source: main/trunk/gli/src/org/greenstone/gatherer/metadata/DocXMLFile.java@ 34510

/**
 *############################################################################
 * A component of the Greenstone Librarian Interface, part of the Greenstone
 * digital library suite from the New Zealand Digital Library Project at the
 * University of Waikato, New Zealand.
 *
 * Author: Michael Dewsnip, NZDL Project, University of Waikato, NZ
 *
 * Copyright (C) 2004 New Zealand Digital Library Project
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 *############################################################################
 */

package org.greenstone.gatherer.metadata;


import java.io.*;
import java.util.*;
import java.net.URLDecoder;
import org.greenstone.gatherer.DebugStream;
import org.greenstone.gatherer.Gatherer;
import org.greenstone.gatherer.util.Utility;

//import org.greenstone.gatherer.feedback.Base64; // decode() from Base64 didn't work
import org.apache.commons.codec.binary.Base64; // decoding from Base64 works

/** This class represents one doc.xml file */

public abstract class DocXMLFile extends File
{
    static boolean isWin = Utility.isWindows();
    
    protected HashMap source_file_name_to_description_elements_mapping = new HashMap();
    
    protected final String MetadataWrap;
    protected final String MetadataItem;

    protected final String FILE_RENAME_METHOD_NONE = "none";
    protected final String FILE_RENAME_METHOD_URL = "url";
    protected final String FILE_RENAME_METHOD_BASE64 = "base64";

    public DocXMLFile(String doc_xml_file_path, String metaWrap, String metaItem)
    {
    super(doc_xml_file_path);
    this.MetadataWrap = metaWrap;
    this.MetadataItem = metaItem;
    }

    /** On Windows, file_relative_path will be hex-encoded for codepts beyond ASCII.
     * But keys into the source_file_name_to_description_elements_mapping will then also match on Windows */
    public ArrayList getMetadataExtractedFromFile(File file, String file_relative_path)
    {
    // Build up a list of metadata extracted from this file
    ArrayList metadata_values = new ArrayList();
    
    ///for (Object relFilename : source_file_name_to_description_elements_mapping.keySet()) {
    ///    System.err.println("\n@@@ relFilename: " + relFilename);
    ///}
    
    // Check whether this file (i.e. doc.xml or docmets.xml on inheritance) file contains extracted metadata for the specified file
    ArrayList description_elements_list = (ArrayList) source_file_name_to_description_elements_mapping.get(file_relative_path);
    if (description_elements_list == null) {
            // ...it doesn't
            ///System.err.println("Unable to find meta for file path form " + file_relative_path);
            return metadata_values; // we're done
    } ///else { System.err.println("@@@ file rel path: " + file_relative_path + " matched" ); }

    MetadataSet extracted_metadata_set = MetadataSetManager.getMetadataSet(MetadataSetManager.EXTRACTED_METADATA_NAMESPACE);

    // Parse the file
    DebugStream.println("Applicable file: " + this);
    try {
        BufferedReader buffered_reader = new BufferedReader(new InputStreamReader(new FileInputStream(this), "UTF-8"));

        int description_element_num = 0;
        int next_description_element_start = ((Integer) description_elements_list.get(description_element_num)).intValue();
        boolean in_relevant_description_element = false;

        String line = null;
        for (int line_num = 0; (line = buffered_reader.readLine()) != null; line_num++) {
        // Check if this line contains the start of a relevant "Description" element
        // (mets:xmlData in METS parlance, Description in GreenstoneArchive format) 
        if (line_num == next_description_element_start) {
            in_relevant_description_element = true;
            continue;
        }

        // If we're not in a relevant Description element we don't care about anything
        if (in_relevant_description_element == false) {
            continue;
        }

        // Check if this line contains the end of the relevant Description element
        if (line.indexOf("</"+MetadataWrap+">") != -1) {
            description_element_num++;
            if (description_element_num == description_elements_list.size()) {
            break;
            }

            next_description_element_start = ((Integer) description_elements_list.get(description_element_num)).intValue();
            in_relevant_description_element = false;
            continue;
        }

        // If this line doesn't contain a complete Metadata element, we're not interested
        if (line.indexOf("<"+MetadataItem+" ") == -1 || line.indexOf("</"+MetadataItem+">") == -1) {
            continue;
        }

