as <code>File:&lt;name&gt;-portrait.&lt;ext&gt;</code> and linked from the item with an <code>image</code> statement.


== Fictional characters ==
== Fictional characters ==


[[Special:AddFictionalCharacter]] searches Wikidata by name (search → review candidates → create). The review step has '''Given name''' / '''Family name''' and '''Appears in''' (entity search, several works allowed). The '''label auto-generates''' as <syntaxhighlight lang="text" inline>{given name} {family name} (fictional character)</syntaxhighlight>; the '''description''' auto-generates as <syntaxhighlight lang="text" inline>fictional character in {works}</syntaxhighlight> when left blank. Characters get no classic page — just the item.
[[Special:AddFictionalCharacter]] searches Wikidata by name (search → review candidates → create). The review step has '''Given name''' / '''Family name''' and '''Appears in''' (entity search, several works allowed). The '''label auto-generates''' as <syntaxhighlight lang="text" inline>{given name} {family name} (fictional character)</syntaxhighlight>; the '''description''' auto-generates as <syntaxhighlight lang="text" inline>fictional character in {works}</syntaxhighlight> when left blank. Characters get no classic page — just the item.


== Collectives: common classes ==
== Collectives: common classes ==
Line 40: Line 33:


* '''Parent organization''' (optional) — an entity search box: the organization this collective is a subsidiary or subdivision of (type a name to pick an existing item, or type a Q-id).
* '''Parent organization''' (optional) — an entity search box: the organization this collective is a subsidiary or subdivision of (type a name to pick an existing item, or type a Q-id).
* '''Logo''' (optional) — upload a local image file or paste an image URL (toggle), plus the '''logo license''' (required when a logo is provided). The file is saved as <code>File:&lt;name&gt;-logo.&lt;ext&gt;</code> and linked from the item with an <code>image</code> statement + a <code>license</code> statement.
* '''Official website''' (optional) — the collective's own website (a plain URL field). The value is written as an <code>official website</code> statement and renders in the '''Collective: page infobox'''.
* '''Logo''' (optional) — behind the ''"I will upload a logo image for this collective"'' toggle. '''Choose the source yourself''' — upload a local image file, paste an image URL, or '''reuse an existing file on this wiki''' (two collectives often share a logo, e.g. a website and its TV channel — pick the already-uploaded file instead of re-uploading). The reuse box '''searches existing files as you type''' — any fragment of the file name matches (''astro'' finds ''Astronomy and Astrophysics-logo.svg''; a thumbnail previews each suggestion; the same fragment search powers the portrait on [[Special:AddPerson]] and the logo on [[Special:AddSoftware]]). Nothing is pre-selected: the matching input appears after you pick a source. A pasted URL has a '''Validate''' button with preview + auto-fill, like the portrait. Fill the '''logo license''' (required when a logo is provided) and optionally '''Logo author''' / '''Additional license information'''. The file is saved as <code>File:&lt;name&gt;-logo.&lt;ext&gt;</code> and linked from the item with <code>image</code>, <code>license</code>, <code>image author</code> and <code>additional license information</code> statements; the logo renders in the '''Collective: page infobox''' (the infobox image cell reads the item's <code>image</code> statement — the logo shows on every page sitelinked to an item that has one, including pages created before the logo existed).


== Special:AddSource is class-first ==
== Special:AddSource is class-first ==
Line 64: Line 58:
|-
|-
| Web page || child of website || '''enter the URL''' — metadata fetched automatically
| Web page || child of website || '''enter the URL''' — metadata fetched automatically
|-
| Newspaper article · Magazine article || parent || search: title or author
|-
| Conference paper || parent || search: DOI, title, or author
|-
| Report · Document || parent || manual: title, publisher, report number
|-
| Thesis || parent || search: title or author
|-
| Manuscript || parent || manual: title, author
|-
| Patent || parent || manual: title, patent number
|-
| Legal case || parent || manual: court, territorial jurisdiction, case number
|-
| Legislation · Bill || parent || manual: jurisdiction, legislation number
|-
| Treaty || parent || manual: jurisdiction
|-
| Interview || parent || manual: title, author
|-
| Map || parent || manual: title, author, publisher
|-
| Presentation || parent || manual: title, author
|-
| Dataset || parent || manual: title, author, publisher
|-
| Text || parent || manual: title (catch-all — author, year, URL and access optional)
|}
|}


