Tox is failing 2 out of 4 tests..(??)

[riccardoAlbertoni]

Tox seems to fail 2 out of 4 tests, though I think I have tested the same things manually via http://localhost:8080/v1.3.0/ui/ and I am quite confident that everything is working... instructions I used and the related screenshots in the last sections of the notebook ([1] or [2]).
@dgarijo and @mosoriob: Any chance that the tests get misconfigured? Or have I misunderstood the instructions?

[1]https://colab.research.google.com/drive/1XKk_feh4Eb-fC3FGQ9kGwfiP7KieswT3?usp=sharing
[2]https://github.com/riccardoAlbertoni/ISWC_reproducibility607/blob/master/Reproducing607.ipynb

[dgarijo]

Thanks, I will have a look. Maybe there is something wrong with the tests. @mosoriob do you have time to look at this?

Otherwise I'll have a look during the weekend. In any case, @riccardoAlbertoni the important thing is that you got the API working (even though I would like to have this working for you as well)

[mosoriob]

You need to modify the tests because they use a dummy id.
For example, the GET test is trying to get the resource with the id id_example. However, the resource doesn't exist in dbpedia.

        response = self.client.open(
            '/v1.3.0/bands/{id}'.format(id='id_example'),

You can found good examples in the model catalog api

[dgarijo]

Thanks @mosoriob I will review and update the documentation in case it's necessary to reflect this.
It makes sense, because it's impossible for to know a priori what are the sample resources in the endpoint, unless we do a query to retrieve them.

[dgarijo]

@riccardoAlbertoni if you want to make sure the tests run, I would select an example of a band and an example of a genre and give it a try. Just making sure you complete the notebook :)

[riccardoAlbertoni]

@riccardoAlbertoni, if you want to make sure the tests run, I would select an example of a band and an example of a genre and give it a try. Just making sure you complete the notebook :)

I have updated the notebook. Thanks a lot. In the notebook, you can find some suggestions on how you could improve the documentation on that specific point if you want.

[fbvinid]

Hi everyone,

Not sure if this is just related to my system, but I had some problem running the experiments with tox.

The error I got was: py.error.EACCES: [Permission denied]: mkdir('/home/vinid/Downloads/replicate_review/OBA/outputs/dbpedia_music/servers/python/server/.tox',) . This might just be my docker installation that has created the folder with a different permission setting.

I solved the problem by quickly changing the permissions setting of the server folder. Eventually, I also tried running the experiments with pytest installed in a local virtualenv, just to try something different (see here)

Everything else seems to run/to build correctly: I could run the first part flawlessly (see here and here), for the second one I have to thank @riccardoAlbertoni and the authors for the great work, I was a bit lost when I had to run the server, but the commands in @riccardoAlbertoni notebook colab notebook were super helpful to put the server up.

My configuration:

OS: Ubuntu 20.04
JAVA Version: 11.0.8
Docker Version: 19.03.8

[dgarijo]

Thanks @riccardoAlbertoni , @fbvinid
glad you could all reproduce everything in the end. I will try to use your feedback to improve the documentation. This was very helpful.

[dgarijo]

I updated the documentation for the testing part: https://oba.readthedocs.io/en/latest/test/
@fbvinid what do you think was confusing in the instructions for running the server? Is there anything else we should add to the documentation?

[fbvinid]

Hi @dgarijo :),

I think the most difficult part for me was running docker. I used it in the past, but not at a level that allows me to mess too much with the commands I find on the README files... I was able to get everything in place thanks to @riccardoAlbertoni's notebook.

It might be worth adding a visual example here on how to use the APIs, but I don't think that is a big issue.

Apart from the fact that it is impossible to run docker in the notebook, I think that having something like that and link it in the README could be useful for the users (something along the lines of Running APIs for DBpedia Music). For example, a step by step tutorial, all in one page, that describe how to set up everthing from start to end with some pictures (essentially, what's already in the notebook).

[dgarijo]

Got it, thanks!