array-of-arrays-into-ast

Purpose

It consumes array of arrays and produces a trie-like AST from them. This library was a piece of a one experimental code generator of ours.

Principles

Every object’s key will have a value of array.

• null inside that array means it’s the tip of the branch.

• An object inside that array means the branch continues.

API — generateAst()

The main function generateAst() is imported like this:

It is a function which takes two arguments: an array and an optional options object:

The optional options object has the following shape:

Here are all defaults in one place for copying:

The function returns a plain object, the assembled AST.

API — defaults

You can import defaults:

It's a plain object:

The main function calculates the options to be used by merging the options you passed with these defaults.

API — version

You can import version:

opts.dedupe

If you generate the AST with default settings, dedupe setting will be active and duplicate paths won’t be created:

import generateAst from "array-of-arrays-into-ast";
const res = generateAst([[1], [1], [1]]);
console.log(`${`\u001b[${33}m${`res`}\u001b[${39}m`} = ${JSON.stringify(res, null, 4)}`);
// res = {
//   1: [null]
// }

Now, see what happens when you turn off opts.dedupe:

import generateAst from "array-of-arrays-into-ast";
const res = generateAst([[1], [1], [1]], { dedupe: false });
console.log(
  `${`\u001b[${33}m${`res`}\u001b[${39}m`} = ${JSON.stringify(res, null, 4)}`
);
// res = {
//   1: [null, null, null]
// }
}

Notice how entries for each branch were created.

Generally, we don’t see the reason why you’d want duplicates, but the setting is there if you ever need it.

Compared vs. datastructures-js

There are libraries that produce and manage trie data structures, for example, datastructures-js. In particular case, the problem is, the data structure is abstracted behind the let trie = ds.trie(); and you can’t access it directly, traversing the nested tree of arrays and objects.

datastructures-js trie would limit to search(), traverse() and count() methods. However, we need to recursively traverse every node and look up and down, what’s around it.

Here’s where this library comes in. It doesn’t abstract the data it’s producing — you get a nested plain object which you can traverse and further process any way you like, using a vast ocean of object- processing libraries.