As a technical user of image suggestions, I want to understand how it works, and what logic is included.
This task is to ensure that the image suggestions pipeline is documented end to end.
The data platform team has created the following for their side of the documentation:
https://www.mediawiki.org/wiki/Platform_Engineering_Team/Data_Value_Stream/Data_Pipeline_Onboarding/
https://www.mediawiki.org/wiki/Platform_Engineering_Team/Data_Value_Stream/Projects/Image_Suggestions
https://www.mediawiki.org/wiki/Platform_Engineering_Team/Data_Value_Stream/Data_Gateway
The Structured Data team needs to document their side of the pipeline, and how it interacts with the Data Platform.
Acceptance Criteria:
- A new page is created as a subpage of https://www.mediawiki.org/wiki/Structured_Data_Across_Wikimedia/Image_Suggestions with the technical documentation for the image suggestions pipeline.
- Please consult with @Sannita about where on the main page to post a link to this documentation.
- The documentation is clear about logic on the Structured Data side of the pipeline, including things like how it's determined if an article is unillustrated; how image suggestions are found; how the confidence score is calculated; what confidence score is used; etc.
Documentation at https://www.mediawiki.org/wiki/Structured_Data_Across_Wikimedia/Image_Suggestions/Data_Pipeline