        // Extract the metadata element name
        int name_index = line.indexOf(" name=\"") + " name=\"".length();
        String metadata_element_name_full = line.substring(name_index, line.indexOf("\"", name_index));

        // If the metadata has a namespace it isn't extracted metadata, so we're not interested
        // Actually, if it is ex. then we are interested 
        String metadata_set_namespace = MetadataTools.getMetadataSetNamespace(metadata_element_name_full);
        
        if (!metadata_set_namespace.equals("") && !metadata_set_namespace.equals("ex")) {
            continue;
        }

        // Extracted metadata!
        // do it like this just in case we have ex.
        String metadata_element_name = MetadataTools.getMetadataElementName(metadata_element_name_full);

        // We completely ignore bibliographic data
        if (metadata_element_name.equals("SourceSegment")) {
            buffered_reader.close();
            return new ArrayList();
        }

        // Ignore metadata starting with gsdl (used to be lowercase metadata and /metadata)
        if (metadata_element_name.startsWith("gsdl")) {
            continue;
        }

        MetadataElement metadata_element = extracted_metadata_set.getMetadataElementWithName(metadata_element_name);

        // Value trees are not stored for extracted metadata, so create a new value tree node now
        int value_index = line.indexOf(">", name_index) + ">".length();
        String metadata_element_value = line.substring(value_index, line.lastIndexOf("</"+MetadataItem+">"));

        metadata_element.addMetadataValue(metadata_element_value);
        MetadataValueTreeNode metadata_value_tree_node = metadata_element.getMetadataValueTreeNode(metadata_element_value);

        // Add the new metadata value to the list
        MetadataValue metadata_value = new MetadataValue(metadata_element, metadata_value_tree_node);
        metadata_values.add(metadata_value);
        }

        buffered_reader.close();
    }
    catch (FileNotFoundException exception) {
        DebugStream.printStackTrace(exception);
    }
    catch (IOException exception) {
        DebugStream.printStackTrace(exception);
    }

    return metadata_values;
    }




    /**
     * Every file must be skimmed when a collection is opened, for two reasons:
     *   - To build a mapping from source file to its corresponding doc.xml file
     *   - To get a complete list of all extracted metadata elements
     */
    public void skimFile()
    {
    String fileRenameMethod = null;
    String gsdlsourcefilename_value = null;
    boolean is_unix_path = false;
    int description_element_start_gsdlsourcefilename_value = -1;
    
    MetadataSet extracted_metadata_set = MetadataSetManager.getMetadataSet(MetadataSetManager.EXTRACTED_METADATA_NAMESPACE);

    // Skim the file as quickly as possible (don't parse as XML), looking at the Metadata elements
    DebugStream.println("Skimming " + this + "...");
    try {
        BufferedReader buffered_reader = new BufferedReader(new InputStreamReader(new FileInputStream(this), "UTF-8"));
        int description_element_start = -1;

        String line = null;
        for (int line_num = 0; (line = buffered_reader.readLine()) != null; line_num++) {
        // This line contains the start of a "MetadataWrap" element 
        // (mets:xmlData in METS parlance, Description in GreenstoneArchive format) 
        if (line.indexOf("<"+MetadataWrap+">") != -1) {
            if (description_element_start != -1) {
            System.err.println("Parse error: previous " + MetadataWrap + " element unfinished!");
            }
            description_element_start = line_num;
            continue;
        }

        // This line contains the end of a "MetadataWrap" element
        if (line.indexOf("</"+MetadataWrap+">") != -1) {
            if (description_element_start == -1) {
            System.err.println("Parse error: "+MetadataWrap+" element unstarted!");
            }
            description_element_start = -1;
            continue;
        }

        // If we're not in a"MetadataWrap" element there shouldn't be any Metadata elements
        if (description_element_start == -1) {
            continue;
        }

        // This line doesn't contain a Metadata element, so we're not interested
        if (line.indexOf("<"+MetadataItem+" ") == -1) {
            DebugStream.println("Warning: "+MetadataWrap+" element line doesn't contain Metadata element.");
            continue;
        }

        // Extract the metadata element name
        int name_index = line.indexOf(" name=\"") + " name=\"".length();
        String metadata_element_name_full = line.substring(name_index, line.indexOf("\"", name_index));