* '''The Zotero/CSL-aligned classes''' (newspaper/magazine article, conference paper, report, document, thesis, manuscript, patent, legal case, legislation, bill, treaty, interview, map, presentation, dataset) follow the same class-first flow with adapted fields. The '''legal texts''' (legal case, legislation, bill, treaty) carry one or more '''territorial jurisdictions''' — an OpenStreetMap search box like the person places: type a place name, pick the suggestion, and the OSM id + its display label are stored (several jurisdictions are allowed — separate them with a comma; each renders as a link in the <code>Source:</code> page infobox). For a text that applies internationally (a treaty signed by many countries, an international law), tick '''"This is an international …"''' instead — the checkbox '''replaces''' the jurisdiction (the box is hidden and cleared) and the infobox shows ''International''. A legal case also takes a '''court''' (entity search) and a '''case number''', a patent a '''patent number''', reports/documents a '''report number''', and legislation/bills a '''legislation number'''.
* '''Text''' is the '''catch-all class''' for any text that does not fit the others — a historical text, an inscription, a document of uncertain nature. Only the '''title''' is required (author(s), year, URL and access are optional).
* '''The entity search boxes are class-scoped''': typing in a license box suggests only license items, the author box only people/organizations/groups, the publisher box only publishers, and so on — not every item on the instance.
* '''Website / web page URL entry''': type a public web address (<code>http://</code>/<code>https://</code>) and the title, description and intro are fetched automatically when the site allows it — you correct everything on the next pages. For '''websites''' the URL is collapsed to the site root (a website is the site, not a page) and there is '''no year''' field (a website is dynamic); for '''web pages''' the full URL is kept.
* '''Website / web page URL entry''': type a public web address (<code>http://</code>/<code>https://</code>) and the title, description and intro are fetched automatically when the site allows it — you correct everything on the next pages. For '''websites''' the URL is collapsed to the site root (a website is the site, not a page) and there is '''no year''' field (a website is dynamic); for '''web pages''' the full URL is kept.
* '''Publisher is an entity''' (book, scholarly article): the publisher field is an entity search box — type a name to pick an existing publisher item (or type a Q-id). Create a missing publisher first (e.g. via [[Special:AddCollective]]). The item's publisher statement then points at that entity.
* '''Publisher is an entity''' (book, scholarly article, newspaper/magazine article, conference paper, report, document, thesis, interview, map, dataset): the publisher field is an entity search box — type a name to pick an existing publisher item (or type a Q-id). Create a missing publisher first (e.g. via [[Special:AddCollective]]). The item's publisher statement then points at that entity.
* '''Journal is an entity''' (scholarly article): the journal field is an entity search box, like the publisher — pick the journal item (create a missing one first, e.g. via [[Special:AddCollective]]).
* '''Journal is an entity''' (scholarly article): the journal field is an entity search box, like the publisher — pick the journal item (create a missing one first, e.g. via [[Special:AddCollective]]).
* '''Access''' (book, scholarly article, song, film, book excerpt): how this source can be accessed — a toggle with four modes:
* '''Access''' (book, scholarly article, song, film, book excerpt, and the new newspaper/magazine article, conference paper, report, document, thesis, manuscript, map and dataset): how this source can be accessed — a toggle with four modes:
   * '''Access URL (non-direct)''' — a link to a page where the work can be accessed (a landing page, not the file itself).
   * '''Access URL (non-direct)''' — a link to a page where the work can be accessed (a landing page, not the file itself).
   * '''Direct download link''' — a URL whose file is fetched and saved on this wiki automatically (as <code>File:&lt;title&gt;.&lt;ext&gt;</code>).
   * '''Direct download link''' — a URL whose file is fetched and saved on this wiki automatically (as <code>File:&lt;title&gt;.&lt;ext&gt;</code>).
Line 76: Line 101:
   The download/file modes expand a '''license''' field (pick an existing license item — CC0, CC BY, GNU, etc., or type to search). Uploaded files are named after the item (your original filename is ignored) and you should not upload copyrighted content unless you have the right to distribute it publicly.
   The download/file modes expand a '''license''' field (pick an existing license item — CC0, CC BY, GNU, etc., or type to search). Uploaded files are named after the item (your original filename is ignored) and you should not upload copyrighted content unless you have the right to distribute it publicly.
* '''The Access row on the Source: page''' — the infobox of every <code>Source:</code> page shows an '''Access''' row: a stored copy renders as the file name (click it to open '''[[Special:SourceFile]]''' — an embedded PDF preview when the file is a PDF, the licence, and a '''download button that requires accepting the licence conditions'''), a non-direct access URL renders as a clickable link, and a source without access shows '''N/A'''.
* '''The Access row on the Source: page''' — the infobox of every <code>Source:</code> page shows an '''Access''' row: a stored copy renders as the file name (click it to open '''[[Special:SourceFile]]''' — an embedded PDF preview when the file is a PDF, the licence, and a '''download button that requires accepting the licence conditions'''), a non-direct access URL renders as a clickable link, and a source without access shows '''N/A'''.
* '''Child classes need a parent''': for a book excerpt, YouTube video or web page you first pick the parent item (book, YouTube channel, website) with an entity search box. The line under the box — ''"Book not yet imported to RonzzWikiBase? import it yourself"'' — jumps to the parent's import page. On creation the child is linked to the parent with a <code>part of</code> statement.
* '''Child classes need a parent''': for a book excerpt, YouTube video or web page you first pick the parent item (book, YouTube channel, website) with an entity search box — the line under the box, ''"Book not yet imported to RonzzWikiBase? import it yourself"'', jumps to the parent's import page; on creation the child is linked to the parent with a <code>part of</code> statement. For a '''web page''' the parent is '''auto-assigned from the URL's site root''': the root URL's host is matched against the website items' stored URL addresses — a page on a '''subdomain''' of a recorded website matches its parent site too (a page on ''scifa.univ-lorraine.fr'' belongs to the ''univ-lorraine.fr'' website; a page on the recorded host itself still matches first). An '''exact host (or ancestor) match silently fills the parent box''' (you can still change it before submitting; no confirmation needed); a '''close name match''' fills it with the ''"we think this corresponds to …"'' confirmation ('''Yes, that's right''' / '''No, let me correct'''); and no match leaves the box empty with ''"No record found for {URL} — the website exists, but we don't have a record of it yet. Add the website first, then pick it above."'' (add the website via [[Special:AddSource/website]]).
* '''Book excerpts''': after picking the parent book, fill the '''pages''' and optionally the '''volume''' and '''chapters'''. A blank '''description''' auto-generates as ''Pages a-b (Volume c) of {book}'' from the pages/volume fields and the book's title; blank '''year''' and '''author(s)''' fall back to the parent book's own values (''Leave blank if same as the parent book'').
* '''Book excerpts''': after picking the parent book, fill the '''pages''' and optionally the '''volume''' and '''chapters'''. A blank '''description''' auto-generates as ''Pages a-b (Volume c) of {book}'' from the pages/volume fields and the book's title; blank '''year''' and '''author(s)''' fall back to the parent book's own values (''Leave blank if same as the parent book'').
* '''Website vs web page''': add an '''entire website''' once, then add specific pages as child items (see the explanation on [[Special:AddSource/website]]).
* '''Website vs web page''': add an '''entire website''' once, then add specific pages as child items (see the explanation on [[Special:AddSource/website]]).
* '''Duration''' (film, song, video, YouTube video): type <syntaxhighlight lang="text" inline>MM:SS</syntaxhighlight> or <syntaxhighlight lang="text" inline>HH:MM:SS</syntaxhighlight> — stored in seconds.
* '''Duration''' (film, song, video, YouTube video): type <syntaxhighlight lang="text" inline>MM:SS</syntaxhighlight> or <syntaxhighlight lang="text" inline>HH:MM:SS</syntaxhighlight> — stored in seconds.
* '''A source needs at least one author''' — except a book excerpt that inherits its authors from the parent book: the review step requires picking author '''entities''' (person, organization or group — type a name to search, or type a Q-id). Create missing people first via [[Special:AddPerson]]. This is enforced at creation.
* '''Title is the only required field''' — every other fact (author(s), year, publisher, …) is optional, because it can genuinely be unknown (a text of unknown authorship, an anonymous work). Authors are picked as '''entities''' (person, organization or group — type a name to search, or type a Q-id); create missing people first via [[Special:AddPerson]]. The '''legal texts''' (legal case, legislation, bill, treaty) take no author (the court / jurisdiction carry the attribution); a book excerpt inherits its authors from the parent book when left blank. Child classes (book excerpt, YouTube video, web page) still require their parent.
 
