Closed sameetandpotatoes closed 6 years ago
@ASankaran ^
Does it make sense to integrate this into the ACM infrastructure this way? Having two backends for the frontend especially when one is just for a specific use case doesn't make a ton of sense to me. It seems like we are adding a very heavy dependency on the frontend alone, especially because now there are multiple authentication schemes flying around (past the two already supported)
I think there are three major options that seem reasonable in my mind:
I think @narendasan's second solution is the best. The RP / HackIllinois API was not designed to interface directly with ACM infrastructure.
I think we should have a separate front end for the resume book (and use dns to run it on resumebook.reflectionsprojections.org) or integrate it into the RP site (at reflectionsprojections.org/resumebook). Separate would be preferable to that we can run the same frontend (with a different theme) for HackIllinois.
Discussed with @ASankaran and @apache8080 - we'll table this. The UI for rp and hackillinois doesn't really need to happen at this point - DB queries can suffice for that, until we have the resources to make a separate front end completely separate from Groot (and not tied to net ids, etc)
This integrates into the R|P api with OAuth and allows users to view R|P resumes. It's a little difficult right now to actually do filtering, as I'll need to integrate much further into the R|P api and move away from Groot. Also it doesn't seem like R|P's schema has netids defined which makes things tricky to "fallback" on ACM resumes.
Note: All of this was tested locally, without using the actual Google OAuth app set up. I created a fake OAuth app, ran the R|P api locally, and tested the flow like that.
To get this to actually work, we'll need to add: http://acm.illinois.edu/rp/auth to the list of authorized redirect URIs. Despite that, there may be some other issues that need to be ironed out.