documentparser.cpp

OSB-PIK-OpenVXI-3.0.0源代码 “中国XML论坛 - 专业的XML技术讨论区--XML在语音技术中的应用”

 /****************License************************************************
  *
  * Copyright 2000-2003.  ScanSoft, Inc.    
  *
  * Use of this software is subject to notices and obligations set forth 
  * in the SpeechWorks Public License - Software Version 1.2 which is 
  * included with this software. 
  *
  * ScanSoft is a registered trademark of ScanSoft, Inc., and OpenSpeech, 
  * SpeechWorks and the SpeechWorks logo are registered trademarks or 
  * trademarks of SpeechWorks International, Inc. in the United States 
  * and other countries.
  *
  ***********************************************************************/
 

 #include "DocumentParser.hpp"
 #include "DocumentStorage.hpp"
 #include "CommonExceptions.hpp"
 #include "SimpleLogger.hpp"
 #include "VXML.h"                    // for attribute names
 #include "DocumentConverter.hpp"     // for DocumentConverter
 #include "PropertyList.hpp"
 #include "XMLChConverter.hpp"
 #include "ValueLogging.hpp"     // for VXIValue dumping
 #include "VXIinet.h"
 
 // Internal documents
 #include "Schema.hpp"
 
 // Xerces related
 #ifndef HAVE_XERCES
 #error Need Apache Xerces to build the VoiceXML interpreter
 #endif
 
 #include <util/PlatformUtils.hpp>
 #include <util/TransService.hpp>
 #include <sax2/XMLReaderFactory.hpp>
 #include <framework/MemBufInputSource.hpp>
 #include <sax/SAXParseException.hpp> // by DOMTreeErrorReporter
 #include <sax/EntityResolver.hpp>    // by DTDResolver
 #include <sax/ErrorHandler.hpp>      // by DOMTreeErrorReporter
 #include <validators/common/Grammar.hpp>
 using namespace xercesc;
 
 //#############################################################################
 // Utilities - these are specific to Xerces
 //#############################################################################
 
 class DOMTreeErrorReporter : public ErrorHandler {
 public:
   DOMTreeErrorReporter()  { }
   ~DOMTreeErrorReporter() { }
   
   void warning(const SAXParseException& toCatch)     { /* Ignore */ }
   void fatalError(const SAXParseException& toCatch)  { error(toCatch); }
   void resetErrors() { }
 
   void error(const SAXParseException & toCatch)
   { throw SAXParseException(toCatch); }
 
 private:
   DOMTreeErrorReporter(const DOMTreeErrorReporter &);
   void operator=(const DOMTreeErrorReporter &);
 };
 
 
 class DTDResolver : public EntityResolver {
 public:
   virtual ~DTDResolver() { }
   DTDResolver() { }
 
   virtual InputSource * resolveEntity(const XMLCh * const publicId,
                                       const XMLCh * const systemId)
   {
 
     if (Compare(publicId, L"SB_Defaults")) {
       VXIcharToXMLCh name(L"VXML Defaults DTD (for SB 1.0)");
       return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_DEFAULTS_DTD,
                                    VALIDATOR_DEFAULTS_DTD_SIZE,
                                    name.c_str(), false);
     }
 
     if (Compare(systemId, L"http://www.w3.org/TR/voicexml20/vxml.xsd")) {
       VXIcharToXMLCh name(L"http://www.w3.org/TR/voicexml20/vxml.xsd (SB)");
       return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_VXML,
                                    VALIDATOR_VXML_SIZE,
                                    name.c_str(), false);
     }
 
     if (Compare(systemId, L"vxml-datatypes.xsd")) {
       VXIcharToXMLCh name(L"vxml-datatypes.xsd (SB)");
       return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_VXML_DATA,
                                    VALIDATOR_VXML_DATA_SIZE,
                                    name.c_str(), false);
     }
 
     if (Compare(systemId, L"vxml-attribs.xsd")) {
       VXIcharToXMLCh name(L"vxml-attribs.xsd (SB)");
       return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_VXML_ATTR,
                                    VALIDATOR_VXML_ATTR_SIZE,
                                    name.c_str(), false);
     }
 
     if (Compare(systemId, L"vxml-grammar-extension.xsd")) {
       VXIcharToXMLCh name(L"vxml-grammar-extension.xsd (SB)");
       return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_SRGF_EXTN,
                                    VALIDATOR_SRGF_EXTN_SIZE,
                                    name.c_str(), false);
     }
 
