Changeset 267 for trunk/gsdl/src/recpt/documentaction.cpp

Timestamp:

1999-06-15T13:55:32+12:00 (25 years ago)

Author:

sjboddie

Message:

• got text highlighting working
• got multiple collections working (now outputs error message if an attempt is made to get a document when the "c" arg isn't set.

File:

• trunk/gsdl/src/recpt/documentaction.cpp (modified) (5 diffs)

trunk/gsdl/src/recpt/documentaction.cpp

@@ -12 +12 @@
 /*
    $Log$
+   Revision 1.4  1999/06/15 01:55:29  sjboddie
+   - got text highlighting working
+   - got multiple collections working (now outputs error message if an
+     attempt is made to get a document when the "c" arg isn't set.
+
    Revision 1.3  1999/06/08 04:29:37  sjboddie
    added argsinfo to the call to check_cgiargs to make it easy to set
@@ -38 +43 @@
 #include "browsetools.h"
 #include "OIDtools.h"
+#include "querytools.h"
+#include "unitool.h"
 
 documentaction::documentaction () {
@@ -163 +170 @@
   // _cgiargcl_
   
+  // can't do anything if collectproto is null (i.e. no collection was specified)
+  if (collectproto == NULL) return;
 
   outconvertclass text_t2ascii;
@@ -312 +321 @@
   // _javaimagescontent_    this is the javascript code to shove in to make the 
   //                        flashy buttons work
+
+  
+  // can't do anything if collectproto is null (i.e. no collection was specified)
+  if (collectproto == NULL) return;
 
