spoken

Spoken

This module is capable of parsing english and portuguese phrases using a stack automata based mechanism. The phrases are parsed to commands, that can be executed later.

The diagram below illustrates how this module works:

How to use it

You can learn how to use this module with the tests under ./src/tests.

The basic usage is as follows:

import Spoken from 'spoken'

// command 'variableAssigment'
const result = Spoken.recognizePhrase('new constant doctorWho equals number 42', 'en-US')

// true
_.isEqal(result, {
    comand: 'variableAssigment',
    args: {
        memType: 0, // 0 == 'const', 1 == 'let'
        varName: 'doctorWho',
        value: {
            command: 'number',
            args: { value: 42 },
            impl: "function(args, editor, common) {...}"
        }
    },
    impl: "function(args, editor, common){
        // something like: <args.memType> <args.varName> = <args.value>
        ...
    }"
})

How to add new commands

A command is a folder with 2 archives: impl.ts (what will happen when this command gets executed) and a phrase_en-US.dot (an automata responsible for recognizing english phrases that will trigger this command).

An example of a simple command can be found in: ./src/modules/typescript/write

You can view a visual representation of .dot files using VSCode extension named Graphviz Preview

The README.md of each command is generated automatically along with its png image representation with the command: npm run build-docs.

Useful Scripts

1. npm run build to create a file called grammar.json. Every time a command is edited, removed or added this should be done otherwise the changes will have to effect.

2. npm run test to run all the tests.

3. npm run build-docs to build the documentation of each command, the documentation is built using the .dot files. It also generates a png representations of the automato.