It's nice - maybe not yet final, but I like the idea and how it currently works!

You could probably reuse the same engine to do the same with cocktails and maybe other stuff.

Thank you. Sometimes it feels like development progress is slow and that the UI could do with a complete overhaul (the project does have some design volunteers offering time and improved designs which are being incorporated slowly - I'm the bottleneck!); but it's gradually improving and enjoyable to work on.

All the components and dependencies are open source, so there's a social and learning element involved when integrating new features and finding bugs / improvements in other projects.

Cocktails could certainly be another domain this could work for, and homebrewing is another that springs to mind. I'm not currently intentionally keeping the core services extensible to those, but forks and/or later generalization of the application are both possible.

Just some additional thoughts:

1) you might have to do in the future some deduplication of entries in your master index - e.g. by searching https://www.reciperadar.com/#search&action=search&include=ba... I see that the entries pointing to thepioneerwoman.com & tastykitchen.com for "Pasta Carbonara" are identical (the 2 sites might be mirrors of each other or one site might just have copied the contents of the other one) => depending on how much your index will grow you'll maybe want to show only 1 result pointing to 2 different URLs.

2) The scraping of contents of other websites coupled with displaying it directly on our own website might become problematic. Meaning that you would be an "ally" of the source websites if you would just show the results and the overview of the needed ingredients but then just point to the recipe hosted on the other website, but you would become an "enemy/opponent" if you would display the recipe directly on your site.

Meaning that on one hand I would not be happy at all if somebody would steal my contents and stop users from getting to my website (I would therefore e.g. blacklist their bots and IPs from which they would try to scrape my site), but on the other hand for somebody that might increase traffic to my website I would potentially even provide a special API so that that bot could get better quality data (instead of scraping the html).

Cheers & good luck :)

The automatic translation is very weird. If I open the website in my browser it's in french, but the search and recipe is only english keywords, and most of the translated text is plain wrong, like it is straight from google translate. Is there a platform where I could help with the translations ?

Hello, and apologies for the late reply!

Yes, the application text is localized using Apertium[1] machine translation, and it's... not perfect yet :)

At some point I would like to integrate 'Weblate'[2] as a translation platform; for now I'm afraid the best I could do is point you to the PO files for RecipeRadar here:

https://github.com/openculinary/internationalization

If you're able to add some corrections under the 'locales/corrections/fr' directory then your help would be greatly appreciated :)

NB: Thanks also for highlighting the need for ingredient/equipment name translations too; that'll need a little design & planning but is important to handle. I've opened an issue to track it here: https://github.com/openculinary/knowledge-graph/issues/51

[1] - https://www.apertium.org

[2] - https://weblate.org/

a search category would be helpful. just recently i was searching for cake recipes, but if i enter the ingredients list i get all sorts of suggestions and only a few cakes.

possible categories could be eg. breakfast, lunch/dinner, snack, desert, cakes, cookies, sweet, savory, soup, dry/portable (for the road), fingerfood, cold, warm, no cooking (like sandwiches, salads, etc)

the list likely depends on what categories you are able to detect in the recipe data, but you get the idea.

It seems you create the data by scraping other websites?

Correct! With many thanks and credit due to the excellent recipe website crawling community behind the recipe-scrapers[1] Python library.

[1] - https://pypi.org/project/recipe-scrapers/

I would think re-publishing that data violates the copyright of the original publishers.

Ingredient lists and bare directions are not copyrightable, though the wording and style might make the directions copyrightable (IANAL, but see https://duckduckgo.com/?t=canonical&q=recipe+copyright&ia=we... ). https://masterwiki.how/courses/how-to-make-perfect-scrambled... probably has some copyrightable phrases =P

So direct copying without any transformation could be a bit risky.

Not only because of the wording. Collections are prone to copyright. So scraping a bunch of receipes and republishing it would violate the copyright holder of that collection of receipes.

Also - what about the images? Where are they from? It looks like they are scraped too.

How can I search for recipes with:

cans of tuna tuna can canned tuna can of tuna

(none of the above works)

AND

teriyaki sauce

OR

ketchup

?

Hey - thanks for mentioning this. I've filed an issue at https://github.com/openculinary/backend/issues/24 - in short; RecipeRadar has crawled some tuna + teriyaki recipes, so you've uncovered a bug indexing those and making them searchable.

(I've added some debug SQL and output to the issue in case you're curious)

Substitutions (i.e. 'ingredient x OR ingredient y satisfy my constraints') are planned. I'm yet to fully spec the feature out and consider all the use cases, so I'll make sure that's one of them.

[1] - https://github.com/openculinary/knowledge-graph

Update: the situation should now be improved; some recipe content has been re-indexed, and an ingredient parsing problem was discovered and fixed too.

There are still only three recipes that contain both tuna and teriyaki sauce appearing, so coverage could likely be improved further. But it's a start, I hope!

How can I browse the recipes?

As a fairly intentional design decision, you need to provide context from your kitchen - a search engine query, essentially - before you see results.

The concept is that RecipeRadar will ultimately help you track your kitchen inventory (all on your own device; there are no user accounts), and then the search step will disappear into the background - you'll be straight into browsing recipes that match your context.

It's not there yet and I'd appreciate hearing about how you like to browse recipes in case that's something else that can be supported.

You may want to provide 'suggested cupboards', with a limited selection of multi-use ingredients, to show how the app can help make the most of ingredients.

That's a great idea - "kitchen staples", perhaps? Thanks!

No problemo!

Sure, you may want to come up with a few different sets, perhaps Italian stuff, Asian-focused, vegetarian, etc.

I've translated this into a rough use case / spec at https://github.com/openculinary/frontend/issues/172 in case you'd like to follow along.

I definitely will! Thanks for the heads-up!