ConLite/conlite/plugins/pluginmanager/includes/functions/simplexml_tree.php

250 Zeilen
7.9 KiB
PHP

<?php
/**
* Output a tree-view of the node or list of nodes referenced by a particular SimpleXML object
* Unlike simplexml_dump(), this processes the entire XML tree recursively, while attempting
* to be more concise and readable than the XML itself.
* Additionally, the output format is designed as a hint of the syntax needed to traverse the object.
*
* @param SimpleXMLElement $sxml The object to inspect
* @param boolean $include_string_content Default false. If true, will summarise textual content,
* as well as child elements and attribute names
* @param boolean $return Default false. If true, return the "dumped" info rather than echoing it
* @return null|string Nothing, or output, depending on $return param
*
* @author Rowan Collins
* @see https://github.com/IMSoP/simplexml_debug
* @license None. Do what you like with it, but please give me credit if you like it. :)
* Equally, no warranty: don't blame me if your aircraft or nuclear power plant fails because of this code!
*/
function simplexml_tree(SimpleXMLElement $sxml, $include_string_content=false, $return=false)
{
$indent = "\t";
$content_extract_size = 15;
// Get all the namespaces declared at the *root* of this document
// All the items we're looking at are in the same document, so we only need do this once
$doc_ns = $sxml->getDocNamespaces(false);
$dump = '';
// Note that the header is added at the end, so we can add stats
// The initial object passed in may be a single node or a list of nodes, so we need an outer loop first
// Note that for a single node, foreach($node) acts like foreach($node->children())
// Numeric array indexes, however, operate consistently: $node[0] just returns the node
$root_item_index = 0;
while ( isset($sxml[$root_item_index]) )
{
$root_item = $sxml[$root_item_index];
// Special case if the root is actually an attribute
// It's surprisingly hard to find something which behaves consistently differently for an attribute and an element within SimpleXML
// The below relies on the fact that the DOM makes a much clearer distinction
// Note that this is not an expensive conversion, as we are only swapping PHP wrappers around an existing LibXML resource
if ( dom_import_simplexml($root_item) instanceOf DOMAttr )
{
// To what namespace does this attribute belong? Returns array( alias => URI )
$ns = $root_item->getNamespaces(false);
if ( key($ns) )
{
$dump .= key($ns) . ':';
}
$dump .= $root_item->getName() . '="' . (string)$root_item . '"' . PHP_EOL;
}
else
{
// Display the root node as a numeric key reference, plus a hint as to its tag name
// e.g. '[42] // <Answer>'
// To what namespace does this attribute belong? Returns array( alias => URI )
$ns = $root_item->getNamespaces(false);
if ( key($ns) )
{
$root_node_name = key($ns) . ':' . $root_item->getName();
}
else
{
$root_node_name = $root_item->getName();
}
$dump .= "[$root_item_index] // <$root_node_name>" . PHP_EOL;
// This function is effectively recursing depth-first through the tree,
// but this is managed manually using a stack rather than actual recursion
// Each item on the stack is of the form array(int $depth, SimpleXMLElement $element, string $header_row)
$dump .= _simplexml_tree_recursively_process_node(
$root_item, 1,
$include_string_content, $indent, $content_extract_size
);
}
$root_item_index++;
}
// Add on the header line, with the total number of items output
$dump = 'SimpleXML object (' . $root_item_index . ' item' . ($root_item_index > 1 ? 's' : '') . ')' . PHP_EOL . $dump;
if ( $return )
{
return $dump;
}
else
{
echo $dump;
}
}
/**
* "Private" function to perform the recursive part of simplexml_tree()
* Do not call this function directly or rely on its function signature remaining stable
