search

opensafely / covid-vaccine-effectiveness-research

MIT License
0 stars 0 forks source link

Update flow chart #42

Closed acagreen17 closed 3 years ago

remlapmot commented 3 years ago

Looks good, thanks. It's sometimes helpful to add a screenshot of the finished result in the PR if it doesn't show properly in github (mainly html files!)

Just to say GitHub deliberately do not render .html files within their code view of a repo.

The confusion arises because they do render Jupyter notebook .ipynb files which are basically html files (I'm guessing renaming .html to .ipynb would probably work as a hack, but you don't really want to do that).

For public repos you can view the rendered html file by viewing the link to a .html file in https://htmlpreview.github.io/ (or enabling githubpages for the repo - although githubpages is a bit of hassle unless you're making a website and not an option for private repos). So downloading and viewing locally is probably what most people do for private repos.

wjchulme commented 3 years ago

Thanks @remlapmot. We've had a lot of discussions within OpenSafely about how to resolve this for private repos. Plus other annoying gaps in rendering for standard .md files.

I hadn't considered converting to .ipynb! Might be worth doing in certain circumstances so thanks for the tip.

acagreen17 commented 3 years ago

Hadn't thought about viewing the html via github...probably because I've only run it locally and not on the server and as such haven't noticed the issue yet. No reason to render to .html (just my default output for .rmd files), would a pdf or work? If not, quick bit of googling tells me that you can 'knit' a rmd to a ipynb if that's going to work better. @wjchulme let me know if it's best to leave as is or if a different output type would be better.

sebbacon commented 3 years ago

We have started a new feature for outputs to be published to the Job Server, rather than Github (where they can still be archived, but not viewed).

The primary use-case is to make it easier to back out of accidental disclosures.

However, several other beneficial side-effects will accrue, such as the ability for researchers to publish outputs and code independently of each other, and most relevant here, we'll have complete control over rendering of outputs. I anticipate a nice CSV viewer, an image gallery, and HTML and notebook viewers.