== Uploading files (Special:Upload) ==
 
[[Special:Upload]] is the general file-upload page (also reachable from the '''Upload file''' link in the sidebar). It saves the file to a <code>File:</code> page '''and''' — for every form upload — creates (or reuses) a '''linked image item''' holding the file's structured facts, so uploaded files are first-class, queryable entities (the ''facts in the item, prose on the page'' model).
 
* '''Source''' — the radio defaults to '''URL''' (paste a link — the common case); switch to '''file from your device''' when the image is only on your computer.
* '''File name + description''' — filled in by hand, or auto-filled: for a pasted URL, click '''Validate''' next to the URL field — a preview thumbnail, the pixel + file size and a best-effort auto-fill of the name / description / author / licence fields appear (clicking Validate again shows only the latest result). '''Wikimedia image URLs''' (Wikipedia, Commons, upload.wikimedia.org) are read '''from your browser''' — they are not rate-limited, and the image bytes are downloaded by your browser too (the server never fetches Wikimedia). For files larger than 50 MB from Wikimedia, save the image to your device and upload it instead.
* '''License''' — a '''searchable combobox of the instance's license items''' (CC0, CC BY 4.0, CC BY-SA 4.0, GNU licenses, MIT, … — pick from the list or type to search). The license is recorded as a semantic statement on the image item and renders on the File page as a link to the license item.
* '''Image author''' (optional) — the person or organization that created the image, as free text.
* '''Additional license information''' (optional) — free text (attribution lines, terms, credits — e.g. what a CC BY license requires).
* The single ''Maximum file size: 1 GB'' note applies to both upload methods.
* The File page carries an '''Attribution''' section with the author / license / source; the image item holds the same facts as statements (<code>image</code>, <code>license</code>, <code>image author</code>, <code>additional license information</code>, source URL) and is sitelinked to the File page.
 
 
== Re-editing an imported item ==


== Fetched page content ==
Imported items can be edited in place with the '''same form that created them''' — no need to re-import or hand-edit statements. On the item page (<code>/wiki/Item:Q1</code>), the '''Update basic information''' button under the title opens the matching page (<code>Special:UpdatePerson</code>, <code>Special:UpdateSource</code>, <code>Special:UpdateCollective</code>, <code>Special:UpdateSoftware</code>, <code>Special:UpdateFictionalCharacter</code> — chosen from the item's class), prefilled with the item's current values. Change what needs changing and submit: the label, description and the form's statements update; everything else (sitelinks, other statements, uploaded portraits/logos you did not replace) stays. A changed label renames the classic page too. See [[Help:Contributing/entities#Updating basic information with a form]].


When the authority has page prose, it is fetched automatically and shown on a dedicated '''content review''' step after the record review — confirm, correct or clear each block; only what you keep is written to the page. Each block carries an attribution line (''from Wikipedia (CC BY-SA 4.0):'', ''from OpenAlex:'', …).
* '''Replacing an image''' — the upload toggle on the Update pages reads ''"I will upload a '''new''' portrait/logo image … (replacing existing)"'': tick it and pick a source to replace the current image; leave it unticked to keep it.
* '''The access section on Special:UpdateSource is collapsed''' behind the ''"I will update the resource access instructions"'' checkbox — tick it to see (and change) the access URL / download link / local file / licence fields; leave it unticked and the item's access statements stay exactly as they are (no re-upload, no replacement).
* '''Fields you leave blank are kept as they are''' — the update form never overwrites a value you did not provide (a blank description keeps the existing one). Removing a statement is an explicit edit on the item page itself.


