html5-walker

A thin Clojure wrapper around jfiveparse, this lets you find and replace in HTML5 strings.

Install

add [html5-walker "2022.03.07"] to :dependencies in your project.clj

add html5-walker {:mvn/version "2022.03.07"} to :deps in your deps.edn

Usage

html5-walker exposes these functions:

find-nodes

Signature: (find-nodes html-string path)

It returns a sequence of Nodes matching the path.

A path is a vector of keywords or string of hiccup-esque element selectors. Like this:

[:a] matches all anchor tags.
[:form :input] matches all input tags nested inside a form.
[:div.foo] matches all div tags with "foo" in its class name.
[:.button] matches all elements with the "button" class.
[:div#content] matches the div with "content" as its id.
[:first-child] matches any element that is the first child.
[:last-child] matches any element that is the last child.
["meta[property]"] matches all meta tags with the property attribute.
["meta[property=og:title]"] matches all meta tags with the property attribute set to "og:title".

The following additional attribute selectors are also supported, and work like they do in CSS: *=, $=, ~= and ^=.

So running:

(find-nodes "<ul><li>1</li><li>2</li></ul>" [:ul :li])

would return a sequence with two li nodes in it. See the javadoc for more information about these nodes.

replace-in-fragment

Signature: (replace-in-fragment html-string path->fn)

This returns a new html-string with any changes performed by the functions in the path->fn map applied.

So running:

(replace-in-fragment
  "<ul><li>1</li><li>2</li></ul>"
  {[:ul :li] (fn [node] (.setInnerHTML node (str (.getInnerHTML node) "!!!")))})

would return:

"<ul><li>1!!!</li><li>2!!!</li></ul>"

replace-in-document

Just like replace-in-fragment, except it works on an entire html document. This means that html, head and body tags are expected to be there. They will be added if missing.

Note that replace-in-fragment will actually remove these tags when found.

More usage

Take a look at the tests if you'd like more examples.

Changes

2022.03.07

Upgrade jfiveparse to version 0.9.0
Switch versioning separator to dot, for more Maven friendly version numbers

2020-01-08

Support selecting only by class name, like so: :.myclass

License

Distributed under the Eclipse Public License, the same as Clojure.

cjohansen / html5-walker