Subscribe Free

Join 2670+ others. No spamming.
I promise!

We are currently under high development. Follow us at github.

Looking for Python Tutorials?
Check these awesome tutorials



cgrand / enlive


a selector-based (à la CSS) templating and transformation system for Clojure


Enlive Build Status

Enlive is a selector-based (à la CSS) templating library for Clojure.

David Nolen wrote a nice tutorial.

Another tutorial is by Brian Marick.

There's a quickstart section in Clojure Cookbook.

Where do I get support?

On the Enlive Google Group


All artifacts are published to clojars. Latest version is 1.1.6:

[enlive "1.1.6"]

What's new in Enlive?

(most recent first)


  • ADD: exception message when html-resource not found.
  • FIX: auto-reload on windows (also works with chestnut).

Auto-reloading (1.1.2)

(net.cgrand.reload/auto-reload *ns*)

Each time a resource or file used by a template/snippet is updated the namespace is reloaded (as per (require ... :reload)).


  • Perf improvements
  • Fixes to ${vars} substitutions

Pluggable parsers! (1.1.1)

The *parser* dynamic var controls the parser to be used by html-resource at runtime. (or you can pass {:parser XXX} as an additional arg).

For templates and snippets whose sources are not read dynamically, you can opt for another parser either locally:

(deftemplate ugh {:parser xml-parser}
  ( "<a><div>hello</div></a>")

or globally for the declaring ns:

(set-ns-parser! xml-parser)

A parser is a function from InputStream to nodes. xml-parser, net.cgrand.tagsoup/parser and net.cgrand.jsoup/parser are the three builtin ones.

${vars} substitutions (1.1.1)

The following selector + function is going to replace any ${var} in text and attributes by the value found in the map (or any function).

[:#container any-node] (replace-vars {:name "world" :class "hello"})

hiccup-style helper (1.1.0)

(content (html [:h3#hello "Hello worls"]))

older stuff

By default selector-transformation pairs are run sequentially. When you know that several transformations are independent, you can now specify (as an optimization) to process them in lockstep. Note that this doesn't work with fragments selectors.


  [:a :selector] a-transformation
  [:another :selector] another-transformation
  [:a :dependent :selector] yet-another-transformation

If the first two transformations are independent you can rewrite this code as:

  {[:a :selector] a-transformation
   [:another :selector] another-transformation}
  [:a :dependent :selector] yet-another-transformation

Transformations are now slightly restricted in their return values: a node or a collection of nodes (instead of freely nested collections of nodes).

Dynamic selectors: selectors aren't compiled anymore. It means that you don't need to wrap them in (selector ...) forms anymore nor to eval them in the most dynamic cases.

Fragment selectors allow to select adjacent nodes. They are denoted by a map of two node selectors (eg {[:h1] [:p]}), bounds are inclusive and they select the smallest matching fragments.

Transformations (the right-hand parts of rules) are now plain old closures. These functions take one arg (the selected node) and return nil, another node or a collection of nodes.

Rules are applied top-down: the first rule transforms the whole tree and the resulting tree is passed to the next rules.

Nodes are transformed deep-first, that is: if a selector selects several nodes, descendants are transformed first. Hence, when the transformation is applied to an ancestor, you can "see" the transformed descendants (but you can not see your transformed siblings).

   /B                                                                             /(T B)
  A    if A and B are selected and transformed by T the the resulting tree is (T A      )
   \C                                                                             \C


snippet is a unit of your page. It may be logical or visual entry, such as header, footer, page element. Snippet is usually a part of a template, and may serve as a container for other snippets. For example, for navigation on the web page. For that, let’s first define an html template for the navigation. Snippets are created by using net.cgrand.enlive-html/defsnippet function and, same as templates, they require a corresponding HTML template file to be availble in a classpath.

So, snippet function returns a seq of nodes, it can be used as a building block for more complex templates.

templates combine snippets together, they serve like a basement for the snippets. In order to create a template, you can use net.cgrand.enlive-html/deftemplate function. deftemplate is used as something what you would call layout in some other templating systems. In essence, it’s either a self-contained page (rarely true in bigger applications), or a container for snippets.

That said, a template is a returns a seq of string -- basically it's a snippet whose output is serialized. Templates return a seq of strings to avoid building the whole string.

Templates and snippets transform a source (specified as a path (to access resources on the classpath), a File, a Reader, an InputStream, a URI, a URL, an element or a seq of nodes).

Next concept is selectors, which are used within snippets and templates to identify the block of HTML code the transformation would be applied to. They’re very similar to CSS selectors, but also allow more sophisticated, predicate-based selections, for example, you can select a tag based on some part of content, or an attribute. Transformations are functions that triggered on the elements found by selectors. They receive content obtained selector, and modify it in some way.

Quickstart tutorial


If you want to go see the compiled version of the following steps all in one place, you can check out an example Ring application.

First thing you need to start, is to define your first template:

(require '[net.cgrand.enlive-html :as html])

(html/deftemplate main-template "templates/application.html"

Now, you can start writing selectors and transformations for the given selectors. Let's add a title to the template. Given that your template already has <head> and <title> tags, let's insert a title.

