Hypothesis + Array API blog post

[honno]

This article gives a brief demo of how one would use Hypothesis to test array-consuming methods. The structure and contents is pretty much finished, but this is still a draft and thus I would recommend a proper review take place once I remove the draft status. As well as going through the checks below, I will solicit feedback and generally mull it over before this happens.

This is part of the internship series of blog posts, but do note that it's purely technical. I'll leave personal musing out of this (and possibly write something for my own personal blog to link in this article), which I'd prefer anyway and gives breathing room for fellow interns to go more personal for their blog posts.

nbviewer, GitHub's preview

---

Text styling

• [x] The blog is written with plain language (where relevant).
• [x] If there are headers, they use the proper header tags in order to do so (with only one level-one header).
• [x] All links describe where they link to (for example, check the Quansight labs website).
• [x] Any kind of styling that the author uses (for example, bold for emphasis) is consistent throughout the blog.

Non-text contents

• [x] All content is represented as text (for example, images need alt text and videos need captions or descriptive transcripts).
• [x] If there are emojis, there are not more than three in a row.
• [x] Don't use flashing gifs or videos.
• [x] If it were to be read as plain text, the blog still makes sense and no information is missing.

Me

• [ ] Update notebook metadata once I get a publish date
• [x] Link personal blog post (if I write it)
• [x] Sort out blog preview on...
  • [x] blog
  • [x] social

For every commit

• [x] Run custom cleaning script
• [x] Check I can Restart + Run All in an isolated environment
• [x] Check it shows and renders correctly when running nikola
• [x] Update gist underlying nbviewer link (viewing files in PR repos is broken)

[honno]

I'm quite happy with this now (would appreciate reviews of-course). The sheet is clear this month but note #251 might be scheduled for Sep 21. cc @melissawm

[honno]

Thank you Melissa for all your great feedback! Unfortunately I haven't quite got round to resolving everything, but I'll probably submit a "ready" draft tomorrow.

I took Ralf's feedback in a call today to end the blog post in a more helpful manner—I ended up talking a bit about "comparison testing", so you could look at that now... I probably need to add an extra sentence there though, as well as finish the Fin section.

My TODO seems to be:

• [x] Maybe add more context at the start
• [x] Nice quote-list style
• [x] Finishing paragraph(s)
• [x] Social preview image
• [x] Blog preview cut-off
• [ ] Update notebook metadata when I get a publish date

[honno]

I'm happy for another review. I changed the intro slightly (to give a tiny bit more context), and added a bit more to the end.

[honno]

@rgommers Awesome, hope it's ready now.

[honno]

one sec (my force push didnt go through for some reason, so had to do it again—all good now!)

[Zac-HD]

Just found this - awesome post @honno! If you're still taking reviews...

• Header image seems to have mis-rendered the colors of the Hypothesis logo - maybe the .png version would be more stable than .svg?
• Since you're keen on pretty-printing, you might want to re-render the MultipleFailures traceback with our new support for Pytest traceback display 🎨
• I wrote up a short conference paper for SciPy last year about PBT-for-scientific-code, which (at risk of self-promotion) is probably relevant when you suggest comparison-testing for readers' code.

[honno]

Glad you like it @Zac-HD :) Yeah I'd like this post to be more of a reference than anything else, so updating the traceback and referencing your paper (handy!) are definitely good ideas. Will make a new PR at some point™ and ping you for additional review.

Header image seems to have mis-rendered the colors of the Hypothesis logo - maybe the .png version would be more stable than .svg?

Haha that's intentional, thought it'd be interesting to mix in the official Array API (or rather Data APIs) colour scheme. Let me know if you'd prefer I keep with rainbow/blue Scout for this and other things going forward.

[Zac-HD]

Awesome, if it's a deliberate variation I love it! Ping me whenever Hypothesis feedback would be useful 😊