Change repository layout #22
Conversation
Divide into `dic2owl` and `ontology`. Remove `.gitlab-ci.yml` completely as it's not used. Update GitHub Actions CI (`ci_emmocheck.yml`) to point at new paths and to use Python 3.8.
Changing some naming from `domain-crystallography` -> `cif-ontology`.
| [crystallography.owl](https://raw.githubusercontent.com/emmo-repo/domain-crystallography/master/crystallography.ttl) | ||
| in Protege, the correct version of emmo-inferred will be downloaded | ||
| and imported. | ||
| When opening [cif.ttl](https://raw.githubusercontent.com/emmo-repo/CIF-ontology/main/ontology/cif.ttl) in Protégé, the correct version of emmo-inferred will be downloaded and imported. |
There was a problem hiding this comment.
Do you plan to store the generated ontology in the github repo? I think it would be better to create a folder structure on GitHub Pages with the ontology files for each release.
There was a problem hiding this comment.
Yeah, for the README I simply updated the text to match the current files, I didn't update any meaning.
The README needs a proper overhaul all the way through to catch things like this.
Feel free to checkout the branch, update the README as you see fit, and commit it here.
The PR is still a draft, I'm thus not finished tinkering with the layout, but we seem to past the first hurdle (#23). From there I'll wrap the script in some CLI and add a setup.py file, etc.
There was a problem hiding this comment.
Do you plan to store the generated ontology in the github repo? I think it would be better to create a folder structure on GitHub Pages with the ontology files for each release.
Note that GitHub Pages takes all its files from a branch in this repository.
So all the files we serve on GitHub pages will be present in the repository, but only in a specific branch. I don't know if there's a way to let git know locally to not automatically fetch a specific branch from a remote, but if not, then there's no getting around the fact that a lot of data will have to be downloaded every time a git clone is performed.
There was a problem hiding this comment.
Two good points:
- You are right, the README file needs a proper overhaul anyway. Lets fix that in a new PR to not delay this one.
- I guess that you are right that we probably don't want to use the separate repo for GH Pages: https://github.com/emmo-repo/emmo-repo.github.io since that is used for top and mid-level EMMO. In that case we have to use a branch. I am not so concerned about the released versions since they would be rather limited, but rather about the continuous changes to the generated ontology for each push. Is it possible to push to GH Pages without tracking history or does it works exactly like an ordinary branch?
There was a problem hiding this comment.
It works more or less like an ordinary branch, but I think we'll do it in a non-working way. Meaning, we can generate all the files needed for the branch, delete everything else and only have that. This will then be force-pushed, and as such the branch will follow the development of the main branch whenever an ontology file is updated.
Is it correctly understood that your concern is towards the changes that arise when we update an ontology file? I think the size of the repository will not change much from these changes - that's the whole point of git, to only store the diff, not the whole file - only once. The issue could arise if we keep making new large files with new names etc.
There was a problem hiding this comment.
Yes, that is correctly understood. My concern is that the changes in the generated files may be quite large even for a very small change in the code that generates the file.
There was a problem hiding this comment.
Right. Due to the weird way different Protégé versions stores files?
Could we maybe have a pre-commit hook that always re-reads and writes any changed ontology files in a specific way? In this way it should be fine no matter the Protégé version, since the change will happen upon git commit.
There was a problem hiding this comment.
That would be a very nice feature - not only for the CIF ontology but for all our ontology repositories.
Run pre-commit with black that converts all strings to be double-quotations. Update path to ontology folder.
Implement parameters for generate_cif.main() in order to use it from the CLI.
|
@jesper-friis since your last review, I have now added all the boiler plate stuff to make the Python module a proper package and also allowed myself to create a CLI using While this could have been implemented as the CLI being To check out the CLI type The equivalent of just running Note that the git diff can be a bit confusing if you look at all the commits. This is due to the implementation of $ pip install -e ./dic2owl[dev]
$ pre-commit installDone. To get a better git diff for the review, check out individual commits to see the actual content changes. |
Closes #21.
Change and update the repository according to the suggested layout in #21.
Remove/Cull unnecessary files and implement minor updates.