Data folder

[sc0ttj]

An early attempt at solving https://github.com/sc0ttj/mdsh/issues/65

What's the improvement?

New feature: the assets/data folder

Site-wide data: Put .yml, .csv, or .sh files in the assets/data/ folder, and the data they contain will be available to all templates at build time.

Page specific data: Put .yml, .csv, or .sh files in the assets/data/my-cool-page/ folder, and the data they contain will be available to the page my-cool-page only at build time.

Details

• less code, more config
  • site data now defined in YAML (assets/data/site.yml)
  • page/post data defined in YAML in frontmatter (in top of .mdsh files)
• better separation of build data and templates
  • only get site data from assets/data folder (not .site_config)
  • only get post/page data from yaml frontmatter or assets/data
  • fewer custom functions used by templates to generate data, more arrays
• nicer file structure
  • fewer files, simpler file naming/layout
  • changed various scripts into functions
• cleaner templates
  • removed all shell funcs in templates, replaced with:
    • site/page arrays accessed via the foreach iterator
    • better partial handling

Other new features

The foreach iterator

• adds a foreach iterator to the templating system

Easier data debugging

• added a DEBUG_DATA env var, set to true to see site data printed in terminal during build

page_layout variable

Easily render pages with different layouts

Define it in the page front matter, or pass it in via CLI when running create_page.sh (defaults to main if empty, and must be the name of a template file in .app/templates/*/)..

Fixes

• added .nojekyll file, used by GitHub to help identify Jekyll sites (or not) ebc78c3

Implementation of data folder

Added example data files:

site.yml, products.yml, people.csv, arrays-example.sh

Pull in data at build time:

Added a .yml parser & csv parser func, updated cli-interface to parse data files on each build:

Updates in .app.cli-interface.bash to scan the assets/data dir for files, and parse them into vars and arrays that our build scripts and templates can use...

YAML parsers used:

• .app/functions/yay - custom version of yay with added stuff:
  • fixes and added support for multiple nested levels
  • fixes for handling simple lists, and other YAML syntaxes
  • TLDR: better at creating the associative arrays our templates need
• .app/functions/ysh - a parser called yaml.sh
  • used only inside the lookup method
  • the lookup method only calls ysh if no site data vars/arrays found
  • else it returns the matching values from the vars/arrays

CSV parser:

• shell function called csvToData, in .app/functions/liquid_filters.bash
• turns CSV data into bash associative arrays that our templates can use

Much easier to use arrays in templates:

Added a foreach iterator function, to be used in the templates, small fix to mo script to make it work OK.

An example!!

1. assets/data/products.yml contains:

  product1:
    name: Foo
    price: 10
  product2:
    name: Bar
    price: 20

2. so you access it in your templates like so:

  {{#foreach product in products}}
    {{product.name}} costs {{product.price}}
  {{/foreach}}

TODO:

• test foreach iterator with mutiple filters
• test page specifc data