MD-Curcuma

The Data-Wrangler. Copys and transforms Markdown and CSV files from your Obsidian-Vault for usage in Hugo.

• This thing ist done with Typescript.
• Inspired by: https://github.com/accraze/split-md
• Why not golang? So you can use it easier outside the golang unsiverse.

Availible Transporter:

1. MD_Transporter (Markdown to Markdown) - for Markdown Files.
2. CSV_Transporter (CVS to Json) - For files with comma-separated values, where each line is a record. This is a BookBuddy App support: CSV-File conversion & Image Download.
3. XLSX_Transporter (XLSX to Json) - read, write and transform.

Availible Markdown Transformers:

1. Copying and transforming one or more Markdown files.
2. Split Markdown longform documents into several individual documents.
3. Obsidian-Links (Images and Documents), copy Files on the fly.
4. Latex-Formulas, Paragraph and Inline.
5. Frontmatter: Add, Replace, Map and Transform Values.
6. Callouts.
7. Remove Todos.
8. TODO: Take wikilinks (cross-references) into account. Footnotes Endnotes support.

Availible Mappings:

1. Adopt Value - a simple pass through example.
2. ArrayJoin, ArraySplit
3. Boolean Inverse
4. Image Downloader
5. Insert Date (now)
6. Insert UUID
7. Trim Characters from String.

You can build customized transformers and Mappers for different solutions. For planned features and changes take a look into the CHANGELOG.md

Restrictions

• It only runs in Backend, not in Browsers.
• Reading and writing is not arbitrary - (but should be).

How it works

flowchart
    subgraph Transporter
      subgraph Transformer
        Transformers
        Mappers
      end
     Reader --> Transformer --> Tasks --> Writer
    end

Loading

In short:

const my_transporter: MD_Transporter = new MD_Transporter();
my_transporter.perform_job_from("../test-data-obsidian-vault/transport-config.json", "Example-Job No.1");

All tasks are defined in the transport-config.json configuration file.

In long:

const my_transporter: MD_Transporter = new MD_Transporter();

// Basic config
const transporter_parameter: MD_Transport_Parameter_Type = {
  readPath: "test-obsidian-vault/longform.md",
  writePath: "test-hugo/hugo-content-2/",
  doSubfolders: false,
  limit: 1990,
  useCounter: false
};

// Transformer config
const parameter_images: MD_Transformer_Parameter_Type = {
  tag_obsidian_prefix: "![[",
  tag_obsidian_suffix: "]]",
  find_rule: "jpg|png",
  replace_template: `{{< image src="assets/images/{name_full}" >}}`,
  copy_task: {
    source:"test-obsidian-vault/images/",
    target:"test-hugo/hugo-content-2/assets/images/{name}/",
    simulate:false
  }
};

// Add the Transformer to the pipeline
my_transporter.addTransformer(new MD_ObsidianLink_Transformer(parameter_images));

// execute the job
my_transporter.perform_job(transporter_parameter);

Take a look in the test folder for running examples.

Transformations

I already have a number of transformations on offer, but if you need your own, you can and add them to the pipeline.

1. Split a Longform-Markdown File

The Problem:

• I have a long Markdown document that was compiled with the Obsidian Longform plugin.
• I would like to make this available on my website.
• The website is built with Hugo.
• For this purpose, the longform document is to be split into several individual documents.
  • The splitting should be done on headings: e.g.: #
  • The text of the heading is used as the filename
  • If necessary, with a number in front, if necessary replace special characters url conform
• Be Hugo compatible.
  • The individual documents should be equipped with frontmatter.
  • autogenerate uuids
• Certain content may need to be transformed.

Code-Example 4

var splitter_frontmatter: MD_Frontmatter = new MD_Frontmatter(`---
title: "{title}"
description: ""
url: /docs/{url_prefix}-{url}/
date: {date}
draft: false
images: []
menu:
  docs:
    parent: "docs-9602b15bad02600f3883f55e2ade6b81"
    identifier: "{url_prefix}-{url}-{uuid}"
weight: {weight}
toc: true
---\n\n`);

const parameter_splitter: MD_Splitter_Parameter_Type = {
  pattern: "# ",
  cleanName: "# ",
  limit: 100,
  hasCounter: false,
  weightBase: 8000,
  url_prefix: "test-prefix",
  doRemoveHeadline: true,
  frontmatter_filename: "", // ./test/frontmatter-template.md
  frontmatter: splitter_frontmatter
};

my_transporter.addTransformer(new MD_Splitter_Transformer(parameter_splitter));

