X-Git-Url: http://git.nguyen.vg/gitweb/?a=blobdiff_plain;f=XMLDocShredder.cpp;h=cb70e19590eef87217c0daaa82c62f66061ca7c7;hb=df5fdb22632be887ecd9f5c46a014e7e970148a2;hp=1516c9fc1cd724c346c66877b250af4d94d843f3;hpb=451e60ad59e35344dff62da5ca27fcd5eec1bff9;p=SXSI%2Fxpathcomp.git

diff --git a/XMLDocShredder.cpp b/XMLDocShredder.cpp
index 1516c9f..cb70e19 100644
--- a/XMLDocShredder.cpp
+++ b/XMLDocShredder.cpp
@@ -18,12 +18,22 @@
 
 #include <iostream>
 #include "XMLDocShredder.h"
-#include "SXSIStorageInterface.h"
 #include <libxml++/exceptions/parse_error.h>
 #include "Utils.h"
 
 using namespace Glib;
 
+void XMLDocShredder::doText(){
+
+  if (!buffer.empty()){
+    tb->NewOpenTag(PCDATA_OPEN_TAG);
+    tb->NewText(buffer);
+    tb->NewClosingTag(PCDATA_OPEN_TAG);
+  };
+  buffer.clear();
+
+}
+
 void XMLDocShredder::setProperties(){
   /* instruct the parser to expand entity references and report as 
    * regular PCDATA
@@ -58,74 +68,77 @@ XMLDocShredder::XMLDocShredder(const unsigned char * data,
 			       bool iet, 
 			       bool dtc) 			
 {
-  last_text = false;
+  tree = NULL;
   reader_ = new TextReader(data,size,"");
   setProperties();
-  storageIfc_ = new SXSIStorageInterface(sf,iet,dtc);
-  buffer = "";
+  tb  = new XMLTreeBuilder();
+  buffer.clear();
+  tb->OpenDocument(iet,sf,dtc);
 }
 
 XMLDocShredder::XMLDocShredder(const string inFileName,int sf, bool iet, bool dtc)
 {
-  last_text = false;
+  tree = NULL;
   reader_ = new TextReader(inFileName);
   setProperties();
-  storageIfc_ = new SXSIStorageInterface(sf,iet,dtc);
-  buffer = "";
+  tb = new XMLTreeBuilder();
+  buffer.clear();
+  tb->OpenDocument(iet,sf,dtc);
 }
 
 XMLDocShredder::~XMLDocShredder()
 {
 	delete reader_;
-	delete storageIfc_;
+	reader_ = NULL;
+	delete tb;
+	tb = NULL;
 
 }
 
 
 void XMLDocShredder::processStartElement()
 {
-	// fetch element name; this will be the full qualified name
-	ustring name = reader_->get_name();
-	bool empty = false;
-	
-	storageIfc_->newChild(name);
-
-	/* We must be really carefull here. calling process attributes moves
-	   the document pointer on the last attribute, hence calling reader_->is_empty
-	   afterwards will yield the wrong result. It is better to call it while we are
-	   on the element and generate a nodeFinished() call at the end */
-	empty = reader_->is_empty_element();
-
-
-	// now, process attributes
-	if (reader_->has_attributes())
-	  {
-	    processAttributes();
-	  };
-
-	
-	if (empty){
-	  storageIfc_->nodeFinished(name);	   
-	};
-
-
+  doText();
+  // fetch element name; this will be the full qualified name
+  ustring name = reader_->get_name();
+  bool empty = false;
+  size_t found = name.find_first_of(':');
+  if (found == ustring::npos)  
+    tb->NewOpenTag(name);
+  else
+    tb->NewOpenTag(name.substr(found+1,name.length() - found - 1));
+  
+  /* We must be really carefull here. calling process attributes moves
+     the document pointer on the last attribute, hence calling reader_->is_empty
+     afterwards will yield the wrong result. It is better to call it while we are
+     on the element and generate a nodeFinished() call at the end */
+  empty = reader_->is_empty_element();
+  
+  
+  // now, process attributes
+  if (reader_->has_attributes())
+    processAttributes();
+  
+  
+  if (empty)
+    tb->NewClosingTag(name);
+  
+  
 }
 