        // If the metadata has a namespace it isn't extracted metadata, so we're not interested
        String metadata_set_namespace = MetadataTools.getMetadataSetNamespace(metadata_element_name_full);
        if (!metadata_set_namespace.equals("") && !metadata_set_namespace.equals("ex")) {
            continue;
        }

        // Extracted metadata! May have ex. so make sure we remove that
        String metadata_element_name = MetadataTools.getMetadataElementName(metadata_element_name_full);
        if(metadata_element_name.equals("gsdlsourcefilerenamemethod")) {
            // Extract the element value
            int value_index = line.indexOf(">", name_index) + ">".length();
            fileRenameMethod = line.substring(value_index, line.indexOf("<", value_index));         
        }
        
        // Note which file this is for
        else if (metadata_element_name.equals("gsdlsourcefilename")) {
            // the gsdlsourcefilename metadata field may be encoded by the encoding denoted
            // in fileRenameMethod (and will need decoding)
            
            // Extract the gsdlsourcefilename element value
            int value_index = line.indexOf(">", name_index) + ">".length();
            gsdlsourcefilename_value = line.substring(value_index, line.indexOf("<", value_index));
            
            // We're only interested in the path relative to the import folder          
            int import_index = gsdlsourcefilename_value.indexOf("import");
            if (import_index != -1) {
            
                ///System.err.println("@@@@ Found gsdlsourcefilename: " + gsdlsourcefilename_value);
                ///System.err.println("@@@@ Found description_element_start_gsdlsourcefilename_value: " + description_element_start);
                description_element_start_gsdlsourcefilename_value = description_element_start;
                
            }

            // Warn about an odd gsdlsourcefilename, except if it is the Greenstone "tmp" directory or
            // (as in the case of using FLI) if it is the etc/collect.cfg or etc/collectionConfig.xml file
            // which are the gsdlsourcefilenames for the fedora digital object representing a collection.
            // This (tmp dir) is true when the source files come from a zip file processed by ZIPPlug, for example
            else if (gsdlsourcefilename_value.indexOf("tmp") == -1 
                 && !gsdlsourcefilename_value.endsWith("collect.cfg")
                 && !gsdlsourcefilename_value.endsWith("collectionConfig.xml")) {
                // We don't really know what is going on...
                System.err.println("Warning: Could not understand gsdlsourcefilename " + gsdlsourcefilename_value);
            }
        }
        
        // Ignore metadata starting with gsdl (used to be lowercase metadata and /metadata)
        if (metadata_element_name.startsWith("gsdl")) {
            continue;
        }

        MetadataElement metadata_element = extracted_metadata_set.getMetadataElementWithName(metadata_element_name);
        if (metadata_element == null) {
            // This element isn't defined in ex.mds, so create it for this session
            DebugStream.println("Extracted metadata element not defined: " + metadata_element_name);
            extracted_metadata_set.addMetadataElementForThisSession(metadata_element_name);
        }
        }

        buffered_reader.close();

        // Work out if is_unix_path
        int import_index = gsdlsourcefilename_value.indexOf("import");
        if (import_index != -1) {
            String tempStr = gsdlsourcefilename_value.substring(import_index + "import".length());
            is_unix_path = tempStr.startsWith("/");
        }
        // We're only interested in the path relative to the import folder
        // Lop off "import" folder prefix
        gsdlsourcefilename_value = adjustForRelativeToImportDir(gsdlsourcefilename_value);

        // Now that we're done skimming, we actually need to decode gsdlsourcefilename
        // based on whatever fileRenameMethod was used to encode it, so that we can
        // at last properly compare against filenames on the file system
        // in order to load the correct ex.meta for the file.
        // Now that we should have both gsdlsourcefilename AND fileRenameMethod set,
        // we can finally perform the decoding of gsdlsourcefilename.       
        if(fileRenameMethod == null) {
            fileRenameMethod = FILE_RENAME_METHOD_URL; // default for building
        }
        if(!fileRenameMethod.equals(FILE_RENAME_METHOD_NONE)) {         
            gsdlsourcefilename_value = decodeSourceFilename(gsdlsourcefilename_value, fileRenameMethod, is_unix_path);          
        }