Example JSON for usage in config-file 1

{
  "transformer_class_name": "MD_Splitter_Transformer",
  "transformer_parameter": {
  "pattern": "# ",
  "cleanName": "# ",
  "limit": 100,
  "hasCounter": false,
  "weightBase": 8000,
  "url_prefix": "test-prefix",
  "doRemoveHeadline": true,
  "frontmatter_filename": "./test-data-obsidian-vault/frontmatter-template.md",
  "frontmatter": {}
}

2. Obsidian-Links

Images from

![[my-image.jpg]]

Hugo replacement:

 {{<image folder="images/my-image.jpg" >}}

or Dokuments from

![[docu-1.pdf]]

Hugo replacement:

{{< button href="/getthis.php?id=docu-1" name="download docu-1 (pdf)" >}}

Code-Example 5

const parameter_images: MD_Transformer_Parameter_Type = {
  tag_obsidian_prefix: "![[",
  tag_obsidian_suffix: "]]",
  find_rule: "jpg|png",
  replace_template: `{{< image src="assets/images//{name}/{name_full}" >}}`,
  copy_task: {
    source:"test-data-obsidian-vault/images/",
    target:"test-data-hugo/hugo-content-2/assets/images/{name}/",
    simulate:simulate_copy_job
  }
};

const parameter_docs: MD_Transformer_Parameter_Type = {
  tag_obsidian_prefix: "![[",
  tag_obsidian_suffix: "]]",
  find_rule: "pdf|ods|odp",
  replace_template: `{{< button href="/getthis.php?id={name}" name="download {name} ({name_suffix})" >}}`,
  copy_task: {
    source:"test-data-obsidian-vault/attachments/",
    target:"test-data-hugo/hugo-content-2/static/downloads/",
    simulate:simulate_copy_job
  }
};

my_transporter.addTransformer(new MD_ObsidianLink_Transformer(parameter_images));
my_transporter.addTransformer(new MD_ObsidianLink_Transformer(parameter_docs));

Example JSON for usage in config-file 2

{
  "transformer_class_name": "MD_ObsidianLink_Transformer",
  "transformer_parameter": {
    "tag_obsidian_prefix": "![[",
    "tag_obsidian_suffix": "]]",
    "find_rule": "jpg|png",
    "replace_template": "{{< lightbox-docs id=\"0\" folder=\"images/{name}/*\" showImageNr=0 >}}",
    "copy_task":{
      "source":"test-data-obsidian-vault/images/",
      "target":"test-data-hugo/hugo-content-1/assets/images/{name}/",
      "simulate":false
    }
  }
},
{
  "transformer_class_name": "MD_ObsidianLink_Transformer",
  "transformer_parameter": {
    "tag_obsidian_prefix": "![[",
    "tag_obsidian_suffix": "]]",
    "find_rule": "pdf|ods|odp",
    "replace_template": "{{< button href=\"/getthis.php?id={name}\" name=\"download {name} ({name_suffix})\" >}}",
    "copy_task":{
      "source":"test-data-obsidian-vault/attachments/",
      "target":"test-data-hugo/hugo-content-1/static/downloads/",
      "simulate":false
    }
  }
},

Copy images and documents on the fly

A simulation mode provides information about which images and attachments it expects and where.

3. Latex Formulas

Transform Latex Formulas from Obsidian-Style to Hugo-Style.

• https://www.makeuseof.com/write-mathematical-notation-obsidian/
• https://getdoks.org/docs/built-ins/math/

from this:

$$
W_{kin} = \frac { m \cdot v^2}{2} = \frac {p^2}{ 2 \cdot m}
$$

This is an inline ${(x+y)}^2$ Formula.

to this:

```math {.text-center}
$$
W_{kin} = \frac { m \cdot v^2}{2} = \frac {p^2}{ 2 \cdot m}
$$
```

This is an inline {{< math >}} ${(x+y)}^2$ {{< /math >}} Formlula.

