Skip to content

Latest commit

 

History

History
263 lines (199 loc) · 7.72 KB

README.md

File metadata and controls

263 lines (199 loc) · 7.72 KB

markdown-inject

Add file or command output to markdown documents.

Functionality Demo

Installation

markdown-inject is written in TypeScript and distributed as a node module on the npm ecosystem. It exposes a bin executable, making it a command line offering.

Download and invoke in one command:

npx markdown-inject

Local npm installation:

npm install markdown-inject --save-dev

or with Yarn:

yarn add markdown-inject --dev

Optionally, wire up markdown-inject to a git pre-commit hook tool like husky to automatically update markdown injection as part of your workflow.

Usage

Note: markdown-inject takes no action during pull request builds in CI.

Usage: markdown-inject [options] <glob pattern | -a>

Examples:
  $ npx markdown-inject -a
  $ npx markdown-inject 'README.md'
  $ npx markdown-inject './**/*.{md,mdx}'

Add file or command output to markdown documents.

Options:
  -v, --version                   output the version number
  -a, --all                       applies a globPattern of './**/*.md'
                                  (default: false)
  -b, --block-prefix <prefix>     specifies the prefix for START and END HTML
                                  comment blocks (default: "CODEBLOCK")
  -n, --no-follow-symbolic-links  prevents globs from following symlinks
  -q, --quiet                     emits no console log statements (default:
                                  false)
  -e, --no-system-environment     prevents "command"s from receiving system
                                  environment
  -h, --help                      display help for command

markdown-inject expands a given glob for markdown files. Then it discovers the below CODEBLOCK HTML or MDX (as shown in CONTRIBUTING.mdx) comments within each markdown file, performs the appropriate action (in this case, reading another local file), and writes content back into the markdown file:

<!-- CODEBLOCK_START {"value": ".nvmrc"} -->
<!-- CODEBLOCK_END -->
<!-- CODEBLOCK_START {"value": ".nvmrc"} -->
<!-- prettier-ignore -->
~~~~~~~~~~bash
File: .nvmrc

v20.10.0
~~~~~~~~~~

<!-- CODEBLOCK_END -->

Output is written between the CODEBLOCK_START and CODEBLOCK_END comments. Output includes:

  • A prettier ignore comment introducing the output so that prettier does not further alter existing code.
  • A markdown codeblock is opened with the language specified via configuration.
  • The <type>: <value> line is included by default, labeling the output.
  • The command or file output.

Executing commands follows a similar syntax:

<!-- CODEBLOCK_START {"value": "echo hello world", "type": "command"} -->
<!-- prettier-ignore -->
~~~~~~~~~~bash
$ echo hello world

hello world
~~~~~~~~~~

<!-- CODEBLOCK_END -->

You can hide the <type>: <value> comment from the generated output too:

<!-- CODEBLOCK_START {"value": "echo hello world", "type": "command", "hideValue": true} -->
<!-- prettier-ignore -->
~~~~~~~~~~bash
hello world
~~~~~~~~~~

<!-- CODEBLOCK_END -->

Environment

System environment is automatically passed to commands:

$ echo "My home directory is: $HOME"

My home directory is: /Users/me

In some scenarios, passing system environment to commands may be undesirable. This functionality can be disabled using the --no-system-environment CLI flag. This creates output such as:

$ echo "My password is: $MY_PASSWORD"

My password is:

Sometimes commands need a little extra nudging via environment to receive a usable output. Environment variables can be added using the environment key:

<!--
  CODEBLOCK_START
  {
    "type": "command",
    "value": "node ./deploy.js --dry-run",
    "environment": {
      "SKIP_PRECHECK": "true"
    }
  }
-->
<!-- CODEBLOCK_END -->

The environment key can also be used to overwrite system environment variables with example values:

<!--
  CODEBLOCK_START
  {
    "type": "command",
    "value": "echo \"My password is: $MY_PASSWORD\"",
    "environment": {
      "MY_PASSWORD": "<insert password here>"
    }
  }
-->
<!-- prettier-ignore -->
~~~~~~~~~~bash
$ echo "My password is: $MY_PASSWORD"

My password is: <insert password here>
~~~~~~~~~~

<!-- CODEBLOCK_END -->

Environment variables with values which follow bash variable naming rules will be substituted into the command environment whether or not --no-system-environment is enabled. This can be useful for re-introducing necessary environment variables that would be omitted by --no-system-environment:

<!--
  CODEBLOCK_START
  {
    "type": "command",
    "value": "echo \"My home directory is: $HOME\nMy password is: $MY_PASSWORD\"",
    "environment": {
      "HOME": "$HOME"
    }
  }
-->
<!-- prettier-ignore -->
~~~~~~~~~~bash
$ echo "My home directory is: $HOME
My password is: $MY_PASSWORD"

My home directory is: /Users/me
My password is:
~~~~~~~~~~

<!-- CODEBLOCK_END -->

Codeblock Configuration

The CODEBLOCK_START HTML comment config block has the following properties:

Name Type Required Default Description
value string true Command to execute or file to retrieve
environment object false {} Run command executions with the given environment values
hideValue boolean false false Do not display File: foo.js or $ npx foo on the first line
language string false commandbash, file: File extension Syntax highlighting language
trim boolean false true Trim whitespace from the ends of file or command output
type 'command' | 'file' false 'file' Type of execution

Contributing

See CONTRIBUTING.md for more information.

Similar Projects

  • embedme - embed source files into markdown code blocks
  • mdsh - a similar tool but for the Rust ecosystem