   
@@ -423 +436 @@
 
 bool documentaction::do_action (cgiargsclass &args, recptproto *collectproto,
-                  displayclass &disp, outconvertclass &outconvert, 
-                  ostream &textout, ostream &logout) {
-
-  text_tarray metadata;
-  FilterResponse_t response;
-  DocumentRequest_t docrequest;
-  DocumentResponse_t docresponse;
-  text_t topparent, classifytype, classifytitle;
-  comerror_t err;
-
-  text_t &arg_d = args["d"];
-  text_t &arg_cl = args["cl"];
-  text_t &collection = args["c"];
-
-  text_t OID = arg_d;
-  if (arg_d.empty()) OID = arg_cl;
-
-  textout << outconvert << disp << "_document:header_\n"
-      << "_document:content_\n";
-  
-  if (arg_d.empty() && arg_cl.empty()) {
-    textout << outconvert << disp << "Document contains no data_document:footer_\n";
-    return true;
-  }
-
-  // get the classifytitle and classifytype
-  get_top (OID, topparent);
-  metadata.push_back ("Title");
-  metadata.push_back ("classifytype");
-  if (get_info (topparent, collection, metadata, collectproto, response, logout)) {
-    classifytitle = response.docInfo[0].metadata[0].values[0];
-    classifytype = response.docInfo[0].metadata[1].values[0];
-  }
-
-  if (!arg_d.empty()) {
-    
-    if (classifytype == "Hierarchy")
-      output_hierarchy_toc (classifytitle, args, collectproto, disp, 
-                outconvert, textout, logout);
-    else
-      // Book is the default for a document level toc - AZList and Datelist don't
-      // make sense for document level
-      output_book_toc (classifytitle, args, collectproto, disp,
-               outconvert, textout, logout);
-  } else {
-    
-    if (classifytype == "Hierarchy")
-      output_hierarchy_toc (classifytitle, args, collectproto, disp, 
-                outconvert, textout, logout);
-    else if (classifytype == "Datelist")
-      output_datelist_toc (classifytitle, args, collectproto, disp, 
+                displayclass &disp, outconvertclass &outconvert, 
+                ostream &textout, ostream &logout) {
+
+  if (collectproto == NULL) {
+    logout << "documentaction::do_action called with NULL collectproto\n";
+    textout << outconvert << disp << "_document:header_\n" 
+        << "Error: Attempt to get document without setting collection\n"
+        << "_document:footer_\n";
+  } else {  
+
+    text_tarray metadata;
+    FilterResponse_t response;
+    FilterResponse_t queryresponse;
+    DocumentRequest_t docrequest;
+    DocumentResponse_t docresponse;
+    text_t topparent, classifytype, classifytitle;
+    comerror_t err;
+    bool highlight = false;
+    
+    text_t &arg_d = args["d"];
+    text_t &arg_cl = args["cl"];
+    text_t &collection = args["c"];
+    
+    text_t OID = arg_d;
+    if (arg_d.empty()) OID = arg_cl;
+    
+    textout << outconvert << disp << "_document:header_\n"
+        << "_document:content_\n";
+    
+    if (arg_d.empty() && arg_cl.empty()) {
+      textout << outconvert << disp << "Document contains no data_document:footer_\n";
+      return true;
+    }
+    
+    // if we have a query string and highlighting is turned on we need
+    // to redo the query to get the terms for highlighting
+    if (!args["q"].empty() && args.getintarg("hl")) {
+      FilterRequest_t request;
+      text_t quotedstring; // don't use this here
+      request.filterResultOptions = FRmatchTerms;
+      if (do_query (request, args, collectproto, quotedstring, queryresponse, logout))
+    highlight = true;
+    }
+    
+    // get the classifytitle and classifytype
+    get_top (OID, topparent);
+    metadata.push_back ("Title");
+    metadata.push_back ("classifytype");
+    if (get_info (topparent, collection, metadata, collectproto, response, logout)) {
+      classifytitle = response.docInfo[0].metadata[0].values[0];
+      classifytype = response.docInfo[0].metadata[1].values[0];
+    }
+    
+    if (!arg_d.empty()) {
+      
+      if (classifytype == "Hierarchy")
+    output_hierarchy_toc (classifytitle, args, collectproto, disp, 
+                  outconvert, textout, logout);
+      else
+    // Book is the default for a document level toc - AZList and Datelist don't
+    // make sense for document level
+    output_book_toc (classifytitle, args, collectproto, disp,
+             outconvert, textout, logout);
+    } else {
+    
+      if (classifytype == "Hierarchy")
+    output_hierarchy_toc (classifytitle, args, collectproto, disp, 
+                  outconvert, textout, logout);
+      else if (classifytype == "Datelist")
+    output_datelist_toc (classifytitle, args, collectproto, disp, 
+                 outconvert, textout, logout);
+      else 
+    // AZList is the default for a classification level toc - Book doesn't make 
+    // sense for classification level
+    output_azlist_toc (classifytitle, args, collectproto, disp, 
                outconvert, textout, logout);
-    else 
-      // AZList is the default for a classification level toc - Book doesn't make 
-      // sense for classification level
-      output_azlist_toc (classifytitle, args, collectproto, disp, 
-             outconvert, textout, logout);
-  }
-
-  // get info on this document
-  metadata.erase (metadata.begin(), metadata.end());
-  metadata.push_back("Title");
-  metadata.push_back("hastxt");
-  metadata.push_back("haschildren");
-
-  if (get_info (OID, collection, metadata, collectproto, response, logout)) {
-    text_t &title = response.docInfo[0].metadata[0].values[0];
-    int hastxt = response.docInfo[0].metadata[1].values[0].getint();
-    int haschildren = response.docInfo[0].metadata[2].values[0].getint();
-
-    int gt = args.getintarg("gt");
-    if (gt == 0) {
-      if (hastxt == 1) {
-    // get the text
-    docrequest.OID = OID;
-    collectproto->get_document (collection, docrequest, docresponse, err, logout);
-    
-    textout << outconvert << disp 
-        << "<p>\n<h3>" << title << "</h3>" << docresponse.doc;
-      }
+    }
+    
+    // get info on this document
+    metadata.erase (metadata.begin(), metadata.end());
+    metadata.push_back("Title");
+    metadata.push_back("hastxt");
+    metadata.push_back("haschildren");
+    
+    if (get_info (OID, collection, metadata, collectproto, response, logout)) {
+      text_t &title = response.docInfo[0].metadata[0].values[0];