Code-Example 6

var parameter_math_paragraph: MD_Transformer_Parameter_Type = {
  tag_obsidian_prefix: "$$",
  tag_obsidian_suffix: "$$",
  find_rule: "",
  replace_template: "```math {.text-center}\n$$\n {content} \n$$\n```\n",
};

var parameter_math_inline: MD_Transformer_Parameter_Type = {
  tag_obsidian_prefix: "$",
  tag_obsidian_suffix: "$",
  find_rule: "",
  replace_template: "{{< math >}} ${content}$ {{< /math >}}",
};

my_transporter.addTransformer(new MD_MathParagraph_Transformer(parameter_math_paragraph));
my_transporter.addTransformer(new MD_MathInline_Transformer(parameter_math_inline));

Example JSON for usage in config-file 3

{
  "transformer_class_name": "MD_MathParagraph_Transformer",
  "transformer_parameter": {
  "tag_obsidian_prefix": "$$",
  "tag_obsidian_suffix": "$$",
  "find_rule": "",
  "replace_template": "```math {.text-center}\n$$\n {content} \n$$\n```\n"
  }
},
{
  "transformer_class_name": "MD_MathInline_Transformer",
  "transformer_parameter": {
  "tag_obsidian_prefix": "$",
  "tag_obsidian_suffix": "$",
  "find_rule": "",
  "replace_template": "{{< math >}} ${content}$ {{< /math >}}"          
  }
},

4. Frontmatter: Add, Replace, Map and Transform Values

What happens here?

Adds a frontmatter to a single file (or a batch of single files).

• The frontmatter from the source file is removed.
• The new frontmatter-template is inserted instead.
• Take some frontmatter fields from the source file,
  • and write them to another field in the target file.
  • The value can be converted on the way.

It's not quite finished yet.

Code-Example 1

var document_frontmatter: MD_Frontmatter_Template =
  new MD_Frontmatter_Template(`---
title: ""
description: ""
summary: ""
date:
draft:
weight: 
categories: []
tags: []
contributors: []
pinned: false
homepage: false
seo:
  title: "" 
  description: ""
  canonical: ""
  noindex: false
---\n\n`);

// use one of the predefined tasks like so:
// task: new MD_Mapping_BooleanInverse_Task()
// or write a custom task:
const map_1: MD_Mapping = {
  mapping_items: [{
    source_property_name: "doPublish",
    target_poperty_name: "draft",
  }],
  task: {
    perform: function (mapping_properties: MD_MappingTask_Properties): boolean {
      let target_value = !mapping_properties.source_value;
      return target_value;
    },
  },
};

// An example task that inserts the current date. Source Property isnt used here.
const map_2: MD_Mapping = {
  mapping_items: [{
    source_property_name: "",
    target_poperty_name: "date",
  }],
  task: {
    perform: function (mapping_properties: MD_MappingTask_Properties): any {
      return new Date().toJSON().slice(0, 16);
    },
  },
};

const parameter_frontmatter: MD_Frontmatter_Parameter_Type = {
  frontmatter: document_frontmatter,
  frontmatter_filename: "",
  mappings: [map_1, map_2],
};

my_transporter.addTransformer(new MD_Frontmatter_Transformer(parameter_frontmatter));

Example JSON for usage in config-file 4

{
  "transformer_class_name": "MD_Frontmatter_Transformer",
  "transformer_parameter": {
  "frontmatter_filename": "./test-data-obsidian-vault/frontmatter-template.md",
  "frontmatter": {}
          }
},

5. Callouts

Transform Obsidian-Callouts to Hugo-Callout-Shortcodes.

• Obsidian Callouts https://help.obsidian.md/Editing+and+formatting/Callouts
• Hugo Callouts https://getdoks.org/docs/basics/shortcodes/

I'm not quite finished here either.

from

> [!info] Custom Title
> Here's a callout block.
> It supports **Markdown**, [[Internal link|Wikilinks]], and [[Embed files|embeds]]!
> ![[Engelbart.jpg]]

• note
• abstract(summary, tldr)
• info
• todo
• tip (hint, important)
• success(check, done)
• question (help, faq)
• warning (caution, attention)
• failure (fail, missing)
• danger (error)
• bug
• example
• quote (cite)

to

{{ < callout context="tip" title="Custom Title" icon="rocket" > }}
 Here's a callout block.
{{ < /callout > }}

• context="tip" title="Tip" icon="rocket"
• context="note" title="Note" icon="info-circle"
• context="caution" title="Caution" icon="alert-triangle"
• context="danger" title="Danger" icon="alert-octagon"

