The Signpost

Technology report

View it! A new tool for image discovery

View it! Tool
The depths of Commons, at your fingertips. Or eyetips.

View it! is a new tool to access all the images related to articles on Wikipedia. It was created to show Wikipedians relevant Commons media depicting, or otherwise related to, the article they are looking at.

The goal of View it! is to enrich content pages by illustrating a given subject, and aid Wikipedia editors by surfacing images that can be used in articles. This tool allows readers access to the breadth of images available on Commons, beyond the limited images (if any) curated for inclusion in the article.

How to use

View it! tab
The tab.
View it! results for goat.
The results for goat
Full View it! Launch presentation

The primary way to use View it! is a user script, which can be installed using the instructions on Meta.

View it! has two different versions which can be installed, depending on the type of experience you would like. With the Lite version, the user script simply adds a "View" tab to the page, which will link you to the image results from our Toolforge project. The Full version will actually display the images on the Wikipedia page, with a “View” tab that opens a full-screen gallery. Both of these versions ultimately lead to the view-it.toolforge.org page, where more advanced features can also be accessed.

In the Lite version, which is now mostly stable, the "View" tab will have the number of results (if any) in parentheses. Clicking the button will take users in a new window to Toolforge and the results. In the Full version, which is actively being developed, users will see an image carousel across the top of articles. The carousel can be expanded, while clicking the tab opens a dedicated gallery page. Eventually, we plan for the carousel to be accessible from the edit mode, so that editors can use this tool to click to select images to insert while editing.

How it works

Free text search results.
P170 Q23548 with free text search

View it! uses the Commons API to search for images based on properties associated with the article. The default configuration displays images which are either in an article’s associated Commons category (as listed in its Wikidata item) or where the article matches the value of a Commons image's Wikidata "depicts" statement. When searching directly on Toolforge, you can search any Q-number. And with the advanced search, you can view results with other properties: "Main subject" (P921), "Created by" (P170), and "Commons Category" (P373). You can also add freeform text to the search, which comes in handy when looking for images of buildings or neighborhoods in a particular city, specific images from a creator (like images of the moon from NASA), or similar situations. You can also adjust the results for specific assessments and resolutions. When an image is clicked, you're taken directly to the file on Commons, where you can copy code to add the image to the article. The tool runs on a View it! API we built, which is the same one that powers the user scripts.

NEW: You can now add an image directly from the View it! results! When you edit an article, you will see a copy button on each image in the View it! results. You can copy the code for the selected image and paste it into the editor - works in both visual and source editor.

Development

During development, we held many community discussions about the tool, and collected additional feedback through View it! talk pages on Wikipedia and Meta. From user suggestions, further features were developed, such as the advanced search (which can be accessed through Toolforge). Advanced search allows for more refined searching, including free text, using alternate statements, and Commons category results.

View it! is currently in use by 114 users, and has particularly been helpful on smaller wikis to help add images to articles that previously did not have any. View it! is also valuable for identifying items with inaccurate data statements, and can be used to improve associated metadata and results.

What's next

While the Full version is currently usable, it is under active development. We are working on more editor-focused features, such as improving the inserting an image while editing feature. If you'd like to provide suggestions or feedback on those features, or anything else, please comment on this Signpost article, or visit the View it! Talk page. The team responds to all comments we receive, and we incorporate good ideas into our design. Please check out our Launch video from January 12th!

+ Add a comment

Discuss this story

These comments are automatically transcluded from this article's talk page. To follow comments, add the page to your watchlist. If your comment has not appeared here, you can try purging the cache.
  • I do quite a bit of categorisation on Commons so I'm glad to see the tool taking advantage of this. But I'm also aware of a certain backlog on categorisation there. I think that the vast majority of files on commons have some useful categories. However this sort of change highlights the need for better image finding tools - for example image recognition software that can find similar categorised images to our uncategorised ones and thereby let us concentrate on the unique new images. ϢereSpielChequers 16:45, 16 January 2023 (UTC)[reply]

















Wikipedia:Wikipedia Signpost/2023-01-16/Technology_report