* '''Scholarly article''' — '''Abstract''' + '''Key words''' (OpenAlex, Crossref as fallback).
When a field is auto-filled from fetched data, the wiki asks you to confirm: ''"{field} fetched from source: {value}, we think this corresponds to {label} (Q#)."'' — '''Yes, that's right''' keeps it, '''No, let me correct''' clears the field for you to pick another item. This happens for:
* '''Book''' — '''Summary''' (Wikipedia lead).
* '''Song''' — '''Overview''' (Wikipedia lead) + '''Lyrics''' (the Wikipedia article's lyrics section, when present).
* '''Film''' — '''Overview''' (Wikipedia lead) + '''Plot'''.
* '''Person''' — '''Biography''' (Wikipedia lead) on the <code>Person:</code> page.
* '''Website / web page''' — the intro/summary fetched from the site itself.


The classic pages render sections '''only when they have content''' (no empty placeholders; no ''See also''); websites and YouTube channels show a short intro without headings.
* the '''license''' on Special:Upload and the Add\* portrait/logo URL fields (from the image's metadata);
* the '''publisher''' and '''journal''' on the book / scholarly-article review (a harvested name that matches — exactly or closely — an existing item is prefilled and confirmed);
* the '''place of birth''' / '''place of death''' on the AddPerson review (a harvested place label matched against local items; no match → the "External record: …" hint);
* the '''parent website''' on the web-page entry (the site root of the entered URL matched against website items; no match → the "No record found for {URL} — add the website first" hint);
* the harvested '''developer / license / programming language''' facts on AddSoftware;
* a '''free-text author name''' in the AddSource search: if it matches an existing person, the manual form is prefilled with that person and confirmed (previously a free-text name was only a search filter).


== Workflow ==
== Workflow ==


For classes with an authority: search → review the candidates (same-name disambiguation) → review the record → [review the fetched content, when present] → create. Each candidate carries a ''see record details'' link to its canonical authority page (opens in a new tab). Repeating a lookup for an already-imported label reuses the existing item (create-or-skip). The English label comes from the authority; add <code>fr</code>/<code>eo</code> labels afterwards. Fetch failures degrade to manual entry at <code>Special:AddSource/&lt;class&gt;/manual</code>.
For classes with an authority: search → review the candidates (same-name disambiguation) → review the record → [review the fetched content, when present] → create. Each candidate carries a ''see record details'' link to its canonical authority page (opens in a new tab). The English label comes from the authority; add <code>fr</code>/<code>eo</code> labels afterwards. '''Source labels carry the source type in parentheses''' for disambiguation — e.g. ''The Hobbit (Book)'', ''Example Domain (Website)'', ''Chapter 3 (Book excerpt)'' — the suffix is added automatically (and editable) on the AddSource review/manual forms, and the <code>Source:</code> page title follows the label. While you type the title on those forms, a read-only '''Final label''' box next to it shows exactly what the item label will be (title + suffix) — no surprises at creation. Fetch failures degrade to manual entry at <code>Special:AddSource/&lt;class&gt;/manual</code>.
 
=== Duplicates are confirmed, not silently reused ===
 
When the record you are about to create matches an '''existing item''' — the same authority identifier (Wikidata, OpenAlex, ORCID, DOI, ISBN, VIAF, ISNI, YouTube id), the same web URL (official website, repository, documentation or access URL), or a '''very similar label''' — the wiki asks first:
 
: ''"We think this item may be a duplicate of [[Item:Qxxx|label]]."'' — '''Yes, that's right''' takes you straight to the existing item (nothing is created); '''No''' continues the creation (the final confirmation creates the item anyway, even with the same label — the existing item's classic page keeps rendering the original).
 
The check runs as soon as the duplicate is recognizable: when you '''pick a search result''' whose authority id already exists locally, when you '''enter an existing website / web-page / YouTube URL''', and on '''every final submit''' (the last check re-examines the record you actually submit). A machine client (the API) never creates silently either: it answers <code>duplicate</code> with the existing item's id instead, unless it explicitly confirms.


Notes:
Notes:
Line 103: Line 153:
* '''Author search''' on the book / scholarly-article / film / song / video pages narrows the results for common titles. Toggle between '''free text''' (an author name, e.g. Albert Einstein) and '''semantic entities''' (Wikidata Q-ids, comma-separated — e.g. Q937, Q42). The semantic mode resolves works by the author entities on the Wikidata hub.
* '''Author search''' on the book / scholarly-article / film / song / video pages narrows the results for common titles. Toggle between '''free text''' (an author name, e.g. Albert Einstein) and '''semantic entities''' (Wikidata Q-ids, comma-separated — e.g. Q937, Q42). The semantic mode resolves works by the author entities on the Wikidata hub.
* '''YouTube''': paste a URL for an exact match (no match → ''no match for the provided URL''), or type a name for the top 10 matches.
* '''YouTube''': paste a URL for an exact match (no match → ''no match for the provided URL''), or type a name for the top 10 matches.
* The provenance form on [[Special:AddQuotation]] (and the other content pages) has entity '''search + autofill''' for <code>attributed to</code> and <code>source</code>: type a name, pick the entity, the item id is filled in. Manually typed ids still work.
* The provenance form on [[Special:AddQuotation]] (and the other content pages) has entity '''search + autofill''' for <code>attributed to</code> and <code>source</code>: type a name, pick the entity, the item id is filled in. Manually typed ids still work. The '''label field is prefilled''' with the class marker — ''(quotation)'' / ''(code snippet)'' / ''(math snippet)'' — type the content text in front of it, like the ''The Hobbit (Book)'' source-label convention. The '''Add more''' button creates the item and reopens the page with everything you typed carried over '''except the label and the content''' (the label resets to the marker, the content to empty) — adding several items from the same source or author means typing only the new content each time. Only signed-in users can create content items.
* Citations use the source item's harvested metadata (journal name, publisher, volume/issue/pages, DOI, ISBN) and cite the source '''type''' correctly (a book quote cites as a book, a song quote as a song).
* Citations use the source item's harvested metadata (journal name, publisher, volume/issue/pages, DOI, ISBN) and cite the source '''type''' correctly (a book quote cites as a book, a song quote as a song).


