[gthole]

Yeah, mine is driven by Django Rest and an Angular app for the UI. I'm seriously considering moving to React, though. I just use Sqlite for the DB layer because it's easily performant enough.

I used Drink & Tell (and its sequel Drunk & Told) for the primary source of recipes, which I can't recommend more.

The "related recipes" thing is interesting. If I stick to doing it with a sql query, then I think an algorithm to pick out the base spirit, then filter on that and on whether it's shaken/stirred/scaffa, then sort by common ingredient count? Then again, an Elastic Search stack can do related objects out of the box, but I wouldn't want it to get that complex :)

[getup8]

I haven't tried standing yours up yet but will tomorrow! Do you have all the Drink and Tell/Told recipes as fixtures? Or just a few?

Your repo is nice and clean, definitely learning from that, and your model design.

Any reason you haven't deployed as a web app?

Oh and for related.. an issue is that sometimes related recipes are really quite different (e.g. an old fashioned riff with Gin and St Germain, or tequila/mezcal) so would be tough to automate..

[gthole]

It's deployed alright, I just don't publicize it since I only intend it for personal use.

None of the D&T recipes are fixtures, since they're technically copyright material. I set up classic cocktails as fixtures for people who want to spin up their own or play around with the repo.

Did you try standing it up locally?

Yeah the related cocktails thing is definitely an interesting problem.