Convert HTML into Markdown with Go. It is using an HTML Parser to avoid the use of regexp as much as possible. That should prevent some weird cases and allows it to be used for cases where the input is totally unknown.

go get github.com/JohannesKaufmann/html-to-markdown

import md "github.com/JohannesKaufmann/html-to-markdown"

converter := md.NewConverter("", true, nil)

html = `<strong>Important</strong>`

markdown, err := converter.ConvertString(html)
if err != nil {
  log.Fatal(err)
}
fmt.Println("md ->", markdown)

If you are already using goquery you can pass a selection to Convert.

markdown, err := converter.Convert(selec)

If you want to make use of html-to-markdown on the command line without any Go coding, check out html2md, a cli wrapper for html-to-markdown that has all the following options and plugins builtin.

The third parameter to md.NewConverter is *md.Options.

For example you can change the character that is around a bold text ("**") to a different one (for example "__") by changing the value of StrongDelimiter.

opt := &md.Options{
  StrongDelimiter: "__", // default: **
  // ...
}
converter := md.NewConverter("", true, opt)

For all the possible options look at godocs and for a example look at the example.

converter.AddRules(
  md.Rule{
    Filter: []string{"del", "s", "strike"},
    Replacement: func(content string, selec *goquery.Selection, opt *md.Options) *string {
      // You need to return a pointer to a string (md.String is just a helper function).
      // If you return nil the next function for that html element
      // will be picked. For example you could only convert an element
      // if it has a certain class name and fallback if not.
      content = strings.TrimSpace(content)
      return md.String("~" + content + "~")
    },
  },
  // more rules
)

For more information have a look at the example add_rules.

If you want plugins (github flavored markdown like striketrough, tables, ...) you can pass it to Use.

import "github.com/JohannesKaufmann/html-to-markdown/plugin"

// Use the `GitHubFlavored` plugin from the `plugin` package.
converter.Use(plugin.GitHubFlavored())

Or if you only want to use the Strikethrough plugin. You can change the character that distinguishes the text that is crossed out by setting the first argument to a different value (for example "~~" instead of "~").

converter.Use(plugin.Strikethrough(""))

For more information have a look at the example github_flavored.

Have a look at the plugin folder for a reference implementation. The most basic one is Strikethrough.

This library produces markdown that is readable and can be changed by humans.

Once you convert this markdown back to HTML (e.g. using goldmark or blackfriday) you need to be careful of malicious content.

This library does NOT sanitize untrusted content. Use an HTML sanitizer such as bluemonday before displaying the HTML in the browser.

Godoc

Determines which elements are to be kept and rendered as HTML.

Determines which elements are to be removed altogether i.e. converted to an empty string.

If you find HTML snippets (or even full websites) that don't produce the expected results, please open an issue!

Please first discuss the change you wish to make, by opening an issue. I'm also happy to guide you to where a change is most likely needed.

Note: The outside API should not change because of backwards compatibility...

You don't have to be afraid of breaking the converter, since there are many "Golden File Tests":

Add your problematic HTML snippet to one of the input.html files in the testdata folder. Then run go test -update and have a look at which .golden files changed in GIT.

You can now change the internal logic and inspect what impact your change has by running go test -update again.

Note: Before submitting your change as a PR, make sure that you run those tests and check the files into GIT...

• turndown (js), a very good library written in javascript.
• lunny/html2md, which is using regex instead of goquery. I came around a few edge case when using it (leaving some html comments, ...) so I wrote my own.

This project is licensed under the terms of the MIT license.