diff --git a/extras/scripts/create_structure_files.php b/extras/scripts/create_structure_files.php
new file mode 100644
index 0000000000000000000000000000000000000000..045e4106c19e565535e0bfd2316995f97af2b995
--- /dev/null
+++ b/extras/scripts/create_structure_files.php
@@ -0,0 +1,146 @@
+<?php
+/**
+ * Adds strucutre file to root of compound ojects.
+ * path_to_directory_containing_compound_objects\
+ * compound_object_1\
+ * compound_object_2\
+ * compound_object_3\
+ * .
+ * .
+ * .
+ *
+ *
+ * Usage:
+ *
+ * > php create_strcutre_files.php path_to_directory_containing_compound_objects
+ *
+ *
+ *
+ */
+
+
+$target_directory = trim($argv[1]);
+
+if(!is_dir($target_directory)){
+ exit("Please check that you have provided a full path to a directory as the input argument." . PHP_EOL);
+}
+
+$path_to_xsl = "tree_to_compound_object.xsl";
+
+scanWrapperDirectory($target_directory, 'structure', $path_to_xsl);
+
+// For use with use with get_dir_name(), which is used inside XSLT.
+$compound_obj_path = '';
+
+function scanWrapperDirectory($target_directory, $structurefilename = 'structure', $path_to_xsl) {
+ // basenames to exclude.
+ $exclude_array = array('..', '.DS_Store', 'Thumbs.db', '.');
+
+ $stuffinwrapperdirectory = scandir($target_directory);
+ foreach ($stuffinwrapperdirectory as $compoundObjectOrFile) {
+ $objpath = $target_directory . DIRECTORY_SEPARATOR . $compoundObjectOrFile;
+ if(!in_array($compoundObjectOrFile, $exclude_array) && is_dir($objpath)) {
+ global $compound_obj_path;
+ $compound_obj_path = $objpath;
+ // subdirectories of wrapper directory will be compound object.
+ // create a structure file for each.
+ $structure_xml = compoundObjectStructureXML($objpath);
+
+ // Apply XSLT
+ $structure_xml = treeToCompound($path_to_xsl, $structure_xml);
+ $structure_xml_output_file_path = $objpath . DIRECTORY_SEPARATOR
+ . $structurefilename . '.xml';
+ file_put_contents($structure_xml_output_file_path, $structure_xml);
+
+ }
+
+ }
+}
+
+
+function treeToCompound($path_to_xsl, $tree_output_xml) {
+ // Usage: php tree_to_compound.php tree_to_compound_object.xsl tree_output.xml
+
+ $xsl = $path_to_xsl;
+ // tree_output_xml is an xml string.
+ $xml = $tree_output_xml;
+
+ $xsl_doc = new DOMDocument();
+ $xsl_doc->load($xsl);
+
+ $xml_doc = new DOMDocument();
+ $xml_doc->loadXML($xml);
+
+ $xslt_proc = new XSLTProcessor();
+ $xslt_proc->importStylesheet($xsl_doc);
+ $xslt_proc->registerPHPFunctions();
+
+ $output = $xslt_proc->transformToXML($xml_doc);
+
+ return $output;
+}
+
+/**
+ * Removes path segments leading up to the last segment.
+ *
+ * Called from within the XSLT stylesheet.
+ */
+function get_dir_name() {
+ //global $input_dir;
+ //global $target_directory;
+ global $compound_obj_path;
+ $input_dir = $compound_obj_path;
+ $dir_path = preg_replace('/(\.*)/', '', $input_dir);
+ $dir_path = rtrim($dir_path, DIRECTORY_SEPARATOR);
+ $base_dir_pattern = '#^.*' . DIRECTORY_SEPARATOR . '#';
+ $dir_path = preg_replace($base_dir_pattern, '', $dir_path);
+ $dir_path = ltrim($dir_path, DIRECTORY_SEPARATOR);
+ echo $dir_path;
+ return $dir_path;
+}
+
+
+/**
+ * Recursively create XML string of directory structure/
+ * Based on psuedo-code from http://stackoverflow.com/a/15096721/850828
+ */
+function directoryXML($directory_path) {
+
+ // basenames to exclude.
