Changeset 13576 for trunk/gsdl3/src/java/org/greenstone/gsdl3/service/GS2LuceneRetrieve.java

Timestamp:

2007-01-11T14:55:04+13:00 (17 years ago)

Author:

kjdon

Message:

implemented the core functionality for this.

File:

• trunk/gsdl3/src/java/org/greenstone/gsdl3/service/GS2LuceneRetrieve.java (modified) (4 diffs)

trunk/gsdl3/src/java/org/greenstone/gsdl3/service/GS2LuceneRetrieve.java

@@ -20 +20 @@
 
 // Greenstone classes
-import org.greenstone.mgpp.*;
 import org.greenstone.gsdl3.core.GSException;
 import org.greenstone.gsdl3.util.GSFile;
 import org.greenstone.gsdl3.util.GSXML;
-
+import org.greenstone.gsdl3.util.GDBMWrapper;
+import org.greenstone.gsdl3.util.DBInfo;
+import org.greenstone.gsdl3.util.GSHTML;
+import org.greenstone.gsdl3.util.OID;
 // XML classes
+import org.w3c.dom.Document;
 import org.w3c.dom.Element;
 import org.w3c.dom.Text;
@@ -34 +37 @@
 import org.apache.log4j.Logger;
 
+/** Retrieve documents from a gs2 lucene collection. Note that this doesn't 
+    actually use lucene, as the documents are stored in XML files */
 public class GS2LuceneRetrieve
     extends AbstractGS2DocumentRetrieve
 {
-     static Logger logger = Logger.getLogger(org.greenstone.gsdl3.service.GS2LuceneRetrieve.class.getName());
+    
+    static Logger logger = Logger.getLogger(org.greenstone.gsdl3.service.GS2LuceneRetrieve.class.getName());
+    
+    
+    protected static final String DOC_LEVEL="Doc";
+    protected static final String SEC_LEVEL="Sec";
+    protected static final String ID_ATT = "gs2:id";
 
-     // Parameters used
+    // Parameters used
     private static final String LEVEL_PARAM = "level";
-
+    
     // Elements used in the config file that are specific to this class
     private static final String DEFAULT_LEVEL_ELEM = "defaultLevel";
-
-    private MGPPWrapper mgpp_src = null;
     
     private String default_level = null;
-    private String mgpp_textdir = null;
-    
+    private String text_dir = null;
+
+    private boolean text_available = true;
+    
     public GS2LuceneRetrieve() {
-    this.mgpp_src = new MGPPWrapper();
     }
-
+    
     public void cleanUp() {
     super.cleanUp();
-    this.mgpp_src.unloadIndexData();
     }
-
+    
     /** configure this service */
     public boolean configure(Element info, Element extra_info)
@@ -65 +74 @@
         return false;
     }
-
+    
     // Do specific configuration
     logger.info("Configuring GS2LuceneRetrieve...");
-
-    // TODO - is there anything we need to do here? 
-    // set up XML parser??
-
+    
+    text_dir = GSFile.collectionIndexDir(this.site_home, this.cluster_name) + File.separatorChar+"text"+File.separatorChar;
+    if (!(new File(text_dir).isDirectory())) {
+        logger.error("Text directory "+text_dir+" does not exist, will be unable to retrieve text for "+cluster_name);
+        text_available = false;
+        return true; // return true so that we still get the other services for the collection
+    }
+    // Get the default level out of <defaultLevel> (buildConfig.xml)
+    Element def = (Element) GSXML.getChildByTagName(info, DEFAULT_LEVEL_ELEM);
+    if (def != null) {
+        this.default_level = def.getAttribute(GSXML.NAME_ATT);
+    }
+    if (this.default_level == null || this.default_level.equals("")) {
+        logger.error("Default level not specified for "+this.cluster_name+", assuming "+DOC_LEVEL);
+        this.default_level = DOC_LEVEL;
+    }
+    
     return true;
- 
+    
     }
     
@@ -80 +102 @@
      * <nodeContent>text content or other elements</nodeContent>
      */
-    protected Element getNodeContent(String doc_id) throws GSException {
-
+    protected Element getNodeContent(String doc_id, String lang) throws GSException {   
+    String doc_content = getTextString("TextRetrievalError", lang);
+    try {
+        if (!text_available) {
+        throw new Exception("No text directory available");
+        }
+        
+        long doc_num = this.gdbm_src.OID2Docnum(doc_id);
+        if (doc_num == -1) {
+        throw new Exception("OID "+doc_id +" couldn't be converted to lucene doc num");
+        }
+    
+        DBInfo info=this.gdbm_src.getInfo(OID.getTop(doc_id));
+        if (info == null) {
+        throw new Exception("Couldn't get GDBM database entry for "+OID.getTop(doc_id));
+        }
+        
+        String archivedir=info.getInfo("archivedir");
+        File doc_xml_file = new File(text_dir+archivedir+File.separatorChar+"doc.xml");
+        if (!doc_xml_file.isFile()) {
+        throw new Exception("Doc XML file "+doc_xml_file.getPath()+" does not exist");
+        }
+        Document doc_xml_doc = this.converter.getDOM(doc_xml_file);
+        if (doc_xml_doc == null) {
+        throw new Exception("Couldn't parse file "+doc_xml_file.getPath());
+        }
+        Element full_document = doc_xml_doc.getDocumentElement();
+        if (full_document == null) {
+        throw new Exception("Couldn't parse file "+doc_xml_file.getPath());
+        }
+        Element current_section = null;
+        if (default_level.equals(DOC_LEVEL)) {
+        current_section = full_document;
+        } else {
+        current_section = GSXML.getNamedElement(full_document, SEC_LEVEL, ID_ATT, String.valueOf(doc_num));
+        }
+        if (current_section == null) {
+        throw new Exception("Couldn't find section "+ doc_num+" in file "+doc_xml_file.getPath());
+        }
+        doc_content = GSXML.getNodeText(current_section);
+        if (doc_content == null) {
+        doc_content = "";
+        } else {
+        doc_content = resolveTextMacros(doc_content, doc_id, lang);
+        } 
+    } catch (Exception e) {
+        logger.error("Error trying to get document text for "+doc_id+" in collection "+this.cluster_name+": "+e);
+    }
     
     Element content_node = this.doc.createElement(GSXML.NODE_CONTENT_ELEM);
-    
-    String doc_content = "";
-    try {
-        // TODO get the doc content from the index text dir
-    } catch (Exception e) {
-        logger.info("exception happended getting doc content for " + doc_id); 
-        doc_content = "dummy content for doc "+doc_id +"\n";
-
-    }
-
     Text t = this.doc.createTextNode(doc_content);
     content_node.appendChild(t);
     return content_node;
     }
-
-
 }