nfdi4cat / voc4cat-template Goto Github PK

Merges toplevel-test -> main

Test if everything works in this new repo.

See failure in #56.

Ontospy:

• Test other than default template in ontospy
• set title in ontospy-docs (requires nfdi4cat/voc4cat-tool#112 and nfdi4cat/VocExcel#10)
• Explore how to add multi-vocab / multi-version index page. --> Both version are less good than pyVODE-generated docs.

pyLODE:

• Add as alternative documentation generator. Done in nfdi4cat/voc4cat-tool#115
• Test deployment, see documentation for vocab_example.ttl

This probably broke when the repo was moved to this organization. GitHub changed the defaults in Feb-2023. The repo was created as personal repo before this date but the move to NFDI4Cat happened afterwards in Mar-2023.

Originally reported at nfdi4cat/voc4cat#70.

This occurs only if the published documentation is built in the release pipeline because the new ìndex.html` is not copied to the root of gh-pages.

In contrast, the merge pipeline moves the new ìndex.html` to the root of gh-pages. Therefore, this issue can only be observed between a release and the next PR-merge.

Merges CI-fix-detached-head -> main

The clean stage failed typically. The problem was a consequence of the "detached head" state. The detached head state is caused by the checkout of a specific commit (in above log Checking out cadcc9e8 as refs/merge-requests/10/head...). This is nicely explained and visualized in https://circleci.com/blog/git-detached-head-state/.

The problem is caused by the way gitlab CI does the chekcout which is by commit-hash. In order to avoid the detached-head it is sufficient to checkout the branch of the merge request after the gitlab initiated checkout. But there is an additional complication: The clone in gitlab CI does not know anything about branches. Therefore, we need to fetch branch information first (git fetch --tags) before we can checkout the merge-request´s branch.

This also sets CI to use Python 3.10.8 instead of latest. This is because some dependencies do not yet work with Python 3.11.

Related to https://gitlab.fokus.fraunhofer.de/nfdi4cat/ta1-ontologies/voc4cat-playground/-/issues/3

The action "CI on pull request create or update" is also triggered if a new comment is made an a PR. This is caused by using the activity type "edited".

Using the defaults (opened, synchronize, or reopened) should be sufficient.

Currently lines in the generated turtle file are attributed to the CI.runner but not the user that created the merge request. example

This should be changed. The commits made by CI should be attributed to the merge-request creator instead of CI.runner. This is a simple change (the code is already present, just commented out).

Release 0.6.0 of voc4cat-tool brought a new CLI. The gh-actions here need to be adapted.

• Check artifact: The update xlsx-file included as artifact should have a name derived from the submitted file instead of Updated_VocExcel.xlsx

• describe where to find the xlsx-file that matches the current SKOS/turtle-vocabulary version (see README.md)
• describe where to find the pyLODE-generated HTML-documentation for the current vocabulary version (see README.md)
• construct url with version for formerly released versions of the vocabulary. Whenever a new version is tagged the deployment should use the version in the url. This requires another pipeline triggered by a new tag on main.

Merges gh-actions -> main

The gitlab CI pipeline file in this repo was translated to github actions by a new tool from github action-importer.

No changes were made after the conversion by the tool.

For more complex templates cookiecutter is a popular tool to generate a new project based on a template. It is currently a bit unclear, if our template has enough fields that demand customisation and therefore justify using cookiecutter.

Another aspect of a cookiecutter-template may be that later changes made in the template can be easily transferred to the generated project. To make this happen cookiecutter must be used together with another tool, e.g. cruft.

Merges issue2 -> main

This MR changes CI to make commits to the vocabulary under the name of the user. Only for cleaning the user "CI runner" will be used.

With this change "git blame" can output the correct author for each part of a vocabulary (turtle-file).

Closes #2

Observed in version

v0.7.4

Description

The index.html page that links the different vocabularies/versions is not present at the root of gh-pages (but one level up in dev/ folder). It should be moved down one level after creation.

The root page is linked from the GiHub UI. For this repo it is: https://nfdi4cat.github.io/voc4cat-template/

This link fails currently. We should add an index page that lists all deployed versions. The page should be updated as part of the publish workflow. The code to create the page should be part of https://github.com/nfdi4cat/voc4cat-tool

Related documentation from GitHub:

• https://docs.github.com/en/repositories/archiving-a-github-repository/referencing-and-citing-content

Related documentation from Zenodo:

• https://developers.zenodo.org/#add-metadata-to-your-github-repository-release

Related documentation on citation metadata (optional):

• All metadata that Zenodo should get can be supplied via .zenodo.json. For more info about this file see Zenodo developer docs or https://github.com/zenodraft/metadata-schema-zenodo.
• https://citation-file-format.github.io - The CITATION.cff file is understood by GitHub as well as Zenodo. As of Aug-2023 using both a CITATION.cff file and a .zenodo.json file has issues, see nfdi4cat/voc4cat-tool#166.

Note: The presence of the file and the fields specified therein implies that data, which were auto-generated by GitHub, will be overridden (not merged!) with data from CITATION.cff or .zenodo.json. So new contributors/authors may not show up in Zenodo if authors are given also in CITATION.cff. This may be desired or not. Just be aware of it.
It is also possible to use one to the other citation formats to prevent side-effects on Zenodo and use .zenodo.json for Zenodo-related metadata, for example to specify the related Zenodo communities).

Tutorial-like docs:

• https://inbo.github.io/tutorials/tutorials/git_zenodo/
• https://coderefinery.github.io/github-without-command-line/doi/

See nfdi4cat/voc4cat-tool#131 for possible checks.

Some checks are only possible if the vocabulary repository is restricted to exactly one vocabulary (config-option single_vocab = true in idranges.toml). Setting this option to true is strongly suggested for all real vocabularies.

Our playground will be unrestricted. So there multiple vocabularies can be present at the same time.

A template file to specify ID-ranges has to be added to this repo. Related: nfdi4cat/voc4cat-tool#108

The purpose of the file is to allocate ranges of IDs for each contributor so that IDs an be created in these ranges without further coordination effort.

Committing to PR-branch from actions fails with Error: Invalid status code: 128 failed job

Description

Markdown?

• yes
• no