gulp-markdown-to-json
Parse Markdown + YAML, compile Markdown to HTML, wrap it in JSON.
The neutral format of JSON opens new possibilities for your handcrafted content. Send it onward to other plugins such as gulp-hb. When their powers combine, you get a static site generator! Pipe to request to send it to a search index or import into a CMS. Write a plugin to tap into the stream if you need a client library.
npm install gulp-markdown-to-json --save-dev
This plugin does not bundle any Markdown parser to keep your options open. BYOM!
Pick one of these or write a new parser for fun!
- commonmark.js
- markdown-js
- markdown-it ← the best
- marked
- remark
- remarkable
Install, configure, and pass a rendering method to this plugin with a source string as its first argument. Output goes straight into the JSON file’s body
property. If your parser requires instantiation pass a context to call it with by defining context
in the config object. If yet more hoop jumping is required, write a wrapper function such as this example for remark:
function render (string) {
const remark = require('remark');
const html = require('remark-html');
return remark().use(html).process(string).toString();
}
Note: YAML frontmatter blocks are stripped and handled before Markdown rendering with front-matter
/gulpfile.js
const gulp = require('gulp');
const markdownToJSON = require('gulp-markdown-to-json');
const marked = require('marked');
marked.setOptions({
pedantic: true,
smartypants: true
});
gulp.task('markdown', () => {
gulp.src('./content/**/*.md')
.pipe(markdownToJSON(marked))
.pipe(gulp.dest('.'))
});
Transformed source files flow onward to the destination of your choice with directory structure preserved. Any valid JSON files matched by your gulp.src
glob passthrough.
/blog/posts/bushwick-artisan.md
---
slug: bushwick-artisan
title: Wes Anderson pop-up Bushwick artisan
layout: centered
---
## YOLO
Chia quinoa meh, you probably haven't heard of them sartorial Holowaychuk pickled post-ironic. Plaid ugh vegan, Sixpoint 8-bit sartorial artisan semiotics put a bird on it Mission bicycle rights Club-Mate vinyl.
/blog/posts/bushwick-artisan.json
{
"slug": "bushwick-artisan",
"title": "Wes Anderson pop-up Bushwick artisan",
"layout": "centered",
"body": "<h2 id=\"yolo\">YOLO</h2>\n<p>Chia quinoa meh, you probably haven't heard of them sartorial Holowaychuk pickled post-ironic. Plaid ugh vegan, Sixpoint 8-bit sartorial artisan semiotics put a bird on it Mission bicycle rights Club-Mate vinyl.</p>",
"updatedAt": "1970-01-01T00:00:00Z"
}
Gather Markdown files before piping with list-stream to combine output into a single JSON file. Directory structure is preserved and represented as nested JSON for iteration with Handlebars.js and friends. This is handy for navigation and other global content. Valid JSON files are included in the object if matched by your gulp.src
glob.
The consolidated file is named content.json
by default and optionally renamed.
const gulp = require('gulp');
const ListStream = require('list-stream');
const markdownToJSON = require('gulp-markdown-to-json');
const marked = require('marked');
gulp.task('markdown', () => {
gulp.src('./content/**/*.md')
.pipe(ListStream.obj())
.pipe(markdownToJSON(marked, 'blog.json'))
.pipe(gulp.dest('.'))
});
blog.json
{
"blog": {
"blog": {
"title": "ipsum dipsum",
"body": "<p>From west to \"east\"!</p>",
"updatedAt": "1970-01-01T00:00:00Z"
},
"posts": {
"bushwick-artisan": {
"slug": "bushwick-artisan",
"title": "Wes Anderson pop-up Bushwick artisan",
"layout": "centered",
"body": "<h2 id=\"yolo\">YOLO</h2>\n<p>Chia quinoa meh, you probably haven't heard of them sartorial Holowaychuk pickled post-ironic. Plaid ugh vegan, Sixpoint 8-bit sartorial artisan semiotics put a bird on it Mission bicycle rights Club-Mate vinyl.</p>",
"updatedAt": "1970-01-01T00:00:00Z"
}
}
},
"mission": {
...
}
}
Specify flattenIndex: true
in the config object to unwrap home page/index-style content and merge it into the parent object. Name these files index
or the same as a parent directory.
{
"blog": {
"title": "ipsum dipsum",
"body": "<p>From west to \"east\"!</p>",
"updatedAt": "1970-01-01T00:00:00Z",
"posts": {
...
}
},
"mission": {
...
}
}
This avoids redundant-feeling blog.blog
scenarios when iterating and selecting from this content.
Define titles as title
in the YAML frontmatter. Text of the first <h1>
is assigned to title
automatically if this is not specified.
Specify stripTitle: true
in the config object to remove the first <h1>
from the body. Use this if you are displaying the title outside of the body, in a page header for example.
/blog/posts/bushwick-artisan.md
Wes Anderson pop-up Bushwick artisan
====================================
## YOLO
Chia quinoa meh, you probably haven't heard of them sartorial Holowaychuk pickled post-ironic. Plaid ugh vegan, Sixpoint 8-bit sartorial artisan semiotics put a bird on it Mission bicycle rights Club-Mate vinyl.
/blog/posts/bushwick-artisan.json
{
"title": "Wes Anderson pop-up Bushwick artisan",
"body": "<h2 id=\"yolo\">YOLO</h2>\n<p>Chia quinoa meh, you probably haven't heard of them sartorial Holowaychuk pickled post-ironic. Plaid ugh vegan, Sixpoint 8-bit sartorial artisan semiotics put a bird on it Mission bicycle rights Club-Mate vinyl.</p>",
"updatedAt": "1970-01-01T00:00:00Z"
}
To change or add to the JSON data for each file, specify a transform function and return your desired object. This function is passed the default data and the Vinyl file object for the source file.
For example:
gulp.src('./content/**/*.md')
.pipe(ListStream.obj())
.pipe(markdownToJSON(marked, 'blog.json', (data, file) => {
delete data.body;
data.path = file.path;
return data;
}))
.pipe(gulp.dest('.'))
markdownToJSON((renderer: Function, name?: String, transform?: Function) | config: Object) => TransformStream
config
renderer
Function
accepts Markdown source string, returns an escaped HTML string. Requiredcontext
Object
to use when callingrenderer
name
String
to rename consolidated output file, if using. Default:content.json
flattenIndex
Boolean
unwrap files namedindex
or after parent dirs in consolidated output. Default:false
stripTitle
Boolean
strips the first<h1>
from body, if extracted as title. Default:false
transform
Function
to access and change the JSON data for each file before outputting
Pull requests accepted!
MIT
Copyright © 2017 Sparkart Group, Inc.