cheeriojs/cheerio

Hello Everyone, I am trying to use cheerio in a nodejs application and trying to parse a document to get a structure that I can use to index documents with algolia. I am allowing the developer consuming this to pass in a structure such as h1 h3 p, or h1 h2 h3 p but this will be dynamic. I am then trying to get the elements and create json objects which will be passed to algolia for indexing. Below is a sample html structure and how the json objects are created

<h1>Article Title</h1>
<h3>Section Title</h3>
<p>Content</p>
<h3>Section Two</h3>
<p>Content 2</p>
<h3>Section Three</h3>
<p>Content 3</p>
<h1>Secondary Title</h1>
<p>Secondary Content</p>

so for the options.structure I would pass in h1 h3 p, this would then create the following structures to pass to algolia

{
  "link":"path/article#article-title",
  "importance":0,
  "objectID":"path/article#article-title-816d21744be4034a4bfa3323722024b4"
  "h1":"Article Title",
}

{
  "link":"path/article#section-title",
  "importance":3,
  "objectID":"path/article#section-title-816d21744be4034a4bfa3323722024b4",
  "h1":"Article Title",
  "h3":"Section Title"
}

{
  "link":"path/article#section-title-p0",
  "importance":5,
  "objectID":"path/article#section-title-p0-ae5e3efcdddc387616bbc5b1e5b1b134",
  "h1":"Article Title",
  "h3":"Section Title",
  "p":"Content"
}

{
  "link":"path/article#section-two”,
  "importance":3,
  "objectID":"path/article#section-two-816d21744be4034a4bfa3323722024b4",
  "h1":"Article Title",
  "h3":"Section Two”
}

{
  "link":"path/article#section-two-p0",
  "importance":5,
  "objectID":"path/article#section-two-p0-ae5e3efcdddc387616bbc5b1e5b1b134",
  "h1":"Article Title",
  "h3":"Section Two”,
  "p":"Content 2”
}

{
  "link":"path/article#section-three”,
  "importance":3,
  "objectID":"path/article#section-three-816d21744be4034a4bfa3323722024b4",
  "h1":"Article Title",
  "h3":"Section Three”
}

{
  "link":"path/article#section-three-p0",
  "importance":5,
  "objectID":"path/article-title#section-three-p0-ae5e3efcdddc387616bbc5b1e5b1b134",
  "h1":"Article Title",
  "h3":"Section Three”,
  "p":"Content 3”
}

{
  "link":"path/article#secondary-title",
  "importance":0,
  "objectID":"path/article#secondary-title-816d21744be4034a4bfa3323722024b4"
  "h1”:”Secondary Title",
}

{
  "link":"path/article#secondary-title-p0”,
  "importance”:4,
  "objectID":"path/article#secondary-title-p0-816d21744be4034a4bfa3323722024b4"
  "h1”:”Secondary Title",
  "p”:”Secondary Content”,
}

Now this structure would have to be dynamic meaning rather than the above they could pass in h1 h2 h3 h4 p and it should be parsed accordingly.

There could also be content under h1 such as <h1>Title</h1><p>Content</p> and this would receive a different importance as seen in the last example above.

The trouble I am having is how can I select all paragraphs under each tag in the structure. I know I can get all p tags under h1 and that will return them all but it will not let me know which h1 the p’s belong to, etc.

I am following this article trying to replicate this using cheerio.
https://blog.algolia.com/how-to-build-a-helpful-search-for-technical-documentation-the-laravel-example/