If the html contains inline images (or some other html entity with non text content) with spaces in between, the parsed text contains those spaces.

<p><img></img> <img></img> <img></img> text</p>

would produce <space><space><space>text after parse.

These spaces count as characters in the pruneLength.

A simple regex to trim this could be

parsed = parsed.replace(/^\s+|\s+$|\s+(?=\s)/g, "")

I upgraded and started seeing lots of unescaped characters in my excerpts. Lodash only unescapes a handful of characters and suggests using https://github.com/mathiasbynens/he for more robust support.

Say I have some HTML like this:

<h1>Title</h1>
<p>Hello, world</p>
<p>Hello again, world</p>
<hr>
<p>This is where the content really gets going, world</p>

I want to use <hr> as the excerpt cutoff, like so:

excerpt(myString, {moreRegExp: /<hr>/i})

Unfortunately the excerpt also includes the <h1> content. Would you be open to adding an option to omit h1-h6 tags from the excerpt?

A real-world example of this can be found on the Electron blog, where Jekyll's excerpt creation ignores headings by default:

Hi @martinheidegger! 👋

The docs say:

It will either use the first found paragraph or everything up to a <!-- more -->

But I'm not seeing that behavior:

excerpt_html('first line\n\nsecond line')
// => 'first line\nsecond line'

excerpt_html('#heading\n\nfirst line\n\nsecond line')
//=> '#heading\nfirst line\nsecond line'

Am I missing something?

I'm getting the following error in react-native/expo, was previously working fine.

Hello,

I have markdown content (blog posts) where the first paragraph(s) usually only contains an image and the actual textual content starts after. This module fails to correctly extract the excerpt in that case, returning an empty string instead.

test('ignore empty first paragraphs', function (t) {
  t.equals(excerptHtml('<p><img src="cat.png" /></p><p><img src="dog.png" /></p><p>test</p>', { 
  }), 'test')
  t.end()
})

Would you be interested in a PR fixing this behavior?