Code-Example 2

var parameter_callouts: MD_Transformer_Parameter_Type = {
  tag_obsidian_prefix: "> [!",
  tag_obsidian_suffix: "]",
  find_rule: "",
  replace_template: `{{< callout context="{context}" title="{title}" icon="{icon}" > }} {content} {{< /callout >}}`,
};

my_transporter.addTransformer(new MD_Callout_Transformer(parameter_callouts));

Example JSON for usage in config-file 5

{
  "transformer_class_name": "MD_Callout_Transformer",
  "transformer_parameter": {
    "tag_obsidian_prefix": "> [!",
    "tag_obsidian_suffix": "]",
    "find_rule": "",
    "replace_template": "{{< callout context=\"{context}\" title=\"{title}\" icon=\"{icon}\" > }} {content} {{< /callout >}}"
  }
},

6. Remove TODOs

Remove for example the following Paragraphes:

- [ ] #TODO Some serious stuff to do...

Code-Example 3

const parameter_remove: MD_Transformer_Parameter_Type = {
  tag_obsidian_prefix: "",
  tag_obsidian_suffix: "",
  find_rule: "- [ ] #TODO ",
  replace_template: ``,
};

my_transporter.addTransformer(new MD_RemoveTODOS_Transformer(parameter_remove));

Example JSON for usage in config-file 6

{
  "transformer_class_name": "MD_RemoveTODOS_Transformer",
  "transformer_parameter": {
    "find_rule": "- [ ] #TODO ",
    "replace_template": ""
  }
},

7. Wikilinks / Cross-References

This isnt done yet.

8. Footnotes Endnotes support

Works for footnotes inside the splitted chapter, but not yet if they are outside.

9. BookBuddy-App, CSV Support

• The BookBuddy App exports its contents as a csv file.
• I would like to use the data in Hugo.
• To do this, I convert the csv to json and download the also exported images from URL.
• The Property Cover_Image saves the local Path to the Image for Hugo.

let image_download_mapping_props: MD_ImageDownloader_MappingType = {
  image_target_folder: "test-data-hugo/hugo-content-4/assets/images/",
  image_hugo_path: "images/", 
  filename_property_name: "UUID", // not used by now
  simulate: false
};

const image_download_mapping: MD_Mapping = {
  mapping_items: [{
    source_property_name: "Uploaded_Image_URL",
    target_poperty_name: "Cover_Image",
  }],
  task: new MD_ImageDownloader_Mapping(image_download_mapping_props),
};

const csv_transporter_parameter: CSV_Transporter_Parameter_Type = {
    readPath: "test-data-obsidian-vault/attachments/bookbuddy-export.csv",
    writePath: "test-data-hugo/hugo-content-4/data/bookbuddy-export.json",
    csvSeparator: ',',
    mappings: [image_download_mapping],
};

CSV_Transporter.transform_to_json(csv_transporter_parameter);

Mappings

Adopt Value - a simple pass through example

ArrayJoin, ArraySplit

Boolean Inverse

Image Downloader

Insert Date (now)

Insert UUID

Trim Characters from String

Build a Custom Mapper

For fast prototyping, it can be useful to set up a mapper in this way:

const map_1: MD_Mapping = {
  mapping_items: [
    {
      source_property_name: "spn_test", // read from
      target_poperty_name:  "tpn_test", // write to
    }
  ],
  task: {
    perform: function (mapping_properties: MD_MappingTask_Properties): boolean {
      // The values are also made available to you, 
      // so that you no longer have to read them from the properties:
      let sv:boolean = mapping_properties.source_value; // type any
      let tv:boolean = mapping_properties.target_value; // type any
      // in normal cases you would work with these given values
      // and return the target_value at the end.

      // you also have access to the source Object, and the Target Object.
      // possibly source and target are the same.
      let the_souce:any = mapping_properties.source;
      let the_target:any = mapping_properties.target;

      // and you have access to the names of the preoccessed Properties:
      // possibly they are the same.
      let spn:string = mapping_properties.source_property_name;
      let tpn:string = mapping_properties.target_poperty_name;

      // do your transformations and mappings
      // and return the target_value (which is assigned automattically)
      return !sv;
    },
  }
};

In other cases it my be better to build a real class because you can pass additional parameters. For example take a look at MD_ArraySplit_Mapping

Install and Use

Where to use?

The Transport-Scripts can be used...

1. in the Hugo-Project - and pull from Obsidian Vault.
2. in Obsidian - and push to your Hugo project. I will write an Obsidian-Plugin for that in near future.
3. as stand allone project between Obsidian and Hugo - so it does pulling and pushig.

