Task List

[dansand]

Task list formerly on the Readme.

• Not using GH issues for actual issues on this repo., so we don't clutter the model submission issues

• Fix file size limits, Add note like, larger files can be added to your model repository later
• Is the licence URL being handled properly? Check URL, (Add a set of images to the website, and associate each licence with an image)
• Is there a visual cue for test running again - One option is to have the bot write a on-line "building model information"
• Need to add model submitter as an owner of the repository
• Check this text on this issue template: "Please check our model repository to explore valid licenses."
• UX - explain between model code and software framework
• Can tick yes and no for "intending to submit data"
• If parse issue crashes, a developer/ bug label need to be added.
• Fix the call to Zenodo being hardcoded into the doi/Api
• Fix @id assigment if not found
• Edit permissions for issues (model_reviewers)?
• Is there a way to ensure issues are saved, and cannot be deleted?
• Add some or all of the report to the readme for the new model repo.
• RO-Crate needs to be available to the website. Should we copy to web_material, or is a weblink sufficient? What do crawlers need?
• Add a "published" label to the template repo.
• Set up some mechanism to add a DOI, via an issue, if the doi is validated,it adds the published label.
• Push to the website occurs with published label

Further tasks as they arise

• Add collection DOI and publisher to RO-Crate. Once we have a collection level DOI. This can be done in the defaults_and_customise_ro_crate function
• The add DOI function in the template repo should prepend to the list of values in the identifier field. All identifier items in the RO-Crate are now set up as lists, and multiple entries is fine. Not sure if the same dataset DOI (from NCI) should be added to the other entities (like model_output_data)?
• Need a better way to signal Required/Optional fields in submission form. And also signal what fields can be automatically filled.
• Need to make the M@TE bot a bit quieter with warning emails. A lot the team are receiving these. (can we set up a more restricted team to get messages from the m@te bot)
• Need to draft a policies doc. (https://about.zenodo.org/policies/)
• Move description to Abstract, and use description to describe the crate a license.
• Add embargo section and label
• Crosswalk to ISO metadata
• Publish workflow should also update the Model Code and Model Outputs section of the YAML. (model_files: url:)
• Animation filename path not getting prepended
• Add a how to cite. Needs to be added to the Readme, RO-Crate, YAML, and NCI.
• Quite a few of the parse_issue flow control is based on a No response, entry. We should make this work with empty strings too, in cases where the issue markdown gets edited.
• Automatically add tags to the model repositories
• Deal with the alt. license url if provided
• Add file format section. Had a look at this. Model out formats are not included in typical medial type descriptions, so plain text might have to suffuce.
• A data lineage section that collects things like the github issue #, explains the .metadata_trail folder.
• .metadata_trail, probably shouldn't be hidden, and should have a entry in the RO-Crate.
• parse_issue(issue) error , where response is a MaxRetryError, and we try to concatenate to a log string ( TypeError: can only concatenate str (not "MaxRetryError") to str.)
• if someone forks this repository, would NewModel issue work? Might depend on how teh M@TE organisation is hardcoded. Need to provide some guidance on this.
• If Model Output data are not provided, we should add a short statement to model_output_data notes, and add a 'missing_data' label.
• The workflow around flattenign the ro-crate has become a bit convoluted. Move the pyld functionality to rocrate_utils, and call this in write_repo_contents flatcompact
• parse_issue should have a generic function that can handle csv input, multi-line, etc. All inputs like keywords should be passed through this function, as we can't guarantee the input.
• Add boiler plate section to the Readme that described the default layout of the model repository