@mb3/speck

1.5.0 • Public • Published

Speck

Build Status Code Climate Test Coverage

A simple parser to generate JSON from leading block comments.

See Atom Speck Language for a language package that does syntax highlighting within comments for speck.

Usage

CLI

    Usage: speck [options] -i [dir]

    Options:

      -h, --help             output usage information
      -V, --version          output the version number
      -i, --input [dir]      Input Directory
      -c, --config [config]  The relative path to a directory where a speck.json config file can be found
      -o, --output [dir]     Output directory
      -t, --types [values]   A list of all file types to parse separated by comma[,]
      --validate             Validate speck files
      --php                  Use php parser.

speck.json

You can use speck with a config file like the one below. It can be used to ignore paths by globs and store all other options for a simpler CLI execution. Speck will search the CWD for a speck.json file or you can specify which directory to look in with the -c or --config option.

{
    "ignore": [
        "!**/src{/,*}**",
        "**/*.spec.js"
    ]
}

Generating Jest Test Shells

Input

/* speck
  'name': 'JestSpeckPlugin'
  'extends': 'SpeckPlugin'
  'type': 'Plugin'
  'description': 'Generates a Jest test shell with the given interactions.'
  'interactions': [
    'can parse interactions from a file.',
    'can generate a test shell from a file.',
    'can append to a file cleanly.',
    'can create a new test file cleanly.'
  ]
*/
export class JestSpeckPlugin extends SpeckPlugin {
  // ...
}

Output

import {JestSpeckPlugin} from '../relative/path/to/jest.plugin.js';

describe('JestSpeckPlugin', () => {
  it('should parse interactions from a file.', () => {
    fail('Not Implemented');
  });

  it('should generate a test shell from a file.', () => {
    fail('Not Implemented');
  });

  it('should append to a file cleanly.', () => {
    fail('Not Implemented');
  });

  it('should create a new test file cleanly.', () => {
    fail('Not Implemented');
  });
});

Example

Input:

/* speck
  'class name': 'Comment'
  'description': 'Parses a comment from within a file.'
  'comment structure': 'Block comments should be written in cson'
  'example usage':
     'js': '''let c = new Comment('@title: hello');
       c.parse().toJSON(); // { title: 'hello' }'''
     'markdown': '''# Speck.Comment
       ## Features
       + Multiline
       + Json (nested is also suppoted)
       + Pluggable'''
  'multi line': '''This is a test
   for multiline comments'''
  'list items': [
   0,1,2,3,4
  ]
 */

 export class SomeJSClass {
   // ...
 }

Output:

{
    "class name": "Comment",
    "description": "Parses a comment from within a file.",
    "comment structure": "Block comments should be written in cson",
    "example usage": {
        "js": "let c = new Comment('@title: hello');\nc.parse().toJSON(); // { title: 'hello' }",
        "markdown": "# Speck.Comment\n## Features\n+ Multiline\n+ Json (nested is also suppoted)\n+ Pluggable"
    },
    "multi line": "This is a test\nfor multiline comments",
    "list items": [
        0,
        1,
        2,
        3,
        4
    ]
}

Extending

This repo is a base for the parser, it will take any block comments and generate JSON output. The base class doesn't ship with an implementation and will fail if run as follows:

import Speck from 'speck';

const parser = new Speck(...);

parser
  .gather()
  .parse(); // Throws an error.

However, it does ship with a Speck implemented using the babel parser babylon, this is the default when running the CLI:

import BabelSpeck from 'speck';

const parser = new BabelSpeck(...);

parser
  .gather()
  .parser(); // Generates JSON from ES6 code.

It also ships with a custom php tokenizer to parse leading comments.

To use Speck with other languages to generate JSON from leading block comments you must implement the getComments method.

Example

import Speck from 'speck';

export default BabelSpeck extends Speck {
  getComments(code) {
    // Parse out, and return, leading comments from code.

    return babylon.parse(code, OPTIONS)
      .comments
      .filter(comment => comment.type === 'CommentBlock')
      .map(comment => comment.value);
  }
}

Plugins

Plugins for speck are simple, all you need to do is export a class with a run function that takes three parameters:

  • logger
  • file: the source file path
  • json: the json results from the source file speck comments
export default class ExmplePlugin {
    run(logger, file, json): void {
        // ...
    }
}

If you need to pass options to the plugin at runtime, you can specify the configuration in the speck.json config file as an object that will be passed to the constructor of the plugin.

export default class ExamplePlugin {
    constructor(options = {}) {
        // ...
    }
}

All plugins are run after the base parsing of speck has been complete.

To see an example of how to write a Speck Plugin see MB3-jest-speck-plugin

Development

  • to test yarn test
  • to build yarn run build -> this will also run the tests
  • to run yarn start -- [options] OR ./script/speck.js [options] OR yarn link; speck [options]

Environment Setup

  • make sure you have node version > 6.2.x (n is an easy way to manage your node on MAC only.) run : yarn add -g n; sudo n stable

  • make sure you have php 7 (brew update; brew install homebrew/php/php70)

  • checkout the repo and yarn install; yarn add -g gulp; yarn add -g jasmine

Readme

Keywords

none

Package Sidebar

Install

npm i @mb3/speck

Weekly Downloads

0

Version

1.5.0

License

MIT

Last publish

Collaborators

  • sdavey
  • berniehillier