Project

General

Profile

Actions

Question #2182

closed

Host multiple projects within OpenAtlas

Added by Gary Hsu 8 months ago. Updated 8 months ago.

Status:
Closed
Priority:
Normal
Category:
-
Target version:
-
Start date:
2024-02-13
Estimated time:

Description

Hi,

I was wondering how openatlas would allow the organisation of multiple projects. My thinking is each user may create his own project and enter data accordingly, but now the default workflow seems to firstly define individual entities and then link them together.

which entity should I use to define a project?

Many thanks,
Gary


Related issues 1 (0 open1 closed)

Related to OpenAtlas - Feature #2187: Manual: FAQs about multiple projects and data accessClosedAlexander Watzinger2024-02-16Actions
Actions #1

Updated by Alexander Watzinger 8 months ago

  • Assignee set to Alexander Watzinger
  • Target version set to 8.2.0
  • Subject changed from Host multiple projects within openatlas to Host multiple projects within OpenAtlas
  • Status changed from New to In Progress

Dear Gary, thank you for your interest in OpenAtlas. Depending on the need to share data between the projects there are two main approaches:

Separate instances
If every project has their own instance (installation) of OpenAtlas than you don't have to worry about separating the data later. But it would also mean that data can't be used together, at least not directly.

Same instance
We have some cases where multiple projects want to share their data, e.g. information about places or persons.
They than use the same instance and create a custom type (see manual), usually called Case Study
  • The type can be "attached" to multiple classes (probably most of them)
  • The type can be configured to be used multiple so that e.g. a place can be part of multiple case studies
  • The type can be configured to be required so that users don't forget to specify it when adding data

Later on, e.g. when running analyses or developing a presentation site for one case study, this type can be used to separate the data again.

Please let us know if this answers your question but feel free to ask more if not.

Actions #2

Updated by Gary Hsu 8 months ago

Dear Alex,

Thanks for your reply.
If we choose to use the same instance, can data from different projects/users be kept privately or hidden from each other?

Actions #3

Updated by Alexander Watzinger 8 months ago

Keeping data privately or hidden is not the focus of OpenAtlas. On the contrary, our cooperations are exclusively with open data projects and the idea of open science is very important for us, see our mission statement.

That being said, of course OpenAtlas itself is open source and can be used for any purposes and data doesn't have to be made public. Nevertheless, hiding data from other users isn't implemented and could also lead to several problems, like e.g. duplicates.

Actions #4

Updated by Gary Hsu 8 months ago

I totally agree with the FAIR principle, as I am currently working in NFDI4Object (National Research Data Infrastructure for Archaeological Objects). I am asking this because sometimes researchers would like to own a digital working space to store unpublished data. Some data repository service would set the limit for an embargo period. It happens to scientific data (analytical results). However, I respect what OpenAtlas is aiming for.

On a related note, if I export the whole SQL database, can I import the database to a new instance (assuming the original installation is malfunctioning)? It is like a data transfer between different instances.

Another scenario is that if the new instance contains part of the database from the old instance, will the software detect the changes automatically or will it import the same part twice? Does OpenAtlas have some sort of version control, like a Github repository?

Actions #5

Updated by Alexander Watzinger 8 months ago

Before answering I wanted to thank you for your really interesting questions. Your feedback is appreciated and will be used for further development:
  • Topics that "just" need explaining will be added to the manual (e.g. how to organize multiple projects)
  • Some might turn into issues to be implemented
  • Some might turn into issues but put on the wishlist because of e.g. missing funding or available developers
  • Some we might have to decline for various reasons, e.g. out of scope. But even than we will at least try to explain why

Clarifications about public data

In OpenAtlas it can be chosen if and when to make data public but:
  • Either all data is made public via the API or none
  • On presentation sites it can be implemented which data (e.g. which case study) to show via usage of the API
  • Showing images is a special case because it also depends on licenses (which are reworked currently) and if IIIF is enabled

Until data isn't made public only registered users have access. Like mentioned before it is not possible to hide parts of the data from some users and it isn't planned to be implemented. If this is really needed than using separate instances would likely be a better alternative.

Export/import

Hint: in case you want duplicate/restore OpenAtlas instances don't forget to also copy the files in files/upload.
Exported databases can be setup/imported again. So it is possible to backup and restore or duplicate data.
There is no import script for databases, e.g. a database restore has to be done manually, see manual. That's why nothing can be "detected" which means partial databases would have to be prepared and imported manually too.

In case you have an old database and a new database with more entries which you want in the old one too, I would just replace the old with the new database.
This is assuming there was no data added in the old database which the new one doesn't know about. But in that case (new different data in the old and the new database) it would get complicated anyway.

I hope this explanations could make it a bit clearer.

Data versioning

There is no data versioning in OpenAtlas. The only thing tracked is when and by whom entries are created and updated.
Although I'm aware that this isn't the same, you could of course take regular "snapshots" via the database export just in case.

Actions #6

Updated by Gary Hsu 8 months ago

I apologise for my stupidity.

Should I run codes for Export SQL or Export Databse dump? or are they just step 1 and 2 for restoring the database?

Actions #7

Updated by Bernhard Koschiček-Krombholz 8 months ago

No stupidity here.

Export SQL is plain text. So you can use psql to restore the database. We also have a wiki entry for that

The SQL dump is a binary file, so it is compressed. For the dump you need to use pg_restore to restore the database. We use it only under Windows. -> Wiki entry about restoring a dump

Actions #8

Updated by Alexander Watzinger 8 months ago

  • Related to Feature #2187: Manual: FAQs about multiple projects and data access added
Actions #9

Updated by Alexander Watzinger 8 months ago

  • Status changed from In Progress to Closed
  • Target version deleted (8.2.0)

Thank you again for reporting.
I added FAQs about multiple projects and data access (#2187) to the manual and it can be already viewed at the demo-dev at https://demo-dev.openatlas.eu/static/manual/faq.html

Actions #10

Updated by Alexander Watzinger 8 months ago

I also closed this issue to avoid mixing too much different topics. But feel free to open new ones anytime.

Actions

Also available in: Atom PDF