+ $exclude_array = array('..', '.DS_Store', 'Thumbs.db', '.');
+
+ $dir_name = basename($directory_path);
+ $xml = "<directory name='" . $dir_name . "'>";
+
+ $pathbase = pathinfo($directory_path, PATHINFO_BASENAME);
+ $stuffindirectory = scandir($directory_path);
+
+ foreach($stuffindirectory as $subdirOrfile){
+
+ $subdirOrfilepath = $directory_path . DIRECTORY_SEPARATOR . $subdirOrfile;
+
+ if(!in_array($subdirOrfile, $exclude_array) && is_file($subdirOrfilepath)){
+ $xml .= "<file name='". $subdirOrfile . "' />";
+
+ }
+
+ if(!in_array($subdirOrfile, $exclude_array) && is_dir($subdirOrfilepath)){
+ $xml .= directoryXML($subdirOrfilepath);
+ }
+
+ }
+ $xml .= "</directory>";
+ return $xml;
+}
+
+function compoundObjectStructureXML($dir_path) {
+ $xmlstring = "<tree>";
+ $xmlstring .= directoryXML($dir_path);
+ $xmlstring .= "</tree>";
+ $xml = new DOMDocument( "1.0");
+ $xml->loadXML($xmlstring);
+ $xml->formatOutput = true;
+ return $xml->saveXML();
+}
+
+?>
\ No newline at end of file
diff --git a/extras/scripts/tree_to_compound_object.xsl b/extras/scripts/tree_to_compound_object.xsl
new file mode 100644
index 0000000000000000000000000000000000000000..b046a98806afe11fe6c92071daf445bb38f17b23
--- /dev/null
+++ b/extras/scripts/tree_to_compound_object.xsl
@@ -0,0 +1,43 @@
+<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
+ xmlns:php="http://php.net/xsl" xsl:extension-element-prefixes="php" exclude-result-prefixes="xsl xsi php">
+
+ <!-- XSLT stylesheet to convert the output of the 'tree' utility into XML that represents the structure
+ of an Islandora compound object. Called by the generate_compound_structure_file.php post-write hook
+ script Used by the MIK CSV compound toolchain. -->
+
+ <xsl:output method="xml" encoding="utf-8" indent="yes"/>
+ <xsl:strip-space elements="*"/>
+
+ <xsl:template match="tree/directory">
+ <xsl:comment>Islandora compound structure file generated by MIK. On batch ingest, 'parent' elements
+ will become compound objects, and 'child' elements will become their children. Files in
+ directories named in child elements' 'content' attribute will be added as their datastreams.
+ If 'parent' elements do not contain a MODS.xml file, the value of the 'parent' 'title'
+ attribute will be used as the parent's title/label.</xsl:comment>
+
+ <islandora_compound_object >
+ <xsl:attribute name="title">
+ <xsl:value-of select="php:function('get_dir_name')" />
+ </xsl:attribute>
+ <xsl:apply-templates/>
+ </islandora_compound_object>
+ </xsl:template>
+
+ <!-- We aren't intersted in these nodes, so apply an empty template to them. -->
+ <xsl:template match="report|directories|files"/>
+
+ <xsl:template match="directory">
+ <xsl:choose>
+ <xsl:when test="count(file) > 1">
+ <child content="{./@name}"/>
+ <xsl:apply-templates/>
+ </xsl:when>
+ <xsl:otherwise>
+ <parent title="{./@name}">
+ <xsl:apply-templates/>
+ </parent>
+ </xsl:otherwise>
+ </xsl:choose>
+ </xsl:template>
+
+</xsl:stylesheet>
diff --git a/includes/preprocessor.inc b/includes/preprocessor.inc
index bdde9d15430feee0e33aef7c9fb9004800238f1b..a4f00ed7240b0ed607d7fa4f61671d21aa830fb6 100644
--- a/includes/preprocessor.inc
+++ b/includes/preprocessor.inc
@@ -10,14 +10,14 @@
*/
class IslandoraCompoundBatchPreprocessor extends IslandoraBatchPreprocessor {
// prefixed flag for outputing echo or print statements
- protected $icbp_verbose = false;
+ protected $icbp_verbose = FALSE;
/**
* Function to get the OBJ XML files from the input directory.
*/
public function preprocess() {
if (strtolower($this->parameters['icbp_verbose']) == 'true') {
- $this->icbp_verbose = true;
+ $this->icbp_verbose = TRUE;
}
//$input_path = $this->parameters['target'] . DIRECTORY_SEPARATOR . '*.xml';
@@ -39,6 +39,7 @@ class IslandoraCompoundBatchPreprocessor extends IslandoraBatchPreprocessor {
foreach ($iterator as $fileinfo) {
$filePath = $fileinfo->getPathname();
$fileExtension = pathinfo($filePath, PATHINFO_EXTENSION);
+ $fileNameWithExtension = basename($filePath);
$isObjFile = $cbUtilities->extInContentModelMap($fileExtension);
if ($fileinfo->isFile() && $isObjFile) {
$batch_object = new IslandoraCompoundBatchObject($this->connection, $filePath, $this->parameters);
@@ -49,8 +50,8 @@ class IslandoraCompoundBatchPreprocessor extends IslandoraBatchPreprocessor {
//$files[] = $filePath;
$added[]= $batch_object;
}
- if ($fileExtension == 'cpd') {
-
+ if ($fileNameWithExtension == 'structure.xml' /*$fileExtension == 'cpd'*/) {
+
// The structure file will be in the root directory of the compound object.
// We need to create a parent wrapper object to associate the OBJ child objects too.