        // Now we can finally put the gsdlsourcefilename path relative to import dir into the hashmap
        ///System.err.println("@@@ into map: " + gsdlsourcefilename_value);
        if (source_file_name_to_description_elements_mapping.get(gsdlsourcefilename_value) == null) {
            source_file_name_to_description_elements_mapping.put(gsdlsourcefilename_value, new ArrayList());
        }
        ((ArrayList) source_file_name_to_description_elements_mapping.get(gsdlsourcefilename_value)).add(new Integer(description_element_start_gsdlsourcefilename_value));      
        
        // Next, if Windows, check if dealing with Win 8.3 Short Filename
        // In that case, convert short file name to long filename - works only if the file exists       
        if(isWin /*&& gsdlsourcefilename_value.indexOf("~") != -1*/) {
            
            String long_gsdlsourcefilename = gsdlsourcefilename_value;
            
            // gsdlsourcefilename is stored from import folder onwards: import/opt_subdir/filename.ext
            // This may contain Win 8.3 shortening. To get Win Long filename, prefix current collection dir
            // and if resulting file exists, getCanonicalPath() which produces Win Long filename.
            File currentCollectionFolder = Gatherer.c_man.getCollection().getCollectionDirectory();
            File f = new File(currentCollectionFolder, "import" + File.separator + gsdlsourcefilename_value); // should work even if linux style slashes in gsdlsourcefilename_value
            ///System.err.println("### file: " + f.getAbsolutePath());
            
            if(f.exists()) {
                long_gsdlsourcefilename = f.getCanonicalPath();
                ///System.err.println("### canon: " + long_gsdlsourcefilename);             
            } // else couldn't find a version of the filename stored in doc.xml that exists, giving up, leave gsdlsourcefilename_value as is
            
            // Again, we're only interested in the path relative to the import folder
            long_gsdlsourcefilename = adjustForRelativeToImportDir(long_gsdlsourcefilename);
            if(!gsdlsourcefilename_value.equals(long_gsdlsourcefilename)) { // truly distinct Win long and short file names
                // Put a copy of the ref to gsdlsourcefilename's metadata list under the long filename as well
                ///System.err.println("@@@ long filename into map: " + long_gsdlsourcefilename);        
                Object arrList = source_file_name_to_description_elements_mapping.get(gsdlsourcefilename_value);
                source_file_name_to_description_elements_mapping.put(long_gsdlsourcefilename, arrList);
            }
        }
        
    }
    catch (FileNotFoundException exception) {
        DebugStream.printStackTrace(exception);
    }
    catch (IOException exception) {
        DebugStream.printStackTrace(exception);
    } catch (Exception exception) { // e.g. exception decoding gsdlsourcefilename
        DebugStream.printStackTrace(exception);
    }
    }
    
    private String adjustForRelativeToImportDir(String gsdlsourcefilename_value) {
        int import_index = gsdlsourcefilename_value.indexOf("import");
        if (import_index != -1) {
            gsdlsourcefilename_value = gsdlsourcefilename_value.substring(import_index + "import".length());

            boolean is_unix_path = gsdlsourcefilename_value.startsWith("/");
            gsdlsourcefilename_value = gsdlsourcefilename_value.substring(1);

            // (Will decode gsdlsourcefilename at end of this method, once we know
            // for certain the fileRenameMethod that was used to encode it.)

            // Make sure the path matches the OS that is running
            if (is_unix_path && isWin) {
                // Convert path from Unix to Windows
                gsdlsourcefilename_value = gsdlsourcefilename_value.replaceAll("\\/", "\\\\");
            }
            else if (!is_unix_path && !isWin) {
                // Convert path from Windows to Unix
                gsdlsourcefilename_value = gsdlsourcefilename_value.replaceAll("\\\\", "/");
            }
        }
        return gsdlsourcefilename_value;
    }
    
    protected String decodeSourceFilename(String relative_sourcefile_path,
                      String encodingMethod, boolean is_unix_path)
    throws Exception
    {   

    ///System.err.println("*** relative_sourcefile_path: " + relative_sourcefile_path);

    // First get the file extension. Both in Base64 and URL encoded strings,
    // the full-stop character (.) doesn't get encoded.
    // That means getting the file extension is straightforward.
    