== See also ==
* the '''place of birth''' / '''place of death''' on the AddPerson review (a harvested place label is auto-looked-up on OpenStreetMap; a match prefills the OSM id with confirmation, no match → the "search OpenStreetMap to confirm" hint). The picked place's '''display label is stored with the person''' and shown — linked to the OpenStreetMap map — on the <code>Person:</code> page instead of the raw OSM id. A person who already has places added before this feature gets their labels backfilled automatically.
* the '''parent website''' on the web-page entry (the root URL's host matched against website items — an exact host match auto-assigns silently, a close name match confirms, no match → the "No record found for {URL} — add the website first" hint);
 
== When the classic page already exists ==
 
Every creation flow ([[Special:AddPerson]] / [[Special:AddSource]] / [[Special:AddCollective]] / [[Special:AddSoftware]] / [[Special:AddFictionalCharacter]]) writes a classic wiki page for the new item (<code>Person:</code> / <code>Source:</code> / <code>Collective:</code> / <code>FOSS:</code> / <code>Software:</code>). If a page with that exact name already exists — an editor may have created it by hand — the flow does '''not''' overwrite it and does '''not''' link it silently:
 
* you are asked to confirm: ''"A page named X already exists. Is it the same thing as the item you just created?"''
* '''Yes, link the item to this page''' links the new item to the existing page (the page itself is left untouched);
* '''No, don't link it''' leaves the item unlinked — you can link it later with the '''Sitelink''' tab.


* [[Help:Contributing]] — the hub
The same rule applies to items created at [[Special:NewItem]]: a main-namespace page titled with the label is created and linked automatically; if that page already exists, the item is linked to it (no confirmation there — the hook has no form, the page is never modified, and the Sitelink tab lets you correct it). See [[Help:Contributing/entities#Creating an item]].
* [[Help:Contributing/entities]] — creating and editing items and properties
* [[Help:Contributing/semanticDynamicContent]] — page↔item sitelinks (incl. the Sitelink tab)

Latest revision as of 20:13, 22 September 2026

Instead of hand-creating an item per author or source, use the external-authority pages in the semantic tools sidebar.

They fetch metadata from public services (Wikidata, dblp, OpenAlex, Crossref, Open Library, ORCID, YouTube, Wikipedia) and let you pick the right record. They create a local stub item with the authority identifiers (ORCID, VIAF, ISNI, DOI, ISBN, OpenAlex, YouTube IDs, …) and citation metadata (given/family name, journal, publisher as an entity, pages, volume, issue) — every imported statement carries an imported from <authority> reference. Each created item also gets a classic wiki page (Person:, Source: or Collective:) sitelinked to it, whose infobox renders the item's statements at view time — except book excerpts, which are part of a book and get only an item. Where the authority has page prose (an abstract, a plot, lyrics, a lead intro), it is fetched and offered for review before it is written to the classic page (see Help:Contributing/import#Fetched page content).

Page Creates Input
Special:AddPerson a person item + a Person: page name, ORCID iD, VIAF ID, ISNI, or a Wikidata Q id
Special:AddSource a source item + a Source: page — see the class list below class first, then the class-specific input
Special:AddFictionalCharacter a fictional-character item (no classic page) a name (Wikidata)
Special:AddCollective a non-person agent item (organization, company, band, agency, institution, team…) + a Collective: page a name

Every import flow offers a manual fallback — Create the item manually instead — on the search page and on the candidate-selection page. The manual form is prefilled with what you typed in the search boxes (a person name is split into given/family, e.g. "Marie Curie" → given Marie, family Curie), so you correct instead of retyping. When the form is prefilled from fetched data (the website/web page URL entry), the form is titled Item details — you are verifying fetched content, not entering an item from blank.


  • Place of birth / Place of death — OpenStreetMap search boxes: type a place name and pick the place from the suggestions (searched on OpenStreetMap's Nominatim — villages, cities and regions alike, no local place items needed). Picking a suggestion stores the place's OpenStreetMap id (node/123456, way/… or relation/…) as an place of birth (OSM) / place of death (OSM) statement; the Person: page infobox renders it as a link to openstreetmap.org. A place fetched from the authority is automatically looked up on OpenStreetMap and the best match is proposed: "{field} fetched from source: {place}, we think this corresponds to {name} ({id})." — Yes, that's right keeps it, No, let me correct clears the box for you to search. No match shows "External record: {place} — search OpenStreetMap to confirm the place, then pick it from the suggestions." — the field must hold a picked OpenStreetMap id to save (a raw place name alone is rejected).

