shared status shouldn't be by study but by processed data

[antgonza]

This issue was initially discussed in #850 but we realized that this will require bigger modifications to actually work well.

Description:
Let's imagine that a study has 2 raw datas (A/B) and 2 prep templates (1/2) in each one (for clarity, we will have A1, A2, B1, and B2). Then the users makes public the study because A1 was published, then all the other raw datas + prep templates will also be public, which is not OK. What we will need to have is that the user makes public a process data, then all the files that generated that "biom" become public. We will display public studies based on their public process data.

[josenavas]

The proposed changes will be to add the "status" property in a per-object basis, and the study one can automatically populated (or removed) by looking at the status of its objects.

Pros:

• Once the Metadata has been accepted, the user can make public any analysis, (pre)process data that has been already published while he is working in other unpublished work within the same study.
• We are going to avoid study duplication: in case that a user has some published data but he is working a follow-up unpublished work, he can still work in the same study.

Cons:

• HORRIBLE database/codebase changes.

Even the amount of work behind this issue is considerable. I would suggest that this issue is really important and should be solved ASAP, before the code base continues growing...

[ElDeveloper]

The proposed changes will be to add the "status" property in a per-object basis, and the study one can automatically populated (or removed) by looking at the status of its objects.

🔥 😱 🏃 💀

How much more complex is the system going to become if we add this
feature? Are the benefits much larger than the amount of
work/modifications we need to do to accommodate this?

I'm not worried about the work or modifications, I'm worried about this
complicating the way users interact with their datasets.

On (Feb-13-15|14:05), josenavas wrote:

The proposed changes will be to add the "status" property in a per-object basis, and the study one can automatically populated (or removed) by looking at the status of its objects.

Pros:

• Once the Metadata has been accepted, the user can make public any analysis, (pre)process data that has been already published while he is working in other unpublished work within the same study.
• We are going to avoid study duplication: in case that a user has some published data but he is working a follow-up unpublished work, he can still work in the same study.

Cons:

• HORRIBLE database/codebase changes.

Even the amount of work behind this issue is considerable. I would suggest that this issue is really important and should be solved ASAP, before the code base continues growing...

---

Reply to this email directly or view it on GitHub:
#855 (comment)

[antgonza]

I agree with @ElDeveloper, I think the users should only be able to make public 1 thing: the processed data and the permissions should carry down. Per object seems tooooo complicated, without any benefits, at least that I can see clearly.

[josenavas]

Ok. I agree with the solution that we want to do it per processed data (not by per prep template - as the title says).

The only reason why I suggested in a per object basis is in the case that the user used the system to upload some sequences and then submit those sequences to EBI. Then those sequences are publicly available on EBI, and I think it would be also useful to be publicly available on QIITA.

What do you think about this case?

[antgonza]

Excellent point but we also need to check the map file so we can submit to EBI and we do that by PCoAs ...

[josenavas]

I don't think that @ackermag is using PCoA to validate that the metadata is correct and giving access to upload to EBI... Although probably we want to change that.

Sometimes people is doing analyses outside qiita and they just want to use it to upload their data and make it publicly accessible, rather than performing the entire analysis on qiita...

[antgonza]

If someone that we do not know want to do that, I think they should use their own EBI submission protocol ...

[josenavas]

Fair enough! So if I understand correctly this is the changed proposed:

Add the status to the processed data object and once the user makes that public, that processed data and preprocessed data, raw data and prep/sample template becomes public, right?