    // Valid base64: "The 64 characters (hence the name Base64) are 10 digits,
    // 26 lowercase characters, 26 uppercase characters as well as the
    // Plus sign (+) and the Forward Slash (/).
    int fullstop = relative_sourcefile_path.indexOf(".");
    String file_ext = "";
    if(fullstop != -1) {
        file_ext = relative_sourcefile_path.substring(fullstop);
        relative_sourcefile_path = relative_sourcefile_path.substring(0, fullstop);
    }
    
    String[] importFilePathParts = DocXMLFile.getFilePathParts(relative_sourcefile_path, is_unix_path);
    
    String decoded_gsdlsourcefilename = "";

    String separator = is_unix_path ? "/" : "\\";
    for(int i = 0; i < importFilePathParts.length; i++) {
        String decoded_filePathPart = "";
        if(encodingMethod.equals(FILE_RENAME_METHOD_URL)) {
        // URL decode each part of gsdlsourcefilename.
        // Need to set the decoder to use the default system encoding
        // This is stored in the System's file.encoding property.
        decoded_filePathPart = URLDecoder.decode(importFilePathParts[i], System.getProperty("file.encoding"));
        }
        else{ // if(encodingMethod.equals(FILE_RENAME_METHOD_BASE64)) {
        // Decoding with org.greenstone.gatherer.feedback.Base64 didn't work
        //byte[] bytes = org.greenstone.gatherer.feedback.Base64.decode(importFilePathParts[i]);
        // Using org.apache.commons.codec.binary.Base64 instead
        // https://commons.apache.org/proper/commons-codec/archives/1.7/apidocs/org/apache/commons/codec/binary/Base64.html
        // General info: https://stackoverflow.com/questions/43089541/difference-between-basic-and-url-base64-encoding-in-java-8
        byte[] bytes = Base64.decodeBase64(importFilePathParts[i].getBytes());
        ///System.err.println("Got base64 string: " + importFilePathParts[i]);
        ///System.err.println("Decoded from base64 to bytes: " + new String(bytes, System.getProperty("file.encoding")));
        // Using system file.encoding to interpret the resulting bytestring as a String,
        // just as we always did with URL decoding method
        decoded_filePathPart = (bytes == null) ? importFilePathParts[i] : new String(bytes, System.getProperty("file.encoding"));
        }       
        
        if(i == 0) {
        decoded_gsdlsourcefilename = decoded_filePathPart;
        } else {
        decoded_gsdlsourcefilename = decoded_gsdlsourcefilename + separator + decoded_filePathPart;
        }
        ///System.err.println("Built up: " + decoded_gsdlsourcefilename);
    }

    // add the file extension back in
    decoded_gsdlsourcefilename += file_ext;
    
    ///System.err.println("@@@@ decoded_gsdlsourcefilename: " + Utility.debugUnicodeString(decoded_gsdlsourcefilename));

    return decoded_gsdlsourcefilename;
    }    
    
    /**
     * Given a filepath, returns the parts between each file separator as an array.
     * For example, "/Users/me/pinky.txt" should return {"Users", "me", "pinky.txt"};
     */
    private static String[] getFilePathParts(String filepath, boolean is_unix_path) {   
    StringTokenizer tok;
    if(is_unix_path) {
        tok = new StringTokenizer(filepath, "/");           
    } else {
        tok = new StringTokenizer(filepath, "\\");          
    }
    String[] parts;
    int count = tok.countTokens();      
    if(count <= 0) {
        parts = new String[]{filepath};         
    } else {
        int i = 0;
        parts = new String[count];
        while(tok.hasMoreTokens()) {
        parts[i] = tok.nextToken();
        //System.err.println("Next part: " + parts[i]);
        i++;
        }       
    }
    return parts;       
    }
    
    /*
    public ArrayList getMetadataExtractedFromFile(File file)
    {
    // Build up a list of metadata extracted from this file
    ArrayList metadata_values = new ArrayList();

    String file_relative_path = file.getAbsolutePath();
    int import_index = file_relative_path.indexOf("import");
    if (import_index != -1) {
        file_relative_path = file_relative_path.substring(import_index + "import".length() + 1);
    }

    // Check whether this doc.xml file contains extracted metadata for the specified file
    ArrayList description_elements_list = (ArrayList) source_file_name_to_description_elements_mapping.get(file_relative_path);
    if (description_elements_list == null) {
        // ...it doesn't
        return metadata_values;
    }

