search

atra2396 / READUS

Aggregate and link README files across all of your repositories
GNU General Public License v3.0
0 stars 0 forks source link

Implement support for a real SCM provider #2

Closed atra2396 closed 4 years ago

atra2396 commented 4 years ago

Goal

READUS should be able to read data from at least one real SCM system that lives on the web.

Details

Azure DevOps would be the most immediately useful to us, but GitHub might be easier to work with. I will let whomever picks this up decide where to begin. I don’t think handling user authentication is necessary for this one - assume an existing record exists for the organization/repository in the storage layer. There is currently an in-memory model for the storage layer, so this should be doable without having a real DBMS yet. This may require changes to the domain objects, which is totally fine.

Requirements

eurubkov commented 4 years ago

@atra2396 For MVP, should we allow users to get readmes from the selected repository, or can we just get them all from all repos across the project for Azure Devops and across the organization for Github? For github, to get all of them across the org all you need is the org name and an auth token. To get them just for a particular repo, you'll also need the username to whom the repo belongs (all github repos have the format of {username}/{repoName}. For devops we might have to iterate over all the repos though.

I branched off the AzureDevOps branch and implemented GitHub SCM to get the readmes across all repos. We can discuss the path forward whenever you have time

atra2396 commented 4 years ago

@eurubkov I would lean toward the latter, collecting as many readmes as we can. The site is more valuable the more documents it contains, so I'd err on the side of over-collecting than under-collecting. We can add the option to exclude documents in the future.