/**
* Extracting the data of one content
*
* The parse function splits the content into its entries.
* Then every entry is parsed by this function.
* It parses the entry backwards.
* First the last '=' is searched and the value extracted from that.
* A copy is made of the entry if warnings should be generated. This takes quite
* some memory but it is needed to get good warnings. If nor warnings are generated
* then you don have to worry about memory.
* Then the last ',' is searched and the field extracted from that.
* Again the entry is shortened.
* Finally after all field=>value pairs the cite and type is extraced and the
* authors are splitted.
* If there is a problem false is returned.
*
* @access private
* @param string $entry The entry
* @return array The representation of the entry or false if there is a problem
*/
function _parseEntry($entry)
{
$entrycopy = '';
if ($this->_options['validate']) {
$entrycopy = $entry; //We need a copy for printing the warnings
}
$ret = array('bibtex' => $entry.'}');
if ('@string' == strtolower(substr($entry, 0, 7))) {
//String are not yet supported!
if ($this->_options['validate']) {
$this->_generateWarning('STRING_ENTRY_NOT_YET_SUPPORTED', '', $entry.'}');
}
} elseif ('@preamble' == strtolower(substr($entry, 0, 9))) {
//Preamble not yet supported!
if ($this->_options['validate']) {
$this->_generateWarning('PREAMBLE_ENTRY_NOT_YET_SUPPORTED', '', $entry.'}');
}
} elseif ('@comment' == strtolower(substr($entry, 0, 8))) {
// Just ignores
} else {
// Look for key
$matches = array();
preg_match('/^@\w+\{([\w\d]+),/' ,$entry, $matches);
if ( count($matches) > 0 )
{
$ret['entrykey'] = $matches[1];
}
//Parsing all fields
while (strrpos($entry,'=') !== false) {
$position = strrpos($entry, '=');
//Checking that the equal sign is not quoted or is not inside a equation (For example in an abstract)
$proceed = true;
if (substr($entry, $position-1, 1) == '\\') {
$proceed = false;
}
if ($proceed) {
$proceed = $this->_checkEqualSign($entry, $position);
}
while (!$proceed) {
$substring = substr($entry, 0, $position);
$position = strrpos($substring,'=');
$proceed = true;
if (substr($entry, $position-1, 1) == '\\') {
$proceed = false;
}
if ($proceed) {
$proceed = $this->_checkEqualSign($entry, $position);
}
}
$value = trim(substr($entry, $position+1));
$entry = substr($entry, 0, $position);
if (',' == substr($value, strlen($value)-1, 1)) {
$value = substr($value, 0, -1);
}
if ($this->_options['validate']) {
$this->_validateValue($value, $entrycopy);
}
if ($this->_options['stripDelimiter']) {
$value = $this->_stripDelimiter($value);
}
if ($this->_options['unwrap']) {
$value = $this->_unwrap($value);
}
if ($this->_options['removeCurlyBraces']) {
$value = $this->_removeCurlyBraces($value);
}
$position = strrpos($entry, ',');
$field = strtolower(trim(substr($entry, $position+1)));
$ret[$field] = $value;
$entry = substr($entry, 0, $position);
}
//Parsing cite and entry type
$arr = explode('{', $entry);
$ret['cite'] = trim($arr[1]);
$ret['entrytype'] = strtolower(trim($arr[0]));
if ('@' == $ret['entrytype']{0}) {
$ret['entrytype'] = substr($ret['entrytype'], 1);
}
if ($this->_options['validate']) {
if (!$this->_checkAllowedEntryType($ret['entrytype'])) {
$this->_generateWarning('WARNING_NOT_ALLOWED_ENTRY_TYPE', $ret['entrytype'], $entry.'}');
}
}
// Process accents
foreach($ret as $key => &$value)
if ($key != "bibtex")
Structures_BibTex::process_accents($value);
// Handling pages
if (in_array('pages', array_keys($ret))) {
$matches = array();
if (preg_match("/^\s*(\d+)(?:\s*--?\s*(\d+))?\s*$/", $ret['pages'], $matches)) {
$ret['pages'] = new BibtexPages($matches[1], $matches[2]);
}
}
//Handling the authors
if (in_array('author', array_keys($ret)) && $this->_options['extractAuthors']) {
$ret['author'] = $this->_extractAuthors($ret['author']);
}
//Handling the editors
if (in_array('editor', array_keys($ret)) && $this->_options['extractAuthors']) {
$ret['editor'] = $this->_extractAuthors($ret['editor']);
}
}
return $ret;
}
