From HTML to Markdown. How to get rid of HTML attributes and raw HTML code, but preserve comments

[johnn2025]

I have a following HTML file:

<!-- https://fs.blog/feynman-technique/ -->
<h1 class="entry-title entry-title-single">The Feynman Technique: Master the Art of Learning</h1>
<div class="entry-content entry-content-single">
<p>The Feynman Technique is the most effective method to unlock your potential and develop a deep understanding. </p>
<p><a href="https://fs.blog/intellectual-giants/richard-feynman/">Richard Feynman</a> was not only a Nobel laureate in Physics but also a master of demystifying complex topics. His key learning insight: complexity and jargon often mask a lack of understanding. </p>
<p>Feynman&#8217;s learning technique comprises four key steps:</p>
<ol>
<li>Select a concept to learn.</li>
<li>Teach it to a child.</li>
<li>Review and refine your understanding.</li>
<li>Organize your notes and revisit them regularly.</li>
</ol>
<p>...</p>
<div class="wp-block-image">
<figure class="aligncenter"><img fetchpriority="high" decoding="async" width="1920" height="1080" src="https://149664534.v2.pressablecdn.com/wp-content/uploads/2012/04/FeynmanTechnique.jpg" alt="" class="wp-image-43131" srcset="https://149664534.v2.pressablecdn.com/wp-content/uploads/2012/04/FeynmanTechnique-300x169.jpg 300w , https://149664534.v2.pressablecdn.com/wp-content/uploads/2012/04/FeynmanTechnique-768x432.jpg 768w , https://149664534.v2.pressablecdn.com/wp-content/uploads/2012/04/FeynmanTechnique-1024x576.jpg 1024w , https://149664534.v2.pressablecdn.com/wp-content/uploads/2012/04/FeynmanTechnique-1536x864.jpg 1536w , https://149664534.v2.pressablecdn.com/wp-content/uploads/2012/04/FeynmanTechnique.jpg 1920w " sizes="(max-width: 1920px) 100vw, 1920px" /></figure></div>
<figure class="wp-block-pullquote"><blockquote><p>The person who says he knows what he thinks but cannot express it usually does not know what he thinks.</p><cite>Mortimer Adler</cite></blockquote></figure>
<h2 class="wp-block-heading">Step 1: Select a concept to learn.</h2>
<p>...</p>

Note the first line is a comment, which I want to preserve.

I need to convert this file to Markdown and to have as clear Markdown markup as possible, but also to keep any HTML comments there.

Here is what I tried:

pandoc from.htm -o to.md -f html+raw_html -t markdown --wrap=none --lua-filter=remove-attr.lua

function remove_attr (x)
  if x.attr then
    x.attr = pandoc.Attr()
    return x
  end
end

return {{Inline = remove_attr, Block = remove_attr}}

But I'm not really happy with its ouput:

`<!-- https://fs.blog/feynman-technique/ -->`{=html}

# The Feynman Technique: Master the Art of Learning

<div>

The Feynman Technique is the most effective method to unlock your potential and develop a deep understanding.

[Richard Feynman](https://fs.blog/intellectual-giants/richard-feynman/) was not only a Nobel laureate in Physics but also a master of demystifying complex topics. His key learning insight: complexity and jargon often mask a lack of understanding.

Feynman's learning technique comprises four key steps:

1.  Select a concept to learn.
2.  Teach it to a child.
3.  Review and refine your understanding.
4.  Organize your notes and revisit them regularly.

\...

<div>

![image](https://149664534.v2.pressablecdn.com/wp-content/uploads/2012/04/FeynmanTechnique.jpg)

</div>

<figure>
<blockquote>
<p>The person who says he knows what he thinks but cannot express it usually does not know what he thinks.</p>
<p><cite>Mortimer Adler</cite></p>
</blockquote>
</figure>

## Step 1: Select a concept to learn.

\...

</div>

For example, I would prefer <!-- https://fs.blog/feynman-technique/ --> over <!-- https://fs.blog/feynman-technique/ -->{=html} and

> The person who says he knows what he thinks but cannot express it usually does not know what he thinks
>
> Mortimer Adler

over

<figure>
<blockquote>
<p>The person who says he knows what he thinks but cannot express it usually does not know what he thinks.</p>
<p><cite>Mortimer Adler</cite></p>
</blockquote>
</figure>

Changing markdown to gfm makes output better but not a lot.

pandoc from.htm -o to.md -f html+raw_html -t gfm --wrap=none --lua-filter=remove-attr.lua

How can I achieve better output?

[jgm]

If you don't the {=html} raw attributes, then just specify -t markdown-raw_attribute.

As for the figure, there is currently no syntax in pandoc markdown for a figure that contains anything other than a single image. So we fall back to raw HTML.

You can disable that too: -t markdown-raw_attribute-raw_html - and then you'll get

::: {.figure .wp-block-pullquote}
> The person who says he knows what he thinks but cannot express it
> usually does not know what he thinks.
>
> Mortimer Adler
:::

but then you won't get the HTML comment at all.

So to get everything you want, you'll need to:

• use -f markdown-raw_attribute
• create a small filter that matches a Figure element whose contents are a BlockQuote, and replaces it with the BlockQuote.

[johnn2025]

Thanks! But isn't it simpler to use

pandoc from.htm -o to.md -t gfm-raw_html --wrap=none

# The Feynman Technique: Master the Art of Learning

The Feynman Technique is the most effective method to unlock your potential and develop a deep understanding.

[Richard Feynman](https://fs.blog/intellectual-giants/richard-feynman/) was not only a Nobel laureate in Physics but also a master of demystifying complex topics. His key learning insight: complexity and jargon often mask a lack of understanding.

Feynman’s learning technique comprises four key steps:

1.  Select a concept to learn.
2.  Teach it to a child.
3.  Review and refine your understanding.
4.  Organize your notes and revisit them regularly.

...

![](https://149664534.v2.pressablecdn.com/wp-content/uploads/2012/04/FeynmanTechnique.jpg)

> The person who says he knows what he thinks but cannot express it usually does not know what he thinks.
>
> Mortimer Adler

and to write a Lua filter that will preserve HTML comments? Or maybe it won't work because we don't have -f html+raw_html now?

[jgm]

The raw HTML comments won't be in the AST, so the filter can't preserve them.

[johnn2025]

Thanks! John, may I ask one another question about comments?

According to the Pandoc's user guide, the default value of the --strip-comments option is true. But then why are comments preserved when I convert Markdown to HTML without using this option at all?

pandoc from.md -o to.htm --wrap=none

<!-- commment -->

# head

body

<!-- commment -->
<h1 id="head">head</h1>
<p>body</p>

[jgm]

The default is false, and the manual doesn't say otherwise. The section you link to is just showing equivalences between specifying options on the command line and in a defaults file. It says nothing about pandoc's built-in defaults.