User:Magnus Manske/Author strings

This is a documentation page for the author strings gadget.

Many Wikidata items about scholarly article (Q13442814) have author name string (P2093) statements, which should be converted to author (P50) statements. This gadget aims to help with this process.

The gadget runs on two types of items: scholarly article (Q13442814) items, and items about people with certain properties (eg ORCID iD (P496)) indicating a scientific author item.

Article items

The gadget will gather the names from the author name string (P2093) statements for an article item. It will keep the last name, and (up to) the first three. It will internally "simplify" (normalize) these names, eg by removing initials. Then, it will search Wikidata for other articles, using every permutation of each two author names. It will load the resulting article items for all these searches (up to 45, ranked by how often an item was found), and normailze&group all author name string (P2093) statements in these items. Each "author group" will be shown as a block in a list. For each block, you can then (un)select occurrences of the name in a paper, and change all the selected author name string (P2093) statements into author (P50). This will be done in all the checked papers. References and qualifiers will be preserved, and a subject named as (P1810) qualifier will be added to the new author statement. You can either specify an existing author item ID, or create one; the statement change will occur automatically after item creation.

Important: Please use the "Search for author" function before creating a new author item, in order to avoid duplicates!

Author items

On author items, the gadget will search for scholarly article (Q13442814) items containing the (normalized) author name. It will then load the article items and proceed with normalizing&grouping as described above.

Author/author name string duplicates

If a section with this heading shows up, it means there are statements for the same "author serial number" for both author (P50) and author name string (P2093). This is usually a leftover from some bot migration gone wrong. Make sure only the same author pairs (not necessarily the exact name but similar) are selected, and click "Remove selected statements". This will remove the obsolete author name string (P2093) statements.