Changeset 9195

Timestamp:

2005-02-25T15:16:56+13:00 (19 years ago)

Author:

mdewsnip

Message:

Now rips everything except the metadata out of doc.xml files before sending them back. By Matthew Whyte.

File:

• trunk/gli/src/org/greenstone/gatherer/util/Utility.java (modified) (6 diffs)

trunk/gli/src/org/greenstone/gatherer/util/Utility.java

@@ -45 +45 @@
 import javax.swing.tree.*;
 import org.apache.xerces.parsers.*;
+import org.apache.xerces.dom.DocumentImpl;
 import org.apache.xml.serialize.*;
 import org.greenstone.gatherer.Configuration;
@@ -70 +71 @@
     static final public int BUFFER_SIZE = FACTOR * 1024;
     /** Definition of an important directory name, in this case the base dir, or the working directory of the Gatherer. */
+    /** The regular expression that marks if we only want doc.xml files. Will this work in Windows?? --Matthew */
+    static final public String METADATA_ONLY_EXPR = "\".*" + File.separator + "doc\\.xml\"";
+
     static public String BASE_DIR  = System.getProperty("user.dir") + File.separator;
     static final public String BUILD_CFG_FILENAME = "build.cfg";
@@ -619 +623 @@
     static protected boolean zipFunc (ZipOutputStream zos, String file_path, int prefix_strip, GShell source, boolean encountered_file, String accept_expr, String reject_expr) 
     {
-    
+    String new_file_path = file_path;
+
     if((reject_expr != "") && (file_path.matches(reject_expr))) {
         // matches reject expression
@@ -630 +635 @@
         DebugStream.println("File \'" + file_path + "\' doesn't match accept expression \'" + accept_expr + "\'");
         return encountered_file;
+    }
+
+    //Special case: we only want the metadata. Parse the xml files.
+    if(accept_expr.compareTo(".*doc.xml") == 0) {
+        System.err.println("Only want to get metadata"); //debug
+        Document old_document = XMLTools.parseXMLFile(new File(file_path));
+        NodeList content = old_document.getElementsByTagName("Content");
+
+        try {
+        for(int i=content.getLength()-1; i >= 0; i--) {
+            //Remove all content - just leave the metadata.
+            content.item(i).getParentNode().removeChild(content.item(i));
+        }
+        }
+        catch (Exception ex) {
+        ex.printStackTrace();
+        }
+
+        new_file_path = file_path.substring(0, file_path.lastIndexOf(File.separator) + 1) + "new.xml";
+        //Is there any better way than writing out as a new file, then reading back in??
+        XMLTools.writeXMLFile(new File(new_file_path), old_document);
     }
 
@@ -650 +676 @@
 
         // Create a file input stream and a buffered input stream.
-        FileInputStream fis = new FileInputStream(file_path);
+        FileInputStream fis = new FileInputStream(new_file_path);
         BufferedInputStream bis = new BufferedInputStream(fis);
         
@@ -783 +809 @@
         else {
             String zip_full_file = zip_dir_or_file; 
-
             String zip_path = zip_full_file.substring(prefix_strip);