XMLDocShredder.cpp

   1 /**********************************************************
   2  * XMLDocShredder.cpp
   3  * ---------------------
   4  * Implementation of the class that receives events from the XML parser and
   5  * invokes corresponding construction methods of the storage interface.
   6  *
   7  * Author: Greg Leighton
   8  * Date: 02/11/08
   9  * Changes:
  10  *              05/11/08 -- Fixed bug related to parsing empty elements
  11  *                               -- Set parser properties to automatically resolve
  12  *                                      entity references and load external DTD if present
  13  *                               -- Modified processEndDocument() by adding a nodeFinished()
  14  *                                      call to the storage interface to close off the
  15  *                                      document node
  16  *
  17  */
  18
  19 #include <iostream>
  20 #include "XMLDocShredder.h"
  21 #include "SXSIStorageInterface.h"
  22 #include <libxml++/exceptions/parse_error.h>
  23 #include "Utils.h"
  24
  25 using namespace Glib;
  26
  27 void XMLDocShredder::setProperties(){
  28   /* instruct the parser to expand entity references and report as
  29    * regular PCDATA
  30    */
  31   reader_->set_parser_property(
  32                                TextReader::SubstEntities, true);
  33
  34   /* instruct parser to read external DTD, if present.  This is
  35          * needed to obtain any entity definitions in the DTD
  36          */
  37   reader_->set_parser_property(
  38                                TextReader::LoadDtd, true);
  39
  40
  41   /*
  42    */
  43   reader_->set_parser_property(
  44                                TextReader::DefaultAttrs, true);
  45
  46
  47   /* but we don't want to do validation since it would slow us down
  48    */
  49
  50
  51   reader_->set_parser_property(
  52                                TextReader::Validate, false);
  53
  54 }
  55 XMLDocShredder::XMLDocShredder(const unsigned char * data,
  56                                TextReader::size_type size)
  57 {
  58   last_text = false;
  59   reader_ = new TextReader(data,size,"");
  60   setProperties();
  61   storageIfc_ = new SXSIStorageInterface();
  62   buffer = "";
  63 }
  64
  65 XMLDocShredder::XMLDocShredder(const string inFileName)
  66 {
  67   last_text = false;
  68   reader_ = new TextReader(inFileName);
  69   setProperties();
  70   storageIfc_ = new SXSIStorageInterface();
  71   buffer = "";
  72 }
  73
  74 XMLDocShredder::~XMLDocShredder()
  75 {
  76         delete reader_;
  77         delete storageIfc_;
  78
  79 }
  80
  81
  82 void XMLDocShredder::processStartElement()
  83 {
  84         // fetch element name; this will be the full qualified name
  85         ustring name = reader_->get_name();
  86         bool empty = false;
  87
  88         storageIfc_->newText(buffer); //prevText
  89         buffer.erase();
  90
  91         storageIfc_->newChild(name);
  92
  93         /* We must be really carefull here. calling process attributes moves
  94            the document pointer on the last attribute, hence calling reader_->is_empty
  95            afterwards will yield the wrong result. It is better to call it while we are
  96            on the element and generate a nodeFinished() call at the end */
  97         empty = reader_->is_empty_element();
  98
  99
 100         // now, process attributes
 101         if (reader_->has_attributes())
 102           {
 103             processAttributes();
 104           };
 105
 106
 107         if (empty){
 108             storageIfc_->newText("");  //myText
 109             storageIfc_->nodeFinished(name);
 110         };
 111
 112
 113 }
 114
 115 void XMLDocShredder::processEndElement()
 116 {
 117   // tell the storage interface that the current node has been completely processed
 118   storageIfc_->newText(buffer); //prevText
 119   buffer.erase();
 120   storageIfc_->nodeFinished(reader_->get_name());
 121 }
 122
 123 void XMLDocShredder::processPCDATA()
 124 {
 125         // send the content of this PCDATA node to the storage interface as a text node
 126
 127         if (reader_->has_value())
 128         {
 129           buffer += reader_->get_value();
 130
 131         };
 132
 133 }
 134
 135 void XMLDocShredder::processAttributes()
 136 {
 137         reader_->move_to_first_attribute();
 138
 139         string nspaceStr = "xmlns";
 140         storageIfc_->newText(""); //prevText
 141         storageIfc_->newChild("<@>");
 142         do
 143         {
 144                 ustring name = reader_->get_name();
 145                 ustring value = reader_->get_value();
 146
 147                 /* the libxml++ TextReader treats the xmlns attribute like an ordinary attribute,
 148                 * so we have to extract it and build a namespace uri node out of it before
 149                 * passing to the storage interface */
 150
 151                 if ((name.find(nspaceStr.c_str(), 0, 5)) == 0)
 152                 {
 153                         storageIfc_->newChild(":" + value);
 154                         storageIfc_->nodeFinished(":" + value);
 155                 }
 156