 void XMLDocShredder::processEndElement()
 {
-  // tell the storage interface that the current node has been completely processed  
-  storageIfc_->nodeFinished(reader_->get_name());
+  doText();
+  ustring name = reader_->get_name();
+  tb->NewClosingTag(name);
 }
 
 void XMLDocShredder::processPCDATA()
 {
   // send the content of this PCDATA node to the storage interface as a text node
-  
-  if (reader_->has_value()){
-    storageIfc_->newChild("<$>");
-    storageIfc_->newText(reader_->get_value());
-    storageIfc_->nodeFinished("<$>");
-  };
+  if (reader_->has_value())
+    buffer += reader_->get_value();
+
 }
 
 void XMLDocShredder::processAttributes()
@@ -133,9 +146,9 @@ void XMLDocShredder::processAttributes()
 	reader_->move_to_first_attribute();
 		
 	string nspaceStr = "xmlns";
-	storageIfc_->newChild("<@>");
+	tb->NewOpenTag(ATTRIBUTE_OPEN_TAG);
 	do
-	{
+	  {
 		ustring name = reader_->get_name();
 		ustring value = reader_->get_value();
 		
@@ -145,8 +158,7 @@ void XMLDocShredder::processAttributes()
 		
 		if ((name.find(nspaceStr.c_str(), 0, 5)) == 0)
 		{
-			storageIfc_->newChild(":" + value);
-			storageIfc_->nodeFinished(":" + value);	
+		  //TODO 
 		}
 		
 		/* otherwise, this is an ordinary attribute, so we construct a new child node of the 
@@ -157,40 +169,40 @@ void XMLDocShredder::processAttributes()
 		else
 		{
 		  string attname = "<@>"+name;
-		  storageIfc_->newChild(attname);
-		  storageIfc_->newChild("<@$>");
-		  storageIfc_->newText(value);
-		  storageIfc_->nodeFinished("<@$>");
-		  storageIfc_->nodeFinished(attname);
+		  tb->NewOpenTag(attname);
+		  tb->NewOpenTag(ATTRIBUTE_DATA_OPEN_TAG);
+		  tb->NewText(value);
+		  tb->NewClosingTag(ATTRIBUTE_DATA_OPEN_TAG);
+		  tb->NewClosingTag(attname);
 		}
 	}
 	while (reader_->move_to_next_attribute());
-	storageIfc_->nodeFinished("<@>");
+	tb->NewClosingTag(ATTRIBUTE_OPEN_TAG);
 }
 
 void XMLDocShredder::processSignificantWhitespace()
 {
-  
-  if (reader_->has_value()){
-    storageIfc_->newChild("<$>");
-    storageIfc_->newText(reader_->get_value());
-    storageIfc_->nodeFinished("<$>");
-  };  
+  if (reader_->has_value())
+    buffer += reader_->get_value();
+
 }
 
 void XMLDocShredder::processStartDocument(const string docName)
 {
   // tell storage interface to construct the document name
-  storageIfc_->newChild("");  
+
+  tb->NewOpenTag(DOCUMENT_OPEN_TAG);
   
 }
 
 void XMLDocShredder::processEndDocument()
 {
-	/* tell the storage interface that document parsing has finished, and structures
-	 * can now be written to disk. */
-  storageIfc_->nodeFinished("");
-  storageIfc_->parsingFinished();	
+  doText();
+  /* tell the storage interface that document parsing has finished, and structures
+   * can now be written to disk. */
+  tb->NewClosingTag(DOCUMENT_OPEN_TAG);
+  tree = tb->CloseDocument();
+
 }
 
 void XMLDocShredder::processComment()
@@ -227,12 +239,8 @@ void XMLDocShredder::processCDATASection()
 	 * model.  Instead, we simply pass the converted text value to the storage interface as 
 	 * a text node attached to the current context node.
 	 */
-  if (reader_->has_value()){
-    storageIfc_->newChild("<$>");
-    storageIfc_->newText(reader_->get_value());
-    storageIfc_->nodeFinished("<$>");
-  };
-
+  if (reader_->has_value())
+    buffer+= reader_->get_value();
 }
 
 void XMLDocShredder::processUnknownNodeType()