Content of templates/application.html:

<!DOCTYPE html>
<html lang="en">
    <title>This is a title placeholder</title>
(html/deftemplate main-template "templates/application.html"
  [:head :title] (html/content "Enlive starter kit"))

Here, [:head :title] is a selector, pretty much like a css selector. If you're coming from jQuery, you can write same selector as $("head title"). html/content is a transformation. It puts the given content into the element specified by your selector.


Let's add several snippets. For example, navigation and some content. For that, let's first define a template for the navigation. Content of templates/header.html

<!DOCTYPE html>
<html lang="en">
      <h1>Header placeholder</h1>
      <ul id="navigation">
        <li><a href="#">Placeholder for navigation</a></li>
(html/defsnippet main-template "templates/header.html"
  [heading navigation-elements]
  [:h1] (html/content heading)
  [:ul [:li html/first-of-type]] (html/clone-for [[caption url] navigation-elements]
                                                 [:li :a] (html/content caption)
                                                 [:li :a] (html/set-attr :href url)))


Enlive selectors can match either nodes or fragments (several adjacent nodes).

At the core, every selector is a vector. The items of this vector are called steps.

A step is a predicate, for example :h1, :p.some-class or even (attr? :lang).

To select elements which match several predicates, you need to group predicates into a vector: inside steps, vectors mean "and". This may seem confusing but the rule is simple: the outer-most vector hierarchically chains steps, all other vectors denote intersection (and) between steps.

So [:p (attr? :lang)] is going to match any elements with a lang attribute inside a :p element. On the other hand, [[:p (attr? :lang)]] is going to match any p with a lang attribute.

Similarly, sets group predicates in an union. Hence inside steps, sets mean "or". So [#{:div.class1 :div.class2}] match every div which has either class1 or class2. This can alternatively be written as [[:div #{:.class1 .class2}]]. Indeed you can have nested "ors" and "ands" which means nested sets and vectors.

At the top level you can have a big "or" between selectors by wrapping several selectors in a set. #{[:td :em] [:th :em]} is going to match any em insides either a th or a td. This is equivalent to [#{:td :th} :em].

Selector Syntax

See syntax.html

Some examples:

Enlive                                       CSS
[:div]                                       div
[:body :script]                              body script
#{[:ul.outline :> :li] [:ol.outline :> li]}  ul.outline > li, ol.outline > li
[#{:ul.outline :ol.outline} :> :li]          ul.outline > li, ol.outline > li
[[#{:ul :ol} :.outline] :> :li]              ul.outline > li, ol.outline > li
[:div :> :*]                                 div > *
[:div :> text-node]                          (text children of a div)
[:div :> any-node]                           (all children (including text nodes and comments) of a div)
{[:dt] [:dl]}                                (fragments starting by DT and ending at the *next* DD)

The at form

The at form is the most important form in Enlive. There are implicit at forms in snippet and template.

(at a-node
  [:a :selector] a-transformation
  [:another :selector] another-transformation
  ;; ...

The right-hand value of a rule can be nil. It's the idiomatic way to remove an element.

Transformations are closures which take one arg (the selected node) and return nil, another node or an arbitrarily nested collection of nodes.

Rules are applied top-down: the first rule transforms the whole tree and the resulting tree is passed to the next rules.


A transformation is a function that returns either a node or collection of node.

Enlive defines several helper functions:

;; Replaces the content of the element. Values can be nodes or collection of nodes.
(content "xyz" a-node "abc")

;; Replaces the content of the element. Values are strings containing html code.
(html-content "<blink>please no</blink>")

;; Wraps selected node into the given tag
(wrap :div)
;; or
(wrap :div {:class "foo"})

;; Opposite to wrap, returns the content of the selected node

;; Sets given key value pairs as attributes for selected node
(set-attr :attr1 "val1" :attr2 "val2")

;; Removes attribute(s) from selected node
(remove-attr :attr1 :attr2)

;; Adds class(es) to the selected node
(add-class "foo" "bar")

;; Removes class(es) from the selected node
(remove-class "foo" "bar")

;; Chains (composes) several transformations. Applies functions from left to right.
(do-> transformation1 transformation2)

;; Clones the selected node, applying transformations to it.
(clone-for [item items] transformation)
;; or
(clone-for [item items]
  selector1 transformation1
  selector2 transformation2)

;; Appends the values to the content of the selected element.
(append "xyz" a-node "abc")

;; Prepends the values to the content of the selected element.
(prepend "xyz" a-node "abc")

;; Inserts the values after the current selection (node or fragment).
(after "xyz" a-node "abc")

;; Inserts the values before the current selection (node or fragment).
(before "xyz" a-node "abc")

;; Replaces the current selection (node or fragment).
(substitute "xyz" a-node "abc")

;; Takes all nodes (under the current element) matched by src-selector, removes
;; them and combines them with the elements matched by dest-selector.
(move [:.footnote] [:#footnotes] content)

Known limitations/problems

  • No namespaces support (hence unsuitable for most XML)