+      int hastxt = response.docInfo[0].metadata[1].values[0].getint();
+      int haschildren = response.docInfo[0].metadata[2].values[0].getint();
+      
+      int gt = args.getintarg("gt");
+      if (gt == 0) {
+    if (hastxt == 1) {
+      // get the text
+      docrequest.OID = OID;
+      collectproto->get_document (collection, docrequest, docresponse, err, logout);
+      
+      textout << outconvert << disp << "<p>\n<h3>" << title << "</h3>";
+      if (highlight)
+        highlighttext(docresponse.doc, queryresponse.termInfo, disp,
+              outconvert, textout, logout);
+      else
+        textout << outconvert << disp << docresponse.doc;
+    }
+      } else {
+    
+    // text is to be expanded
+    int tmp; // this doesn't get used here
+    if (!haschildren)
+      OID = get_parent (OID);
+    
+    get_contents (OID, classifytype, tmp, collection, collectproto, response, logout);
+    
+    ResultDocInfo_tarray::const_iterator sechere = response.docInfo.begin();
+    ResultDocInfo_tarray::const_iterator secend = response.docInfo.end();
+    
+    if (gt == 1) {
+      // check if there are more than 10 sections containing text to be expanded - 
+      // if there are output warning message - this isn't a great way to do this
+      // since the sections may be very large or very small - one day I'll fix it 
+      // -- Stefan.
+      int seccount = 0;
+      while (sechere != secend) {
+        if ((*sechere).metadata[3].values[0] == "1") seccount ++;
+        if (seccount > 10) break;
+        sechere ++;
+      }
+      if (seccount > 10) textout << outconvert << disp << "_document:textltwarning_";
+      else gt = 2;
+    }
+
+    if (gt == 2) {
+    // get the text for each section
+      sechere = response.docInfo.begin();
+      int count = 0;
+      while (sechere != secend) {
+        textout << outconvert << disp << "\n<p><a name=" << count << "><h3>" 
+            << (*sechere).metadata[0].values[0] << "</h3></a>\n";
+        if ((*sechere).metadata[3].values[0] == "1") {
+          docrequest.OID = (*sechere).OID;
+          collectproto->get_document (collection, docrequest, docresponse, err, logout);
+          if (highlight)
+        highlighttext(docresponse.doc, queryresponse.termInfo, disp, 
+                  outconvert, textout, logout);
+          else
+        textout << outconvert << disp << docresponse.doc;
+        }
+        count ++;
+        sechere ++;
+      }
+    }
+      }
+    }
+    textout << outconvert << disp << "_document:footer_\n";
+  }
+  return true;
+}
+
+
+// highlighttext highlights query terms in text string and outputs the resulting text string
+void documentaction::highlighttext(text_t &text, TermInfo_tarray &terms, displayclass &disp,
+                   outconvertclass &outconvert, ostream &textout, ostream &/*logout*/) {
+
+  text_tmap allterms;
+  text_tmap::const_iterator it;
+
+  // first load all the term variations into a map
+  TermInfo_tarray::const_iterator this_term = terms.begin();
+  TermInfo_tarray::const_iterator last_term = terms.end();
+  while (this_term != last_term) {
+    text_tarray::const_iterator this_var = (*this_term).matchTerms.begin();
+    text_tarray::const_iterator last_var = (*this_term).matchTerms.end();
+    while (this_var != last_var) {
+      allterms[*this_var] = 1;
+      this_var ++;
+    }
+    this_term ++;
+  }
+
+  // get the text to start and end a hightlight
+  text_t starthighlight = "<b><u>";
+  text_t endhighlight = "</u></b>";
+  if (disp.isdefaultmacro("Global", "starthighlight"))
+    disp.expandstring("Global", "_starthighlight_", starthighlight);
+  if (disp.isdefaultmacro("Global", "endhighlight"))
+    disp.expandstring("Global", "_endhighlight_", endhighlight);
+
+
+  text_t::iterator here = text.begin(); 
+  text_t::iterator end = text.end();
+  text_t word, buffer;
+  while (here != end) {
+    if (is_unicode_letdig(*here)) {
+      // not word boundary
+      word.push_back(*here);
+      here++;
+
     } else {
-
-      // text is to be expanded
-      int tmp; // this doesn't get used here
-      if (!haschildren)
-    OID = get_parent (OID);
-
-      get_contents (OID, classifytype, tmp, collection, collectproto, response, logout);
-          
-      ResultDocInfo_tarray::const_iterator sechere = response.docInfo.begin();
-      ResultDocInfo_tarray::const_iterator secend = response.docInfo.end();
-          
-      if (gt == 1) {
-    // check if there are more than 10 sections containing text to be expanded - 
-    // if there are output warning message - this isn't a great way to do this
-    // since the sections may be very large or very small - one day I'll fix it 
-    // -- Stefan.
-    int seccount = 0;
-    while (sechere != secend) {
-      if ((*sechere).metadata[3].values[0] == "1") seccount ++;
-      if (seccount > 10) break;
-      sechere ++;
-    }
-    if (seccount > 10) textout << outconvert << disp << "_document:textltwarning_";
-    else gt = 2;
-      }
-
-      if (gt == 2) {
-    // get the text for each section
-    sechere = response.docInfo.begin();
-    int count = 0;
-    while (sechere != secend) {
-      textout << outconvert << disp << "\n<p><a name=" << count << "><h3>" 
-          << (*sechere).metadata[0].values[0] << "</h3></a>\n";
-      if ((*sechere).metadata[3].values[0] == "1") {
-        docrequest.OID = (*sechere).OID;
-        collectproto->get_document (collection, docrequest, docresponse, err, logout);
-        textout << outconvert << disp << docresponse.doc;
-      }
-      count ++;
-      sechere ++;
-    }
-      }
-    }
-  }
-  textout << outconvert << disp << "_document:footer_\n";
-  return true;
-}
-
+      // found word boundary
+      // add last word if there was one
+      if (!word.empty()) {
+    it = allterms.find(word);
+    if (it != allterms.end()) {
+      word = starthighlight + word + endhighlight;
+    }
+    buffer += word;
+        word.clear();
+      }
+
+      if (*here == '<') {
+        // skip over rest of html tag
+    while ((here != end) && (*here != '>')) {
+      buffer.push_back(*here);
+      here++;
+    }
+      }
+
+      buffer.push_back(*here);
+      here++;
+
+      if (buffer.size() > 1024) {
+    textout << outconvert << disp << buffer;
+    buffer.clear();
+      }
+    }
+  }
+  textout << outconvert << disp << buffer;
+}