I recommend Option 3 at the moment.

Install it

Install-Method 1: Clone the Repository

The easiest way is to clone the repository. You will then also receive the tests, giving you a good starting point for your own adjustments.

• copy the link: https://gitlab.com/glimpse-of-life/md-curcuma.git
• open VS Code
• Click on Clone Git repository.
• Select a directory to clone the target to.
• The directory md-curcuma will be created.
• Open the project in VS-Code
• Open Teriminal
• npm install
• npm run build:ts
• npm run test

Install-Method 2: create new Project

Another option is to create an empty project, an install from npm...

• create a project-folder
• open that folder, and in it...

npm init
# follow the guide
npm install md-curcuma
# check typescript installation
tsc --version

create a tsconfig.json

If you use it inside your Hugo Project, some folders are excluded per default.

{
  "compilerOptions": {
    "module": "commonjs",
    "target": "ES5",
    "noImplicitAny": true,
    "removeComments": true,
    "preserveConstEnums": true,
    "sourceMap": false,
    "baseUrl": "./",
    "rootDir":"transport-scripts",
    "outDir": "transport-scripts-compiled"
  },
  "include": ["transport-scripts/**/*"],
  "exclude": [
    "assets",
    "config",
    "content",
    "data",
    "layouts",
    "node_modules",
    "resources",
    "static",
    "transport-scripts-compiled"
  ]
}

update from npm

# check for updates
npx npm-check-updates

# upgrade package.json
npx npm-check-updates -u
# or 
npx npm-check-updates -u --interactive --format group

# install the new Versions
 npm update --save
# or 
 npm update --save-dev
# and then
 npm install

Create your Skript

Look in the test folder for running examples.

Create two folders for your scripts:

• transport-scripts - holds the typescript source-code
• transport-scripts-compiled

The first is for the Typescript source-code, the second is for the Java script that will later be compiled from it.

Compile to javascript

To simplify handling, include the following build commands in the package.json:

The ts suffix in build:ts separates it from the golag scripts which i name build:go, but of course you can call them whatever you like.

    "build:ts": "rm -rf transport-scripts-compiled && tsc",
    "watch:ts": "tsc --watch",

Create The Example Markdown-Files

• create folder test, and in folder test
  • create folder test/obsidian-fault/ - an put your longform.md in it.
  • create folder test/hugo-content/ - this is the target directory

Create a Transport-Script

In folder transport-scriptscreate the file split-my-longform.ts and paste the code:

const my_transporter: MD_Transporter = new MD_Transporter();
my_transporter.perform_job_from("../test-data/transport-config.json", "Example-Job No.1");

That is all was needed when you provide a config file and a frontmatter file.

You can do it the other - the long - way:

import {
  MD_Frontmatter_Template,
  MD_ObsidianLink_Transformer,
  MD_RemoveTODOS_Transformer,
  MD_Math_Transformer,
  MD_Splitter_Parameter_Type,
  MD_Transformer_Parameter_Type,
} from "md-curcuma";
import {
  MD_Transporter,
  MD_transporter_Parameter_Type,
} from "md-curcuma";


const my_transporter: MD_Transporter = new MD_Transporter();

// Basic instructions for MD_Transporter

const transporter_parameter: MD_Transporter_Parameter_Type = {
  readPath: "test/obsidian-vault/longform.md",
  writePath: "test/hugo-content-2/",
  doSubfolders: false,
  limit: 1990,
  useCounter: false
};

const simulate_copy_job = true;

// The tasks to operate. If you add no tasks you have a simple copy job.

// Placeholders in the Transformer-Template
// as pre defined in MD_Transformer_TemplateValues_Type in the md-transformer module:
// {name_full} {name_suffix} {name}
const parameter_images: MD_Transformer_Parameter_Type = {
  tag_obsidian_prefix: "![[",
  tag_obsidian_suffix: "]]",
  find_rule: "jpg|png",
  replace_template: `{{< image src="assets/images/{name_full}" >}}`,
  copy_task: {
    source:"test/obsidian-vault/images/",
    target:"test/hugo-content-2/assets/images/{name}/",
    simulate:simulate_copy_job
  }
};

