turtleblog v0.0.0-dev.12

turtleblog 🐢

expected site directory structure

source/
	articles/ # contains page dirs
	blog-index/ # is a page dir
	blog-posts/ # contains dated page dirs
	home/ # is a page dir
	templates/
		article.pug
		blog-index.pug
		blog-post.pug
		home.pug

  concept blog generator

here's one way

turtleblog \
  --source source \
  --dist dist \
  --blog devblog

concept walkthrough

• turtle-scan: scan website structure, assign ids, make navigation and links

  	```
  	interface WebsiteData {
  		articles: PageData[]
  		blog: PageData[]
  		blogIndex: PageData
  		home: PageData
  		navigation: [
  			{
  				label: "devblog"
  			}
  		]
  	}
  	```

• turtle-generate: read the files and perform compilations

  	```
  	interface WebsiteOutput {
  		articles: PageOutput[]
  	}
  	```

• turtle-write: output all results to disk and copy files

turtleblog is based on three simple functions

1. turtle-read — read a source directory, return website metadata json
2. turtle-generate — compile pug, sass, and return website output json
3. turtle-write — save the website to disk

turtleblog via command line

#!/bin/bash

# turtleblog tools pipe together

turtle-read | turtle-generate | turtle-write
 #          ^                 ^
 #    {metadata json}    {output json}

# there are two points where the tools pass json data,
# easy opportunity to transform that json yourself

# turtleblog is flexible, so you can tweak arguments and add transformers

turtle-read --source src           # load website source
  | turtle-transform-metadata-pass # (optional) transform metadata
  | turtle-generate                # compile pug, sass, etc
  | turtle-transform-output-pass   # (optional) transform output
  | turtle-write --dist dist       # write website to disk

# there are two types of transformers you can add
#  1. metadata transfomers change the json data before compilations occur
#  2. output transformers change the json before writing to disk

# turtle-read's default source directory is "source"
# turtle-write's default dist directory is "dist"

turtleblog via node

import {turtleRead, turtleGenerate, turtleWrite} from "turtleblog"

async function main() {

	// 1. read the website
	const websiteMetadata = await turtleRead({source: "src"})

	// 2. generate the website
	const websiteOutput = await turtleGenerate({websiteMetadata})

	// 3. write the website
	await turtleWrite({dist: "dist", websiteOutput})
}

turtle-read's expected website source directory structure

{source-dir}/

  layouts/
    page.pug
    blog-index.pug
    blog-post.pug

  pages/
    {page-name}/  # standard page directory
      {section-name}.md  # sections are passed to pug layouts
      {section-name}.md
      ...{any other files are copied}
    {page-name}/  # standard page directory
      {section-name}.md
      {section-name}.md
      {page-name}/  # sub-page directory
        {section-name}.md
        {section-name}.md

  blog/
    index/  # standard page directory
      {section-name}.md
      {section-name}.md
    posts/
      {year}/
        {month}/
          {day}/
            {page-name}/  # standard page directory
              {section-name}.md
              {section-name}.md

roll your own turtle

turtleblog exposes its typescript interfaces, so you extend turtleblog and make your own interoperable replacement tools

of course, you could wholly replace turtle-read, if you wanted to have control over your directory structure

also note you could wholly replace turtle-generate, for control over the compilation process

pug locals

• all pages - pageMetadata - websiteMetadata

• articles - articleMetadata

• blog posts - blogPostMetadata

• blog index - blogIndexMetadata