*/
function _simplexml_tree_recursively_process_node($item, $depth, $include_string_content, $indent, $content_extract_size)
{
$dump = '';
if ( $include_string_content )
{
// Show a chunk of the beginning of the content string, collapsing whitespace HTML-style
$string_content = (string)$item;
$string_extract = preg_replace('/\s+/', ' ', trim($string_content));
if ( strlen($string_extract) > $content_extract_size )
{
$string_extract = substr($string_extract, 0, $content_extract_size)
. '...';
}
if ( strlen($string_content) > 0 )
{
$dump .= str_repeat($indent, $depth)
. '(string) '
. "'$string_extract'"
. ' (' . strlen($string_content) . ' chars)'
. PHP_EOL;
}
}
// To what namespace does this element belong? Returns array( alias => URI )
$item_ns = $item->getNamespaces(false);
if ( ! $item_ns )
{
$item_ns = array('' => NULL);
}
// This returns all namespaces used by this node and all its descendants,
// whether declared in this node, in its ancestors, or in its descendants
$all_ns = $item->getNamespaces(true);
// If the default namespace is never declared, it will never show up using the below code
if ( ! array_key_exists('', $all_ns) )
{
$all_ns[''] = NULL;
}
// Prioritise "current" namespace by merging into onto the beginning of the list
// (it will be added to the beginning and the duplicate entry dropped)
$all_ns = array_merge($item_ns, $all_ns);
foreach ( $all_ns as $ns_alias => $ns_uri )
{
$children = $item->children($ns_alias, true);
$attributes = $item->attributes($ns_alias, true);
// If things are in the current namespace, display them a bit differently
$is_current_namespace = ( $ns_uri == reset($item_ns) );
if ( count($attributes) > 0 )
{
if ( ! $is_current_namespace )
{
$dump .= str_repeat($indent, $depth)
. "->attributes('$ns_alias', true)" . PHP_EOL;
}
foreach ( $attributes as $sx_attribute )
{
// Output the attribute
if ( $is_current_namespace )
{
// In current namespace
// e.g. ['attribName']
$dump .= str_repeat($indent, $depth)
. "['" . $sx_attribute->getName() . "']"
. PHP_EOL;
$string_display_depth = $depth+1;
}
else
{
// After a call to ->attributes()
// e.g. ->attribName
$dump .= str_repeat($indent, $depth+1)
. '->' . $sx_attribute->getName()
. PHP_EOL;
$string_display_depth = $depth+2;
}
if ( $include_string_content )
{
// Show a chunk of the beginning of the content string, collapsing whitespace HTML-style
$string_content = (string)$sx_attribute;
$string_extract = preg_replace('/\s+/', ' ', trim($string_content));
if ( strlen($string_extract) > $content_extract_size )
{
$string_extract = substr($string_extract, 0, $content_extract_size)
. '...';
}
$dump .= str_repeat($indent, $string_display_depth)
. '(string) '
. "'$string_extract'"
. ' (' . strlen($string_content) . ' chars)'
. PHP_EOL;
}
}
}
if ( count($children) > 0 )
{
if ( $is_current_namespace )
{
$display_depth = $depth;
}
else
{
$dump .= str_repeat($indent, $depth)
. "->children('$ns_alias', true)" . PHP_EOL;
$display_depth = $depth + 1;
}
// Recurse through the children with headers showing how to access them
$child_names = array();
foreach ( $children as $sx_child )
{
// Below is a rather clunky way of saying $child_names[ $sx_child->getName() ]++;
// which avoids Notices about unset array keys
$child_node_name = $sx_child->getName();
if ( array_key_exists($child_node_name, $child_names) )
{
$child_names[$child_node_name]++;
}
else
{
$child_names[$child_node_name] = 1;
}
// e.g. ->Foo[0]
$dump .= str_repeat($indent, $display_depth)
. '->' . $sx_child->getName()
. '[' . ($child_names[$child_node_name]-1) . ']'
. PHP_EOL;
$dump .= _simplexml_tree_recursively_process_node(
$sx_child, $display_depth+1,
$include_string_content, $indent, $content_extract_size
);
}
}
}
return $dump;
}