Add CI/CD pipeline and re-establish existing tests

[JR-1991]

Describe your environment

• pyDataverse: 0.3.1
• Python: 3.10
• Dataverse: 6.0

Describe the PR

• What kind of change does this PR introduce?
  • CI/CD pipeline to test pyDataverse using the dataverse-action
  • Minor changes to fetching Dataverse action outputs for testing
• Why is this change required? What problem does it solve?
  • Re-establish the existing testing framework

ToDos

• Test against multiple Python versions (3.7 to 3.11 )
• Change action branch from test-action to main once Added action tests dataverse-action#8

Testing

• Have you used tox and/or pytest for testing the changes?
• Did the local testing ran successfully?
• Did the Continous Integration testing (GitHub Action) run successfully?

Others

• The tests have been executed successfully (example run).
• @pdurbin @poikilotherm - Currently, the action only tests against version 6.0. However, this pipeline should check multiple versions. @skasberger proposed that we test from version 5.x onwards. Therefore, we need to extend the Dataverse action.
• Tests run exclusively for Python version 3.10 at the moment but will expand to others in the PR progress. Is '>3.7' reasonable?

[poikilotherm]

• @pdurbin @poikilotherm - Currently, the action only tests against version 6.0. However, this pipeline should check multiple versions. @skasberger proposed that we test from version 5.x onwards. Therefore, we need to extend the Dataverse action.

This might actually be harder to pull off than it sounds like.

DV 5.x requires Payara 5, so different base image. We could build that of course, even cache some versions images.

But to actually make this work, it might need doing some backports, because we evolved the images quite a bit when going from 5.x to recent develop, soon to be 6.1.

[JR-1991]

@poikilotherm, it's good to note that the tests on pyDataverse were only done on demo.dataverse.org up to this point. Therefore, no functionality will be lost. Also, the recent changes from version 5.x to 6.x may not affect pyDataverse testing too much since most of the endpoints pyDataverse has implemented remain the same (to my knowledge). However, it is still worth considering tests across versions in the future.

How do you think we should proceed? Stick with 6.x exclusively or work on the backports? I'd remove the ToDo here if we stick with the current case.

[sonarqubecloud]

Kudos, SonarCloud Quality Gate passed!   

0 Bugs
0 Vulnerabilities
0 Security Hotspots
3 Code Smells

No Coverage information
0.0% Duplication

[pdurbin]

I didn't dig into the details but I'm so happy to see some CI working again! Ship it! 🚀

[skasberger]

@poikilotherm, it's good to note that the tests on pyDataverse were only done on demo.dataverse.org up to this point. Therefore, no functionality will be lost. Also, the recent changes from version 5.x to 6.x may not affect pyDataverse testing too much since most of the endpoints pyDataverse has implemented remain the same (to my knowledge). However, it is still worth considering tests across versions in the future.

How do you think we should proceed? Stick with 6.x exclusively or work on the backports? I'd remove the ToDo here if we stick with the current case.

I tested it before on local running docker images and on our two instances we ran at AUSSDA, so something very recent, and some version half a year old or so. But never something really old, like more than 1.5y or one major version back.