Changeset 4287

Timestamp:

2003-05-19T16:00:47+12:00 (21 years ago)

Author:

kjdon

Message:

added some code for highlighting search terms in the document. this involves a call to teh query service, so may not be a good way to do it. phrase highlighting not done properly

File:

• trunk/gsdl3/src/java/org/greenstone/gsdl3/action/DocumentAction.java (modified) (4 diffs)

trunk/gsdl3/src/java/org/greenstone/gsdl3/action/DocumentAction.java

@@ -27 +27 @@
 import org.w3c.dom.Element; 
 import org.w3c.dom.Node; 
+import org.w3c.dom.Text;
 import org.w3c.dom.NodeList;
 
 // General Java classes
 import java.util.HashMap;
+import java.util.HashSet;
 import java.io.File;
 
@@ -236 +238 @@
     Element dc_response_doc = (Element) GSXML.getNodeByPath(dc_response_message, path);
     Element dc_response_doc_content = (Element) GSXML.getChildByTagName(dc_response_doc, GSXML.NODE_CONTENT_ELEM);
+            
+    boolean highlight_query_terms = true;
+    if (highlight_query_terms) {
+        dc_response_doc.removeChild(dc_response_doc_content);
+        
+        dc_response_doc_content = highlightQueryTerms(request, dc_response_doc_content);
+        dc_response_doc.appendChild(dc_response_doc.getOwnerDocument().importNode(dc_response_doc_content, true));
+    }
     if (provide_annotations) {
         // now we can modifiy the response doc if needed
@@ -255 +265 @@
         path = GSPath.createPath(links);
         dc_response_doc_content = (Element)GSXML.getNodeByPath(enrich_response, path);
+        
         } 
     }
@@ -355 +366 @@
         
     }
+
+    /** this involves a bit of a hack to get the equivalent query terms - has to requery the query service - uses the last selected service name. (if it ends in query). should this action do the query or shoul dit send a message to the query action? but that will involve lots of extra stuff. also doesn't handle phrases properly - just highlights all teh terms found in the text.
+     */
+    protected Element highlightQueryTerms(Element request, Element dc_response_doc_content) {
+
+    // do the query again to get term info 
+    Element cgi_param_list = (Element)GSXML.getChildByTagName(request, GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
+    HashMap params = GSXML.extractParams(cgi_param_list, false);
+    
+    String service_name = (String)params.get(GSCGI.SERVICE_ARG);
+    if (service_name == null || !service_name.endsWith("Query")) { // hack for now - we only do highlighting if we were in a query last - ie not if we were in a browse thingy
+        return dc_response_doc_content;
+    }
+    String collection = (String)params.get(GSCGI.COLLECTION_ARG);
+    String lang = request.getAttribute(GSXML.LANG_ATT);
+    String to = GSPath.appendLink(collection, service_name);
+    
+    Element mr_query_message = doc_.createElement(GSXML.MESSAGE_ELEM);
+    Element mr_query_request = GSXML.createBasicRequest(doc_, GSXML.REQUEST_TYPE_PROCESS, to, lang);
+    mr_query_message.appendChild(mr_query_request);
+    
+    // paramList
+    Element query_param_list = (Element)doc_.importNode(cgi_param_list, true);
+    mr_query_request.appendChild(query_param_list);
+
+    // do the query
+        Element mr_query_response = (Element)mr_.process(mr_query_message);
+
+    String path = GSPath.appendLink(GSXML.RESPONSE_ELEM, GSXML.TERM_ELEM+GSXML.LIST_MODIFIER);
+    Element query_term_info_list = (Element) GSXML.getNodeByPath(mr_query_response, path);
+    if (query_term_info_list == null) {
+        // no term info
+        System.err.println("DocumentAction: Warning: No query term information.\n");
+        return dc_response_doc_content;
+    } 
+
+    NodeList equivs = query_term_info_list.getElementsByTagName("equivTermList");
+    HashSet all_terms = new HashSet();
+    for (int i=0; i<equivs.getLength(); i++) {
+        
+        // get the terms
+        String [] terms = GSXML.getAttributeValuesFromList((Element)equivs.item(i), GSXML.NAME_ATT);
+        for (int j=0; j<terms.length; j++) {
+        
+        all_terms.add(terms[j]);
+        }
+    }
+
+    String content = GSXML.getNodeText(dc_response_doc_content);
+
+    StringBuffer temp = new StringBuffer();
+    StringBuffer new_content = new StringBuffer();
+
+    for (int i=0; i<content.length(); i++) {
+        char c = content.charAt(i);
+        if (Character.isLetterOrDigit(c)) {
+        // not word boundary
+        temp.append(c);
+        } else {
+        // word boundary
+        // add the last word if there was one
+        if (temp.length()>0) {
+            if (all_terms.contains(temp.toString())) {
+            new_content.append("<annotation type='query_term'>"+temp+"</annotation>");
+            } else {
+            new_content.append(temp);
+            }
+            temp.delete(0, temp.length());
+        }
+        if (c=='<') {
+            temp.append(c);
+            i++;
+            // skip over html
+            while (i<content.length() && content.charAt(i)!='>') {
+            temp.append(content.charAt(i));
+            i++;
+            }
+            temp.append(content.charAt(i));
+            new_content.append(GSXML.xmlSafe(temp.toString()));
+            temp.delete(0, temp.length());
+            
+        } else {
+            new_content.append(c);
+        }
+        }
+    }
+    
+    String content_string = "<nodeContent>"+new_content.toString()+"</nodeContent>";
+    Element content_elem = converter_.getDOM(content_string).getDocumentElement();
+    return content_elem;
+    }
 }