I’m building a Zavod crawler for a custom wanted-persons dataset that includes publicly available photos for each individual, and I would like these images to appear in my OpenSanctions/Yente server. I’m unsure what the recommended approach is for handling photos in FollowTheMoney — should I simply add the image URL as a sourceUrl/noteson the Person, or is it appropriate to create a separate Image entity and link it?
The FtM docs mention that Image is a generated schema, so I’m not sure whether it’s supported or recommended to store images within FtM entities at all. Any guidance or best practices on how images should be represented in a crawler would be greatly appreciated.
This might very well be the oldest open issue we’ve never gotten around to address (Add image URLs for politicians · Issue #218 · opensanctions/opensanctions · GitHub) - in part because I’m a bit afraid of the copyright implications of including source image links (which implies that people should use them). Haven’t honestly spoken to a lawyer about it yet, though … In any case, we’ll want good attribution and hot-linking…
Regarding how you’d do it: Image objects seem fine (the “generated” flag just refers to the idea that interactive web UIs for FtM editing shouldn’t prompt users to create this entity type). Similarly, you’re not affected by this, but: there are tools for rendering FtM data (Aleph) that expect Image entities to actually reference an uploaded image blob using their checksum field. It might be nice to leave that prop empty to make it clear that no such blob is available.