Fictional characters

Special:AddFictionalCharacter searches Wikidata by name (search → review candidates → create). The review step has Given name / Family name and Appears in (entity search, several works allowed). The label auto-generates as {given name} {family name} (fictional character); the description auto-generates as fictional character in {works} when left blank. Characters get no classic page — just the item.


Collectives: common classes

The collective class is inferred from the harvested Wikidata instance-of hints and chosen on the review step (the search-results selection step no longer shows a class field — pick the record first, then classify it when the harvested data is in front of you): private company, public company, non-profit organization, governmental agency, music band, educational institution, research institute, political party, trade union, religious organization, sports team — plus the generic organization / group of humans.

The collective review step also offers:

  • Parent organization (optional) — an entity search box: the organization this collective is a subsidiary or subdivision of (type a name to pick an existing item, or type a Q-id).
  • Official website (optional) — the collective's own website (a plain URL field). The value is written as an official website statement and renders in the Collective: page infobox.
  • Logo (optional) — behind the "I will upload a logo image for this collective" toggle. Choose the source yourself — upload a local image file, paste an image URL, or reuse an existing file on this wiki (two collectives often share a logo, e.g. a website and its TV channel — pick the already-uploaded file instead of re-uploading). The reuse box searches existing files as you type — any fragment of the file name matches (astro finds Astronomy and Astrophysics-logo.svg; a thumbnail previews each suggestion; the same fragment search powers the portrait on Special:AddPerson and the logo on Special:AddSoftware). Nothing is pre-selected: the matching input appears after you pick a source. A pasted URL has a Validate button with preview + auto-fill, like the portrait. Fill the logo license (required when a logo is provided) and optionally Logo author / Additional license information. The file is saved as File:<name>-logo.<ext> and linked from the item with image, license, image author and additional license information statements; the logo renders in the Collective: page infobox (the infobox image cell reads the item's image statement — the logo shows on every page sitelinked to an item that has one, including pages created before the logo existed).

Special:AddSource is class-first

Special:AddSource opens on a class picker — choose the source type (the page title is Special:AddSource); the page then shows the adapted first step (search, or a URL entry for websites/web pages) and the adapted review fields for that class.

Class Kind First step
Book parent search: ISBN-13, title, or author
Scholarly article parent search: DOI, title, or author
Website parent enter the URL — metadata (title, description, intro) is fetched automatically
Song · Film · Video parent search: title or author (Wikidata)
YouTube channel parent search: YouTube name or URL (exact match for a URL)
YouTube video child of YouTube channel search: YouTube name or URL
Book excerpt child of book manual: parent book + pages
Web page child of website enter the URL — metadata fetched automatically
Newspaper article · Magazine article parent search: title or author
Conference paper parent search: DOI, title, or author
Report · Document parent manual: title, publisher, report number
Thesis parent search: title or author
Manuscript parent manual: title, author
Patent parent manual: title, patent number
Legal case parent manual: court, territorial jurisdiction, case number
Legislation · Bill parent manual: jurisdiction, legislation number
Treaty parent manual: jurisdiction
Interview parent manual: title, author
Map parent manual: title, author, publisher
Presentation parent manual: title, author
Dataset parent manual: title, author, publisher
Text parent manual: title (catch-all — author, year, URL and access optional)
  • The Zotero/CSL-aligned classes (newspaper/magazine article, conference paper, report, document, thesis, manuscript, patent, legal case, legislation, bill, treaty, interview, map, presentation, dataset) follow the same class-first flow with adapted fields. The legal texts (legal case, legislation, bill, treaty) carry one or more territorial jurisdictions — an OpenStreetMap search box like the person places: type a place name, pick the suggestion, and the OSM id + its display label are stored (several jurisdictions are allowed — separate them with a comma; each renders as a link in the Source: page infobox). For a text that applies internationally (a treaty signed by many countries, an international law), tick "This is an international …" instead — the checkbox replaces the jurisdiction (the box is hidden and cleared) and the infobox shows International. A legal case also takes a court (entity search) and a case number, a patent a patent number, reports/documents a report number, and legislation/bills a legislation number.
  • Text is the catch-all class for any text that does not fit the others — a historical text, an inscription, a document of uncertain nature. Only the title is required (author(s), year, URL and access are optional).
  • The entity search boxes are class-scoped: typing in a license box suggests only license items, the author box only people/organizations/groups, the publisher box only publishers, and so on — not every item on the instance.
  • Website / web page URL entry: type a public web address (http:///https://) and the title, description and intro are fetched automatically when the site allows it — you correct everything on the next pages. For websites the URL is collapsed to the site root (a website is the site, not a page) and there is no year field (a website is dynamic); for web pages the full URL is kept.
  • Publisher is an entity (book, scholarly article, newspaper/magazine article, conference paper, report, document, thesis, interview, map, dataset): the publisher field is an entity search box — type a name to pick an existing publisher item (or type a Q-id). Create a missing publisher first (e.g. via Special:AddCollective). The item's publisher statement then points at that entity.
  • Journal is an entity (scholarly article): the journal field is an entity search box, like the publisher — pick the journal item (create a missing one first, e.g. via Special:AddCollective).
  • Access (book, scholarly article, song, film, book excerpt, and the new newspaper/magazine article, conference paper, report, document, thesis, manuscript, map and dataset): how this source can be accessed — a toggle with four modes:
 * Access URL (non-direct) — a link to a page where the work can be accessed (a landing page, not the file itself).
 * Direct download link — a URL whose file is fetched and saved on this wiki automatically (as File:<title>.<ext>).
 * Local file — pick a file on your computer to upload.
 * N/A (no direct access) — the work is only reachable via archives or physical copies; no access statement is written.
 The download/file modes expand a license field (pick an existing license item — CC0, CC BY, GNU, etc., or type to search). Uploaded files are named after the item (your original filename is ignored) and you should not upload copyrighted content unless you have the right to distribute it publicly.
  • The Access row on the Source: page — the infobox of every Source: page shows an Access row: a stored copy renders as the file name (click it to open Special:SourceFile — an embedded PDF preview when the file is a PDF, the licence, and a download button that requires accepting the licence conditions), a non-direct access URL renders as a clickable link, and a source without access shows N/A.
  • Child classes need a parent: for a book excerpt, YouTube video or web page you first pick the parent item (book, YouTube channel, website) with an entity search box — the line under the box, "Book not yet imported to RonzzWikiBase? import it yourself", jumps to the parent's import page; on creation the child is linked to the parent with a part of statement. For a web page the parent is auto-assigned from the URL's site root: the root URL's host is matched against the website items' stored URL addresses — a page on a subdomain of a recorded website matches its parent site too (a page on scifa.univ-lorraine.fr belongs to the univ-lorraine.fr website; a page on the recorded host itself still matches first). An exact host (or ancestor) match silently fills the parent box (you can still change it before submitting; no confirmation needed); a close name match fills it with the "we think this corresponds to …" confirmation (Yes, that's right / No, let me correct); and no match leaves the box empty with "No record found for {URL} — the website exists, but we don't have a record of it yet. Add the website first, then pick it above." (add the website via Special:AddSource/website).
  • Book excerpts: after picking the parent book, fill the pages and optionally the volume and chapters. A blank description auto-generates as Pages a-b (Volume c) of {book} from the pages/volume fields and the book's title; blank year and author(s) fall back to the parent book's own values (Leave blank if same as the parent book).
  • Website vs web page: add an entire website once, then add specific pages as child items (see the explanation on Special:AddSource/website).
  • Duration (film, song, video, YouTube video): type MM:SS or HH:MM:SS — stored in seconds.
  • Title is the only required field — every other fact (author(s), year, publisher, …) is optional, because it can genuinely be unknown (a text of unknown authorship, an anonymous work). Authors are picked as entities (person, organization or group — type a name to search, or type a Q-id); create missing people first via Special:AddPerson. The legal texts (legal case, legislation, bill, treaty) take no author (the court / jurisdiction carry the attribution); a book excerpt inherits its authors from the parent book when left blank. Child classes (book excerpt, YouTube video, web page) still require their parent.

