[Feature Request] Prepare test infrastructure for workshop

[bartvanb]

Problem description In the workshop, participants should be able to login to their own vantage6 environment with the UI and do stuff.

Desired solution

• [ ] Server with two collaborations per participant
• [ ] Each collaboration has two organizations and corresponding nodes. One organization is in both collaborations.
• [ ] Each participant needs collaboration admin-like role and researcher role for each of the collaborations (we can probably just create two users per participant if we create them for the organization that is in both collaborations, otherwise maybe four)
• [ ] Ensure access to algorithm store, also to modify stuff there (for algorithm development)
• [ ] create project in harbor2.vantage6.ai/workshop where they can push algorithm images
• [ ] Run nodes for participants somewhere
• [ ] Ensure they can do the appropriate actions but cannot delete too much - probably need to create custom role for this.

Additional context https://github.com/vantage6/vantage6/issues/1314 would make it easier for them to run complete v6 dev networks

[hcadavid]

@bartvanb the challenges on chapter no. 3 (and eventually on chapter 5) are based on a scenario with four organizations and two collaborations (basically to make them explore multiple collaboration configurations with different statuses through the UI). Now I see that four may be too much :), but could we have at least three? (so there are still two collaborations for them to explore). With only one or two nodes these challenges would need to be redefined. Three, on the other hand, would require only a minor update.

For the same exercise, four users are considered (one admin and one researcher for each collaboration).

[bartvanb]

@hcadavid good that you spotted this! The most important for me is that we limit the number of nodes we need to run per participant, as that requires most computational resources. The registration of multiple collaborations/organizations etc at the server is peanuts, it's something we will script anyway and then it doesn't really matter how many loops we run.

Am I understanding correctly that we need 3 nodes divided over 2 collaborations? Is there then one collaboration with 2 nodes and the other one with 1?

[hcadavid]

@bartvanb I'd describe it as three organizations (OrgA, OrbB, OrgC) divided over two collaborations (ColX, ColY). In other words two organizations with one node each one: (OrgA-ColX, OrgB-ColY), and the third organization with two nodes -one for each collaboration- (OrgC-ColX, OrgC-Coly) - so four nodes in total.

As I mentioned, if having (4*#participants) nodes is too computationally expensive, depending on the maximum number of nodes per participant (two, or three?) we may need to change the context given on EP3 (which illustrates, among others, how the same organization can be part of multiple collaborations) and also some of the challenges for EP3 and EP5.

[bartvanb]

OK I updated the TODO list above, please check if you agree!

[hcadavid]

Looks good to me! I'll add an issue to update EP3 scenario with three orgs.