// Along with the structure file in the root directory of the compound object
@@ -70,9 +71,9 @@ class IslandoraCompoundBatchPreprocessor extends IslandoraBatchPreprocessor {
$compound_object_pid = $this->addToDatabase($batch_object);
//echo "Return Value: $returnvalue" . PHP_EOL;
- $this->addCompoundStrcuture($filePath, $compound_object_pid);
+ //$this->addCompoundStrcuture($filePath, $compound_object_pid);
//$this->addCompoundStrcuture($filePath);
-
+ $this->addStrcutureData($filePath, $compound_object_pid);
}
@@ -85,7 +86,7 @@ class IslandoraCompoundBatchPreprocessor extends IslandoraBatchPreprocessor {
* Echos statement if icbp_verbose property is true (set using --icbp_verbose drush option)
*/
public function icbpecho($string) {
- if ($this->icbp_verbose == true){
+ if ($this->icbp_verbose == TRUE) {
echo $string;
}
}
@@ -148,5 +149,68 @@ class IslandoraCompoundBatchPreprocessor extends IslandoraBatchPreprocessor {
}
} // End of addCompoundStrcuture
+
+
+ /**
+ * Parses structure.xml file and adds data to the islandora_compound_batch database table
+ *
+ */
+ public function addStrcutureData($structure_file_path, $compound_object_pid) {
+ static $object_num = 0;
+
+ $batch_id = $this->getSetId();
+
+ // Object ID is the batch set ID concatenated with the object number.
+ $object_id = $batch_id . $object_num;
+
+ // Increment
+ $object_num++;
+
+
+ $doc = new DOMDocument();
+ $doc->load($structure_file_path);
+
+ /*
+ <islandora_compound_object title="895">
+ <child content="893"/>
+ <child content="894"/>
+ </islandora_compound_object>
+ */
+
+ $wanted_tags = array('child');
+ foreach ($doc->getElementsByTagName('*') as $node) {
+ if (in_array($node->nodeName, $wanted_tags)) {
+ $this->icbpecho("Node name:\t" . $node->nodeName . "\n");
+ if ($node->nodeName == 'child') {
+ $xpath = new DOMXPath($doc);
+ $pointers = $xpath->query('@content', $node);
+ //var_dump($pointers);
+ $pointer = $pointers->item(0);
+ $this->icbpecho("Page pointer:\t" . $pointer->nodeValue . "\n");
+ $pointerValue = $pointer->nodeValue;
+ }
+ $path = $node->getNodePath();
+ $this->icbpecho("Path:\t\t$path\n");
+
+ $parentNode = $node->parentNode;
+ $this->icbpecho("Parent path:\t" . $parentNode->getNodePath() . "\n\n");
+
+ $parentNode = trim($parentNode->getNodePath());
+ $insert_result = db_insert('islandora_compound_batch')
+ ->fields(array(
+ 'object_id' => $object_id,
+ 'object_xpath' => $path,
+ 'parent_xpath' => $parentNode,
+ 'pageptr' => $pointerValue,
+ 'parent_pid' => $compound_object_pid,
+ 'batch_id' => $batch_id
+ ))->execute();
+
+ $this->icbpecho("Insert Result: $insert_result" . PHP_EOL);
+ }
+
+ }
+
+ } // End of addStrcutureData
}
\ No newline at end of file
diff --git a/includes/utilities.inc b/includes/utilities.inc
index 041757e0442d1a6035ec935504d8ff8ca7101efc..6c8cba1533a603a22d4193a2d301ae6fe3f6adee 100644
--- a/includes/utilities.inc
+++ b/includes/utilities.inc
@@ -103,5 +103,51 @@ class Utilies {
);
}
}
+
+ /**
+ * Recursively create XML string of directory structure/
+ * Based on psuedo-code from http://stackoverflow.com/a/15096721/850828
+ */
+ private function directoryXML($directory_path) {
+
+ // basenames to exclude.
+ $exclude_array = array('..', '.DS_Store', 'Thumbs.db', '.');
+
+ $dir_name = basename($directory_path);
+ $xml = "<directory name='" . $dir_name . "'>";
+
+ $pathbase = pathinfo($directory_path, PATHINFO_BASENAME);
+
+ $stuffindirectory = scandir($directory_path);
+
+ foreach ($stuffindirectory as $subdirOrfile) {
+
+ $subdirOrfilepath = $directory_path . DIRECTORY_SEPARATOR . $subdirOrfile;
+
+ if (!in_array($subdirOrfile, $exclude_array) && is_file($subdirOrfilepath)) {
+ $xml .= "<file name='" . $subdirOrfile . "' />";
+ }
+
+ if (!in_array($subdirOrfile, $exclude_array) && is_dir($subdirOrfilepath)) {
+ $xml .= directoryXML($subdirOrfilepath);
+ }
+
+ }
+
+ $xml .= "</directory>";
+
+ return $xml;
+ }
+
+ public function createStructureXML($target_directory) {
+ $xmlstring = "<tree>";
+ $xmlstring .= $this->directoryXML($target_directory);
+ $xmlstring .= "</tree>";
+ $xml = new DOMDocument( "1.0");
+ $xml->loadXML($xmlstring);
+ $xml->formatOutput = TRUE;
+ return $xml->saveXML();
+ }
+
}
\ No newline at end of file