    MetadataSet extracted_metadata_set = MetadataSetManager.getMetadataSet(MetadataSetManager.EXTRACTED_METADATA_NAMESPACE);

    // Parse the doc.xml file
    DebugStream.println("Applicable doc.xml file: " + this);
    try {
        BufferedReader buffered_reader = new BufferedReader(new FileReader(this));

        int description_element_num = 0;
        int next_description_element_start = ((Integer) description_elements_list.get(description_element_num)).intValue();
        boolean in_relevant_description_element = false;

        String line = null;
        for (int line_num = 0; (line = buffered_reader.readLine()) != null; line_num++) {
        // Check if this line contains the start of a relevant Description element
        if (line_num == next_description_element_start) {
            in_relevant_description_element = true;
            continue;
        }

        // If we're not in a relevant Description element we don't care about anything
        if (in_relevant_description_element == false) {
            continue;
        }

        // Check if this line contains the end of the relevant Description element
        if (line.indexOf("</Description>") != -1) {
            description_element_num++;
            if (description_element_num == description_elements_list.size()) {
            break;
            }

            next_description_element_start = ((Integer) description_elements_list.get(description_element_num)).intValue();
            in_relevant_description_element = false;
            continue;
        }

        // If this line doesn't contain a complete Metadata element, we're not interested
        if (line.indexOf("<Metadata ") == -1 || line.indexOf("</Metadata>") == -1) {
            continue;
        }

        // Extract the metadata element name
        int name_index = line.indexOf(" name=\"") + " name=\"".length();
        String metadata_element_name_full = line.substring(name_index, line.indexOf("\"", name_index));

        // If the metadata has a namespace it isn't extracted metadata, so we're not interested
        String metadata_set_namespace = MetadataTools.getMetadataSetNamespace(metadata_element_name_full);
        if (!metadata_set_namespace.equals("")) {
            continue;
        }

        // Extracted metadata!
        String metadata_element_name = metadata_element_name_full;

        // We completely ignore bibliographic data
        if (metadata_element_name.equals("SourceSegment")) {
            buffered_reader.close();
            return new ArrayList();
        }

        // Ignore metadata starting with gsdl (used to be lowercase metadata and /metadata)
        if (metadata_element_name.startsWith("gsdl")) {
            continue;
        }

        MetadataElement metadata_element = extracted_metadata_set.getMetadataElementWithName(metadata_element_name);

        // Value trees are not stored for extracted metadata, so create a new value tree node now
        int value_index = line.indexOf(">", name_index) + ">".length();
        String metadata_element_value = line.substring(value_index, line.lastIndexOf("</Metadata>"));

        metadata_element.addMetadataValue(metadata_element_value);
        MetadataValueTreeNode metadata_value_tree_node = metadata_element.getMetadataValueTreeNode(metadata_element_value);

        // Add the new metadata value to the list
        MetadataValue metadata_value = new MetadataValue(metadata_element, metadata_value_tree_node);
        metadata_values.add(metadata_value);
        }

        buffered_reader.close();
    }
    catch (FileNotFoundException exception) {
        DebugStream.printStackTrace(exception);
    }
    catch (IOException exception) {
        DebugStream.printStackTrace(exception);
    }

    return metadata_values;
    }

    */

    /**
     * Every doc.xml file must be skimmed when a collection is opened, for two reasons:
     *   - To build a mapping from source file to its corresponding doc.xml file
     *   - To get a complete list of all extracted metadata elements
     */
    /*
    public void skimFile()
    {
    MetadataSet extracted_metadata_set = MetadataSetManager.getMetadataSet(MetadataSetManager.EXTRACTED_METADATA_NAMESPACE);

    // Skim the doc.xml file as quickly as possible (don't parse as XML), looking at the Metadata elements
    DebugStream.println("Skimming " + this + "...");
    try {
        BufferedReader buffered_reader = new BufferedReader(new FileReader(this));
        int description_element_start = -1;

        String line = null;
        for (int line_num = 0; (line = buffered_reader.readLine()) != null; line_num++) {
        // This line contains the start of a Description element
        if (line.indexOf("<Description>") != -1) {
            if (description_element_start != -1) {
            System.err.println("Parse error: previous Description element unfinished!");
            }
            description_element_start = line_num;
            continue;
        }