 157                 /* otherwise, this is an ordinary attribute, so we construct a new child node of the
 158                  * parent element to store the attribute name, possessing a child text node storing the
 159                  * attribute value.  Then, we close off the attribute node with a call to nodeFinished()
 160                  */
 161
 162                 else
 163                 {
 164                   storageIfc_->newText(""); //prevText
 165                   storageIfc_->newChild(name);
 166                   storageIfc_->newText(value);
 167                   storageIfc_->nodeFinished(name);
 168                 }
 169         }
 170         while (reader_->move_to_next_attribute());
 171         storageIfc_->newText(""); //nextText
 172         storageIfc_->nodeFinished("<@>");
 173 }
 174
 175 void XMLDocShredder::processSignificantWhitespace()
 176 {
 177   // each significant whitespace sequence constructs a text node
 178   buffer += reader_->get_value();
 179
 180 }
 181
 182 void XMLDocShredder::processStartDocument(const string docName)
 183 {
 184   // tell storage interface to construct the document name
 185   storageIfc_->newChild("");
 186
 187 }
 188
 189 void XMLDocShredder::processEndDocument()
 190 {
 191         /* tell the storage interface that document parsing has finished, and structures
 192          * can now be written to disk. */
 193   storageIfc_->newText("");
 194   storageIfc_->nodeFinished("");
 195   storageIfc_->parsingFinished();
 196 }
 197
 198 void XMLDocShredder::processComment()
 199 {
 200   //storageIfc_->newChild("!" + reader_->get_value());
 201   //storageIfc_->nodeFinished();
 202 }
 203
 204 void XMLDocShredder::processProcessingInstruction()
 205 {
 206         ustring name = reader_->get_name();
 207         ustring value = reader_->get_value();
 208
 209         /* Create a child node to store the target of the PI, append a text node to it to store
 210          * the PI data, send to the storage interface.  Close off the PI node with a call to
 211          * nodeFinished
 212          */
 213
 214         // storageIfc_->newChild("?" + name);
 215         // storageIfc_->newText(value);
 216         // storageIfc_->nodeFinished();
 217 }
 218
 219 void XMLDocShredder::processDocTypeDeclaration()
 220 {
 221         /* We currently ignore the DOCTYPE declaration, but we'll provide this method skeleton
 222          * in case we do want to process it in the future.
 223         */
 224 }
 225
 226 void XMLDocShredder::processCDATASection()
 227 {
 228         /* Currently, we don't preserve CDATA sections since they aren't part of the XPath data
 229          * model.  Instead, we simply pass the converted text value to the storage interface as
 230          * a text node attached to the current context node.
 231          */
 232
 233         ustring value = reader_->get_value();
 234         storageIfc_->newText(value);
 235         last_text = true;
 236
 237 }
 238
 239 void XMLDocShredder::processUnknownNodeType()
 240 {
 241         cout << "unknown token encountered during parsing" << endl;
 242         throw xmlpp::parse_error("unknown token encountered during parsing");
 243
 244 }
 245
 246 void XMLDocShredder::parse()
 247 {
 248         while (reader_->read() && (reader_->get_read_state() != TextReader::Error))
 249         {
 250                 switch (reader_->get_node_type())
 251                 {
 252                         case TextReader::Element:
 253                                 processStartElement();
 254                                 break;
 255
 256                         case TextReader::Text:
 257                                 processPCDATA();
 258                                 break;
 259
 260                         case TextReader::EndElement:
 261                                 processEndElement();
 262                                 break;
 263
 264                         case TextReader::SignificantWhitespace:
 265                                 processSignificantWhitespace();
 266                                 break;
 267
 268                         case TextReader::Comment:
 269                                 processComment();
 270                                 break;
 271
 272                         case TextReader::DocumentType:
 273                                 processDocTypeDeclaration();
 274                                 break;
 275
 276                         case TextReader::ProcessingInstruction:
 277                                 processProcessingInstruction();
 278                                 break;
 279
 280                         case TextReader::CDATA:
 281                                 processCDATASection();
 282                                 break;
 283
 284                         case TextReader::None:
 285                                 processUnknownNodeType();
 286                                 break;
 287
 288                         default:
 289                                 int type = reader_->get_node_type();
 290                                 cout << "  Node type: " << type << endl;
 291                                 break;
 292
 293                 }
 294         }
 295 }