Ingredient description parsing

[nerdstep]

Hey @jakeboone02!

Working with various recipes, I've encountered some scenarios where the ingredient description may sometimes also contain a unit.

For example:

1 1⁄2 pounds potatoes (any variety), scrubbed and cut into 1 1/2-inch pieces
3 tomatoes (about 1 pound), chopped

Which is fine, until you start converting units to metric system, which results in something like:

680 g potatoes (any variety), scrubbed and cut into 1 1/2-inch pieces
3 tomatoes (about 1 pound), chopped

It would be nice if we could extract 1 1/2-inch and 1 pound as parsed units that we could then apply conversion against.

[jakeboone02]

This is something I've thought about for a long time, but always came away at a loss as to how to detect those quantities reliably in a procedural way. Like, a natural language processor would easily parse that kind of thing, but structured logic would either be way too complex for a small library like this or too ineffective to be worth the effort.

If you could come up with a simple(-ish), effective algorithm for this case I'd be happy to reconsider!