Uploading files (Special:Upload)

Special:Upload is the general file-upload page (also reachable from the Upload file link in the sidebar). It saves the file to a File: page and — for every form upload — creates (or reuses) a linked image item holding the file's structured facts, so uploaded files are first-class, queryable entities (the facts in the item, prose on the page model).

  • Source — the radio defaults to URL (paste a link — the common case); switch to file from your device when the image is only on your computer.
  • File name + description — filled in by hand, or auto-filled: for a pasted URL, click Validate next to the URL field — a preview thumbnail, the pixel + file size and a best-effort auto-fill of the name / description / author / licence fields appear (clicking Validate again shows only the latest result). Wikimedia image URLs (Wikipedia, Commons, upload.wikimedia.org) are read from your browser — they are not rate-limited, and the image bytes are downloaded by your browser too (the server never fetches Wikimedia). For files larger than 50 MB from Wikimedia, save the image to your device and upload it instead.
  • License — a searchable combobox of the instance's license items (CC0, CC BY 4.0, CC BY-SA 4.0, GNU licenses, MIT, … — pick from the list or type to search). The license is recorded as a semantic statement on the image item and renders on the File page as a link to the license item.
  • Image author (optional) — the person or organization that created the image, as free text.
  • Additional license information (optional) — free text (attribution lines, terms, credits — e.g. what a CC BY license requires).
  • The single Maximum file size: 1 GB note applies to both upload methods.
  • The File page carries an Attribution section with the author / license / source; the image item holds the same facts as statements (image, license, image author, additional license information, source URL) and is sitelinked to the File page.


Re-editing an imported item

Imported items can be edited in place with the same form that created them — no need to re-import or hand-edit statements. On the item page (/wiki/Item:Q1), the Update basic information button under the title opens the matching page (Special:UpdatePerson, Special:UpdateSource, Special:UpdateCollective, Special:UpdateSoftware, Special:UpdateFictionalCharacter — chosen from the item's class), prefilled with the item's current values. Change what needs changing and submit: the label, description and the form's statements update; everything else (sitelinks, other statements, uploaded portraits/logos you did not replace) stays. A changed label renames the classic page too. See Help:Contributing/entities#Updating basic information with a form.

  • Replacing an image — the upload toggle on the Update pages reads "I will upload a new portrait/logo image … (replacing existing)": tick it and pick a source to replace the current image; leave it unticked to keep it.
  • The access section on Special:UpdateSource is collapsed behind the "I will update the resource access instructions" checkbox — tick it to see (and change) the access URL / download link / local file / licence fields; leave it unticked and the item's access statements stay exactly as they are (no re-upload, no replacement).
  • Fields you leave blank are kept as they are — the update form never overwrites a value you did not provide (a blank description keeps the existing one). Removing a statement is an explicit edit on the item page itself.

