posthtml provides a fluent interface to decorating a
string of HTML, like "<ul><li/><li/></ul>" - to add
attributes such as class or id to elements, or to do
custom processing.
(posthtml-decorate "<ul><li/><li/></ul>"
'(:id ul "ID")
'(:class each "ul li" "CLASSNAME")
'((lambda (li)
(posthtml-append li "CONTENT"))
"ul>li"))
The string is parsed into an esxml (DOM) tree structure
with libxml-parse-html-region - the defined decorators are
applied in sequence, and a string rendered via esxml-to- xml is returned. Querying is done with esxml-query.
The first argument to posthtml-decorate is a function or
one of the following tokens; prepend, append, set,
add, id and class. These may be used with or without
an ':' prefix, and additional functionality can be specified
by updating posthtml-decorator-token-alist.
'(:append h1 " APPEND") ;; append to contents of h1 element
'(:set ul :class "COLUMNS") ;; add COLUMNS class to ul element
'(:set h1 :id "TWO" :class "") ;; set h1's id to TWO, remove all classes
'(:set ul :id "") ;; remove ul's id attribute
posthtml-decorate translates the syntax decorators into a
call to posthtml-apply or posthtml-apply-each,
respectively. They are responsible for querying CONTAINER
for SELECTOR, and applying (each) found element - as an
esxml parse-tree - and ARGS to FN.
'((lambda (list-element current-uri)
(let ((link-uri (posthtml-attr (posthtml-find list-element 'a) 'href)))
(when (string= link-uri current-uri)
(posthtml-set list-element :class "CURRENT"))))
each "ul li" "/uri-two")
As the element provided to each function is an esxml parse
-tree, it is available for further querying and
processing. In the following, the ul is updated with a
dynamically created CSS class name, to reflect how many
elements it contains.
'((lambda (ul)
(posthtml-add ul :class
(format "COLUMNSWIDTH%s"
(length (posthtml-find-all ul "li")))))
"ul"))
Here is the result of running this contrived test case, with input on the left.
<html><body> | <html><body>
<h1 class=\"ONE\">headline</h1> | <h1 id=\"TWO\">headline APPEND</h1>
<ul id=\"ONE\"> | <ul class=\"COLUMNS COLUMNSWIDTH3\">
<li><a href="/uri-one">uri</a></li> | <li><a href="/uri-one">uri</a></li>
<li><a href="/uri-two">uri</a></li> | <li class="CURRENT"><a href="/uri-two">uri</a></li>
<li><a href="/uri-three">uri</a></li> | <li><a href="/uri-three">uri</a></li>
</ul> | </ul>
</body></html> | </body></html>
posthtml-find and posthtml-find-all both utilize esxml-
query to provide
querySelector-like functionality in finding (collections
of) nodes with CSS selectors.
Selectors are passed in as a string, as a combination of the following available options.
foo, bar ;; Commas
foo bar ;; Descendant combinator
foo>bar ;; Child combinator
* ;; Universal selector
tag ;; Type selector
#foo ;; ID selector
.foo ;; Class selector
[foo] ;; Attribute selector
[foo=bar] ;; Exact match attribute selector
[foo^=bar] ;; Prefix match attribute selector
[foo$=bar] ;; Suffix match attribute selector
[foo*=bar] ;; Substring match attribute selector
[foo~=bar] ;; Include match attribute selector
[foo|bar] ;; Dash match attribute selector
posthtml-append and posthtml-prepend add child nodes to
elements; the new elements can be strings or esxml
lists. Please also see the documentation for Emacs'
Document Object Model package.
posthtml-attr retrieves, or with a value sets, a dom
node's attribute.
Supplying an empty string to posthtml-set removes the
attribute, posthtml-add appends to the current value.
org-export-filter-final-output-functions provide access to
the final output of any org-export, or org-publishing
sequence; this list consists of the last functions called
with the rendered output, their output being what's written
to disk.
(add-to-list 'org-export-filter-final-output-functions
(lambda (contents backend info)
(posthtml-decorate
contents
'(:set body :style "background-color: black;"))))
emacs -batch \
-l ert \
-l posthtml-tests.el \
-f ert-run-tests-batch-and-exit