const parameter_docs: MD_Transformer_Parameter_Type = {
  tag_obsidian_prefix: "![[",
  tag_obsidian_suffix: "]]",
  find_rule: "pdf|ods|odp",
  replace_template: `{{< button href="/getthis.php?id={name}" name="download {name} ({name_suffix})" >}}`,
  copy_task: {
    source:"test/obsidian-vault/attachments/",
    target:"test/hugo-content-2/static/downloads/",
    simulate:simulate_copy_job
  }
};

var parameter_math: MD_Transformer_Parameter_Type = {
  tag_obsidian_prefix: "$$",
  tag_obsidian_suffix: "$$",
  find_rule: "",
  replace_template: "```math {.text-center}\n$$\n {content} \n$$\n```\n",
};


const parameter_remove: MD_Transformer_Parameter_Type = {
  tag_obsidian_prefix: "", // TODO optional?
  tag_obsidian_suffix: "", // TODO optional?
  find_rule: "- [ ] #TODO ",
  replace_template: ``,
};

// The Markdown-Splitter Task needs a Frontmatter

// The placeholders result from the definition of MD_Frontmatter_Type in md-frontmatter
// A file definition in frontmatter_filename, overwrites frontmatter Property
var splitter_frontmatter: MD_Frontmatter = new MD_Frontmatter(`---
title: "{title}"
description: ""
url: /docs/{url_prefix}-{url}/
date: {date}
draft: false
images: []
menu:
  docs:
    parent: "docs-9602b15bad02600f3883f55e2ade6b81"
    identifier: "{url_prefix}-{url}-{uuid}"
weight: {weight}
toc: true
---\n\n`);

const parameter_splitter: MD_Splitter_Parameter_Type = {
  pattern: "# ",
  cleanName: "# ",
  limit: 100,
  hasCounter: false,
  weightBase: 8000,
  url_prefix: "test-prefix",
  doRemoveHeadline: true,
  frontmatter_filename: "", // ./test/frontmatter-template.md
  frontmatter: splitter_frontmatter
};

my_transporter.addTransformer(new MD_ObsidianLink_Transformer(parameter_images));
my_transporter.addTransformer(new MD_ObsidianLink_Transformer(parameter_docs));
my_transporter.addTransformer(new MD_RemoveTODOS_Transformer(parameter_remove));
my_transporter.addTransformer(new MD_Math_Transformer(parameter_math));
my_transporter.addTransformer(new MD_Splitter_Transformer(parameter_splitter));

my_transporter.perform_job(transporter_parameter);

Compile The Transport-Script

The scripts can then be used as follows. In the terminal:

• npm run build:ts - build the library.
• npm run watch:ts - watch and compile on change to javascript.

The compiled files end up in the transport-scripts-compiled directory.

Run the Script in VSCode

• In Visual Studio Code
• Go to the directory transport-scripts-compiled
• On the compiled .js file
• Mouse right klick -> Menu-Entry Run Code

Look in the test folder to check the result.

Observe the console output on the Output tab.

Build A Custom Transformer

import { MD_Tranporter_Parameter_Type } from "md-curcuma";
import { MD_Transformer_AbstractBase } from "md-curcuma";

interface MD_Custom_Parameter_Type {
  custom_property: string;
}

class MD_Custom_Transformer extends MD_Transformer_AbstractBase {

    private parameter: MD_Custom_Parameter_Type;
  
    constructor(parameter: MD_Custom_Parameter_Type ){
      super();
      this.parameter = parameter;
    }
  
    public set_job_parameter(job_paramter: MD_Transporter_Parameter_Type): void {
      super.set_job_parameter(job_paramter); // this is a hack
    }

    public transform(file_content: MD_FileContent_Interface, index: number):  MD_FileContent_Interface {

      const body_array = file_content.body_array;
      let item = file_content.body_array[index];

      // file_content.frontmatter;
      // file_content.frontmatter_attributes;

      if (body_array[index].indexOf(this.parameter.find_rule) >= 0) {
        console.log(`Transform before: ${body_array[index]}`);
        body_array.splice(index, 1);
        file_content.index = file_content.index - 1;
        console.log(`Transform after: ${body_array[index]}`);
      }

      return file_content;
    }
  }

You currently cannot call custom Transformer from a configuration file, you have to instantiate it in your script.

You can mixup execution from config-file and adding your custom Transformer class like so:

const my_transporter: MD_Transporter = new MD_Transporter();

my_transporter.addTransformer(new MyCustomTransformer(custom_params));

my_transporter.perform_job_from("../test-data/transport-config.json", "Example-Job No.1");