When a field is auto-filled from fetched data, the wiki asks you to confirm: "{field} fetched from source: {value}, we think this corresponds to {label} (Q#)." — Yes, that's right keeps it, No, let me correct clears the field for you to pick another item. This happens for:

  • the license on Special:Upload and the Add\* portrait/logo URL fields (from the image's metadata);
  • the publisher and journal on the book / scholarly-article review (a harvested name that matches — exactly or closely — an existing item is prefilled and confirmed);
  • the place of birth / place of death on the AddPerson review (a harvested place label matched against local items; no match → the "External record: …" hint);
  • the parent website on the web-page entry (the site root of the entered URL matched against website items; no match → the "No record found for {URL} — add the website first" hint);
  • the harvested developer / license / programming language facts on AddSoftware;
  • a free-text author name in the AddSource search: if it matches an existing person, the manual form is prefilled with that person and confirmed (previously a free-text name was only a search filter).

Workflow

For classes with an authority: search → review the candidates (same-name disambiguation) → review the record → [review the fetched content, when present] → create. Each candidate carries a see record details link to its canonical authority page (opens in a new tab). The English label comes from the authority; add fr/eo labels afterwards. Source labels carry the source type in parentheses for disambiguation — e.g. The Hobbit (Book), Example Domain (Website), Chapter 3 (Book excerpt) — the suffix is added automatically (and editable) on the AddSource review/manual forms, and the Source: page title follows the label. While you type the title on those forms, a read-only Final label box next to it shows exactly what the item label will be (title + suffix) — no surprises at creation. Fetch failures degrade to manual entry at Special:AddSource/<class>/manual.

Duplicates are confirmed, not silently reused

When the record you are about to create matches an existing item — the same authority identifier (Wikidata, OpenAlex, ORCID, DOI, ISBN, VIAF, ISNI, YouTube id), the same web URL (official website, repository, documentation or access URL), or a very similar label — the wiki asks first:

"We think this item may be a duplicate of label." — Yes, that's right takes you straight to the existing item (nothing is created); No continues the creation (the final confirmation creates the item anyway, even with the same label — the existing item's classic page keeps rendering the original).

The check runs as soon as the duplicate is recognizable: when you pick a search result whose authority id already exists locally, when you enter an existing website / web-page / YouTube URL, and on every final submit (the last check re-examines the record you actually submit). A machine client (the API) never creates silently either: it answers duplicate with the existing item's id instead, unless it explicitly confirms.

Notes:

  • The authority lookups run only when you click search — never on page load.
  • Author search on the book / scholarly-article / film / song / video pages narrows the results for common titles. Toggle between free text (an author name, e.g. Albert Einstein) and semantic entities (Wikidata Q-ids, comma-separated — e.g. Q937, Q42). The semantic mode resolves works by the author entities on the Wikidata hub.
  • YouTube: paste a URL for an exact match (no match → no match for the provided URL), or type a name for the top 10 matches.
  • The provenance form on Special:AddQuotation (and the other content pages) has entity search + autofill for attributed to and source: type a name, pick the entity, the item id is filled in. Manually typed ids still work. The label field is prefilled with the class marker — (quotation) / (code snippet) / (math snippet) — type the content text in front of it, like the The Hobbit (Book) source-label convention. The Add more button creates the item and reopens the page with everything you typed carried over except the label and the content (the label resets to the marker, the content to empty) — adding several items from the same source or author means typing only the new content each time. Only signed-in users can create content items.
  • Citations use the source item's harvested metadata (journal name, publisher, volume/issue/pages, DOI, ISBN) and cite the source type correctly (a book quote cites as a book, a song quote as a song).
  • the place of birth / place of death on the AddPerson review (a harvested place label is auto-looked-up on OpenStreetMap; a match prefills the OSM id with confirmation, no match → the "search OpenStreetMap to confirm" hint). The picked place's display label is stored with the person and shown — linked to the OpenStreetMap map — on the Person: page instead of the raw OSM id. A person who already has places added before this feature gets their labels backfilled automatically.
  • the parent website on the web-page entry (the root URL's host matched against website items — an exact host match auto-assigns silently, a close name match confirms, no match → the "No record found for {URL} — add the website first" hint);

When the classic page already exists

Every creation flow (Special:AddPerson / Special:AddSource / Special:AddCollective / Special:AddSoftware / Special:AddFictionalCharacter) writes a classic wiki page for the new item (Person: / Source: / Collective: / FOSS: / Software:). If a page with that exact name already exists — an editor may have created it by hand — the flow does not overwrite it and does not link it silently:

  • you are asked to confirm: "A page named X already exists. Is it the same thing as the item you just created?"
  • Yes, link the item to this page links the new item to the existing page (the page itself is left untouched);
  • No, don't link it leaves the item unlinked — you can link it later with the Sitelink tab.

The same rule applies to items created at Special:NewItem: a main-namespace page titled with the label is created and linked automatically; if that page already exists, the item is linked to it (no confirmation there — the hook has no form, the page is never modified, and the Sitelink tab lets you correct it). See Help:Contributing/entities#Creating an item.

Retrieved from "https://wikibase.ronzz.org/index.php?title=Help:Contributing/import&oldid=7826"