        // This line contains the end of a Description element
        if (line.indexOf("</Description>") != -1) {
            if (description_element_start == -1) {
            System.err.println("Parse error: Description element unstarted!");
            }
            description_element_start = -1;
            continue;
        }

        // If we're not in a Description element there shouldn't be any Metadata elements
        if (description_element_start == -1) {
            continue;
        }

        // This line doesn't contain a Metadata element, so we're not interested
        if (line.indexOf("<Metadata ") == -1) {
            DebugStream.println("Warning: Description element line doesn't contain Metadata element.");
            continue;
        }

        // Extract the metadata element name
        int name_index = line.indexOf(" name=\"") + " name=\"".length();
        String metadata_element_name_full = line.substring(name_index, line.indexOf("\"", name_index));

        // If the metadata has a namespace it isn't extracted metadata, so we're not interested
        String metadata_set_namespace = MetadataTools.getMetadataSetNamespace(metadata_element_name_full);
        if (!metadata_set_namespace.equals("")) {
            continue;
        }

        // Extracted metadata!
        String metadata_element_name = metadata_element_name_full;

        // Note which file this doc.xml is for
        if (metadata_element_name.equals("gsdlsourcefilename")) {
            // Extract the gsdlsourcefilename element value
            int value_index = line.indexOf(">", name_index) + ">".length();
            String gsdlsourcefilename_value = line.substring(value_index, line.indexOf("<", value_index));

            // We're only interested in the path relative to the import folder
            int import_index = gsdlsourcefilename_value.indexOf("import");
            if (import_index != -1) {
            gsdlsourcefilename_value = gsdlsourcefilename_value.substring(import_index + "import".length());

            boolean is_unix_path = gsdlsourcefilename_value.startsWith("/");
            gsdlsourcefilename_value = gsdlsourcefilename_value.substring(1);

            // URL decode gsdlsourcefilename. Need to set the decoder to use the default system encoding
            // This is stored in the System's file.encoding property.
            gsdlsourcefilename_value = URLDecoder.decode(gsdlsourcefilename_value, System.getProperty("file.encoding")); 

            // Make sure the path matches the OS that is running
            if (is_unix_path && isWin) {
                // Convert path from Unix to Windows
                gsdlsourcefilename_value = gsdlsourcefilename_value.replaceAll("\\/", "\\\\");
            }
            else if (!is_unix_path && !isWin) {
                // Convert path from Windows to Unix
                gsdlsourcefilename_value = gsdlsourcefilename_value.replaceAll("\\\\", "/");
            }

            // Remember this for quick access later
            if (source_file_name_to_description_elements_mapping.get(gsdlsourcefilename_value) == null) {
                source_file_name_to_description_elements_mapping.put(gsdlsourcefilename_value, new ArrayList());
            }

            ((ArrayList) source_file_name_to_description_elements_mapping.get(gsdlsourcefilename_value)).add(new Integer(description_element_start));
            }

            // Warn about an odd gsdlsourcefilename, except if it is the Greenstone "tmp" directory
            // This is true when the source files come from a zip file processed by ZIPPlug, for example
            else if (gsdlsourcefilename_value.indexOf("tmp") == -1) {
            // We don't really know what is going on...
            System.err.println("Warning: Could not understand gsdlsourcefilename " + gsdlsourcefilename_value);
            }
        }

        // Ignore metadata starting with gsdl (used to be lowercase metadata and /metadata)
        if (metadata_element_name.startsWith("gsdl")) {
            continue;
        }

        MetadataElement metadata_element = extracted_metadata_set.getMetadataElementWithName(metadata_element_name);
        if (metadata_element == null) {
            // This element isn't defined in ex.mds, so create it for this session
            DebugStream.println("Extracted metadata element not defined: " + metadata_element_name);
            extracted_metadata_set.addMetadataElementForThisSession(metadata_element_name);
        }
        }

        buffered_reader.close();
    }
    catch (FileNotFoundException exception) {
        DebugStream.printStackTrace(exception);
    }
    catch (IOException exception) {
        DebugStream.printStackTrace(exception);
    }
    }
    */

}