Experimental TS wrapper for Fauna

[cjol]

A little while ago, I started a FaunaDB project, and quickly grew frustrated with the experience of writing complex FQL queries. Observability of intermediate results has already been discussed as a problem (you typically only get to see the final result, and if it's not what you expect then it's hard to debug along the way). Another problem I was having is that FQL feels like a functional programming language in many ways, but we lack tools like function composition to really write queries in a functional "style".

I tried to solve both problems for myself with a wrapper library imaginatively called fauna-fp. To save you a click, here's a quick demo:

interface User {
  email: string;
  password: string;
}

// inferred type: (x: Arg<{email: string}>) => Query<boolean> 
const emailValidator = compose(select("email"), containsStrRegex("^.+@.+..+$"));

// inferred type: (x: Arg<{password: string}>) => Query<boolean> 
const passwordValidator = compose(select("password"), length, gte(8));

const usersCollection = collection<User>("users");

const validate = <T>(f: (x: Arg<T>) => Query<boolean>) => (x: Arg<T>) =>
  iff(f(x), x, abort("invalid data")); // inferred as Query<T>

// register : (x: Arg<User>) => Query<Document<User>>;
const register = compose(
  validate(emailValidator),
  validate(passwordValidator),
  (x) => ({ data: x }),
  q.create(usersCollection)
);

const newUser = register({ email: 'info@cjol.dev', password: 'letmein' });
const insertedEmail = select(newUser, 'data', 'email');

With almost no type annotations, the transformations at each step of the way infer types such that the inferred type of insertedEmail is Query<string> (and executing the query gives an inferred result of string).

It was a fun project, and I think there's potential to expand on it but I think I bit off more than I can chew by trying to provide both solid typings and the function compositional style in one go. I am not sure that I can justify spending much more time on the project going forward, but I thought it was worth sharing here in case it serves as inspiration for anyone else!