CA as collaborative/crowdsourcing platform?

mpassos

Hi all, I'm curious if anyone ever attempted to use CA as a 'public' platform. Here are some thoughts/considerations/requirements I'm unsure could be met:

Users can create their own account (configured to have a default role), preferably with social logins
Users can describe their own images (and potentially others', in a 'wiki' fashion)
Users can submit their own images (via our own submission interface that communicates with CA's API)

UI-wise we'd probably create our own and consume CA's API, or heavily customize Pawtucket.

I'm considering CA for this because of the available VRA Core profile (our elected data standard), nice date parser, IIIF capabilities and Getty vocabs integration.

Thanks for any inputs and considerations about this scenario.

seth

mpassos Users can create their own account (configured to have a default role), preferably with social logins

There is a registration process. It does support SSO with common providers used in organizations, including LDAP, ActiveDirectory, Shibboleth and SAML. It does not currently support social logins (Google, etc.) because no one has asked for it yet, but it could be done. It's just another authentication adapter.

Users can describe their own images (and potentially others', in a 'wiki' fashion)

There is provision for commenting on records, and people have made plugins to allow users to enter data into specific fields on records. There is also provision for spatial annotation of images by users; at least one theme also implements this for video - the plumbing is there to support it, but the UI has never been standardized.

There's a submission interface built into the current version of Pawtucket. The basic workflow is user logs in, clicks submit, gets a theme-specific form (with whatever fields you want to expose for submission) that can include media uploads. Each submission is queued for review by an archivist on the back-end. It's very basic and meant for classroom/project use. Size of uploads is limited because it's using tradition HTML <form>-based requests.

The next version of iteration Pawtucket, which we're using for our own projects but haven't released yet, also includes a system for upload of very large media (any number of > 1 gb files using the Tus file transfer protocol) with user-provided description for submission to the archive. This system is already in use by a project to support field submission of video.

Users can submit their own images (via our own submission interface that communicates with CA's API)

The GraphQL API only supports indirect loading of images via URLs. Eg. you create a record and set the media to a URL and CA will grab the image off of that URL (assuming it is configured to allow this - by default it's not going to allow this for security reasons). We are exploring better ways to allow for large file uploads in a GraphQL context and would welcome suggestions.

The older REST API can use traditional POST submissions to upload files, if that's what you want.

mpassos

Thanks @seth, looks like there's already more features than I expected in that sense.

seth Each submission is queued for review by an archivist on the back-end

I'm assuming this can be bypassed? Captcha and other security measures (maybe AI to decline adult content, etc) would be welcome but we don't foresee a dedicated person for approving submissions. Of course I don't expect the team to implement these, just checking what the possibilities are.

Regarding file size we are expecting mostly mobile phone photographs so it shouldn't be huge. Even the hi-res digitized archival materials should be under 1GB.

seth people have made plugins to allow users to enter data into specific fields on records

Can you explain why I would need a plugin rather than giving the general user role permission to edit metadata?

Finally, would you advise against using CA as a headless CMS?

seth

You can set the form to mark submissions as approved without action if you want to. It's just not how it's typically used.

Regarding plugins, there is no provision in Pawtucket (the front-end) for editing fields, by design. It's a publishing platform, not an editing platform. In odd cases where such behavior was desirable it was achieved with plugins.

You can use CA as a headless CMS via the API if you want. But for what you're doing why not just use Wordpress?

mpassos

seth Regarding plugins, there is no provision in Pawtucket (the front-end) for editing fields

Got it, thanks. I was still thinking in Providence terms.

seth But for what you're doing why not just use Wordpress?

Project management does not like how Wordpress models data, we were required to implement a VRA Core DB "to the letter". We're likely not using CA either but Laravel, this is just me investigating for future use cases.