git clone ''

(ql:quickload :nathell.clj-tagsoup)

Clojars Project


This is a HTML parser for Clojure, somewhat akin to Common Lisp's cl-html-parse. It is a wrapper around the TagSoup Java SAX parser, but has a DOM interface. It is buildable by Leiningen.


The two main functions defined by clj-tagsoup are parse and parse-string. The first one can take anything accepted by's reader function except for a Reader, while the second can parse HTML from a string.

The resulting HTML tree is a vector, consisting of:

  1. a keyword representing the tag name,
  2. a map of tag attributes (mapping keywords to strings),
  3. children nodes (strings or vectors of the same format).

This is the same format as used by hiccup, thus the output of parse is appropriate to pass to hiccup.

There are also utility accessors (tag, attributes, children).

clj-tagsoup will automatically use the correct encoding to parse the file if one is specified in either the HTTP headers (if the argument to parse is an URL object or a string representing one) or a <meta http-equiv="..."> tag.

clj-tagsoup is meant to parse HTML tag soup, but, in practice, nothing prevents you to use it to parse arbitrary (potentially malformed) XML. The :xml keyword argument causes clj-tagsoup to take into consideration the XML header when detecting the encoding.

There are two other options for parsing XML:


project.clj: clojure (defproject clj-tagsoup-example "0.0.1" :dependencies [[clj-tagsoup/clj-tagsoup "0.3.0"]])

lein repl:

(use 'pl.danieljanus.tagsoup)
=> nil

(parse "")
=> [:html {}
          [:head {}
                 [:title {} "Example Web Page"]]
          [:body {}
                 [:p {} "You have reached this web page by typing \"\",\n\"\",\n  or \"\" into your web browser."]
                 [:p {} "These domain names are reserved for use in documentation and are not available \n  for registration. See "
                     [:a {:shape "rect", :href ""} "RFC \n  2606"]
                     ", Section 3."]]]


Truth be told, I wrote clj-tagsoup prior to discovering Enlive, which is an excellent library. That said, I believe clj-tagsoup has its niche. Here is an à la carte list of differences between the two:

It's for lazy-parse-xml. It's needed because that function uses, which under the hood uses the StAX API. TagSoup is a SAX parser, so a bridge between the two parsing APIs is needed.

If you don't use lazy-parse-xml, you can optionally exclude stax-utils from your project.clj, like this:

    :dependencies [[clj-tagsoup "0.3.0" :exclusions []]]


clj-tagsoup was written by Daniel Janus.