Closed mhamilt closed 3 years ago
I've changed it to the Python Humanities lesson from commit eff61038c69c92b7e17f5de6fabc0b1a71d2d053
The original link is from where the data was being edited. Moved to this repo for now.
This still does not render as a direct download link. I think the best thing to do would really be to wrap everything in a zip and have it as part of release.
Python Inflammation links to a zip within the git itself which is then served out as a download link on the page, though I still think this is a messy way to approach things.
In any case, it should be avoided that learners be required to copy-paste source text or navigate GitHub's interface in any way to download data for a carpentry.
humanities-lesson-data-master.zip
data
branch to give a folder of humanities-lesson-data-data.zip
with a download link of https://github.com/iaine/humanities-lesson-data/archive/data.zipdata
. Messy, but this the approach taken with other carpentries so at least there is precedence. [Link to file]({{ site.url }}/path/to/file/in/repo.csv)
Since browsers will handle different file streams in their own way, we can't guarantee this will result in a download. To guarantee a download would require the usage of something like download.js
the amount of effort required to do that far outweighing any of the other methods.
The data set links directly to another repository.
For ease of use and to avoid confusion with learners, a possible approach would be to include the data set as part of a release within
Linking to particular files in a git repository for download is generally not a great idea. As is having
.zip
files in a repository. For sustainability and to avoid link rot, I'd personally recommend making a release for the dataset repository and including a.zip
of the specific python data as part of the release.