Notifications: "Watch dataset" update alerts (machine and person) subscription

[raprasad]

Functionality

• Allow a researcher to register for dataset updates.
  • If a dataset is updated, the researcher receives an email

Use case:

• genTB (http://gentb.hms.harvard.edu/) needs to regularly pull an updated "master file" of genetic information from Dataverse.
• Each update to the "master file" creates a new DataFile id
• What is the best way to keep retrieving the "latest" version of a file, assuming the file name stays the same?
  • Suggestion: Send an email update to registered users when the "dataset" itself changes. Leave it up to individual users/programs to figure what file they need from the Dataset

(origin: meeting with Merce, Christine, Maha, James, Raman)

[bencomp]

What is TB?

[pdurbin]

What is TB?

Tuberculosis: http://tb.datascience.iq.harvard.edu

[raprasad]

The TB application requires the latest version of a specific dataset. Until DV can send some type of signal/alert, TB app will use the DV API to see if it has the latest version locally.

[pdurbin]

@raprasad I'm not sure how much value this issue has in its current form. Can you add more to the description? Do you really need this? Is genTB even integrated with Dataverse? Do we really need a "genTB" label for GitHub issues?

[bencomp]

Harvesting (#813) is related, as it is one way of publishing and retrieving updates. ResourceSync, with the push notification extension, is another.

@raprasad, don't you just want user notifications for dataset updates? Perhaps there could be one generic way of generating user notifications and more fine-grained control of sending emails from notifications (similar to how various services like GitHub and Discogs let you control what notifications are sent as emails or email digests). In user story terms, it could be an epic:

A user can configure sending of specific types of notifications as emails, digest emails or not as emails.

This would definitely require scalable processing of notifications (see #2892).

Related: #868 #1155 (#1342)

[bencomp]

I would say that this might be a duplicate of #900 and/or #868, except for the request for allowing notifications to be generated from dataset updates.

[pdurbin]

@bencomp this issue is starting to have more value just from you linking together various issues I haven't thought about it months. Thanks! :)

[raprasad]

fyi: updated the main ticket description and removed the genTB tag

@pdurbin

[mheppler]

This issue appears to be a duplicate of an older issue #868. I am closing this issue in favor of tracking the original issue going forward.