     if (Compare(systemId, L"vxml-grammar-restriction.xsd")) {
       VXIcharToXMLCh name(L"vxml-grammar-restriction.xsd (SB)");
       return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_SRGF_RSTR,
                                    VALIDATOR_SRGF_RSTR_SIZE,
                                    name.c_str(), false);
     }
 
     if (Compare(systemId, L"vxml-synthesis-extension.xsd")) {
       VXIcharToXMLCh name(L"vxml-synthesis-extension.xsd (SB)");
       return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_SSML_EXTN,
                                    VALIDATOR_SSML_EXTN_SIZE,
                                    name.c_str(), false);
     }
 
     if (Compare(systemId, L"vxml-synthesis-restriction.xsd")) {
       VXIcharToXMLCh name(L"vxml-synthesis-restriction.xsd (SB)");
       return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_SSML_RSTR,
                                    VALIDATOR_SSML_RSTR_SIZE,
                                    name.c_str(), false);
     }
 
     if (Compare(systemId, L"synthesis-core.xsd")) {
       VXIcharToXMLCh name(L"synthesis-core.xsd (SB)");
       return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_SSML_CORE,
                                    VALIDATOR_SSML_CORE_SIZE,
                                    name.c_str(), false);
     }
 
     if (Compare(systemId, L"grammar-core.xsd")) {
       VXIcharToXMLCh name(L"grammar-core.xsd (SB)");
       return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_SRGF_CORE,
                                    VALIDATOR_SRGF_CORE_SIZE,
                                    name.c_str(), false);
     }
 
     if (Compare(systemId, L"http://www.w3.org/2001/xml.xsd")) {
       VXIcharToXMLCh name(L"http://www.w3.org/2001/xml.xsd (SB)");
       return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_XML,
                                    VALIDATOR_XML_SIZE,
                                    name.c_str(), false);
     }
 
     if (Compare(systemId, L"XMLSchema.dtd")) {
       VXIcharToXMLCh name(L"XMLSchema.dtd (SB)");
       return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_SCHEMA_DTD,
                                    VALIDATOR_SCHEMA_DTD_SIZE,
                                    name.c_str(), false);
     }
     
     if (Compare(systemId, L"datatypes.dtd")) {
       VXIcharToXMLCh name(L"datatypes.dtd (SB)");
       return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_DATATYPE_DTD,
                                    VALIDATOR_DATATYPE_DTD_SIZE,
                                    name.c_str(), false);
     }
 
     if (Compare(publicId, L"-//W3C//DTD VOICEXML 2.0//EN") ||
         Compare(systemId, L"http://www.w3.org/TR/voicexml20/vxml.dtd"))
     {
       VXIcharToXMLCh name(L"VXML DTD (SB)");
       return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_VXML_DTD,
                                    VALIDATOR_VXML_DTD_SIZE,
                                    name.c_str(), false);
     }
 
     /*
     VXIcharToXMLCh name(L"VXML DTD (for SB 1.0)");
     return new MemBufInputSource(VALIDATOR_DATA + VALIDATOR_VXML_DTD,
                                  VALIDATOR_VXML_DTD_SIZE,
                                  name.c_str(), false);
     */
 
     return NULL;
   }
 };
 
 //#############################################################################
 // Document Parser
 //#############################################################################
 
 // xerces crashes when on multi-thread app. that simultaneously load schema
 // grammar therefore use a global mutext to restrict access to only thread
 // at a time
 static VXItrdMutex* gblXMLGrammarMutex = NULL;
 
 bool DocumentParser::Initialize(unsigned int cacheSize)
 {
   try {
     XMLPlatformUtils::Initialize();
     if (!VXMLDocumentModel::Initialize()) return false;
     DocumentConverter::Initialize();
     VXItrdMutexCreate(&gblXMLGrammarMutex);
   }
 
   catch (const XMLException &) {
     return false;
   }
 
   DocumentStorageSingleton::Initialize(cacheSize);
 
   return true;
 }
 
 
 void DocumentParser::Deinitialize()
 {
   DocumentStorageSingleton::Deinitialize();
 
   try {
     VXItrdMutexDestroy(&gblXMLGrammarMutex);
     DocumentConverter::Deinitialize();
     VXMLDocumentModel::Deinitialize();
     XMLPlatformUtils::Terminate();
   }
   catch (const XMLException &) {
     // do nothing
   }
 }
 
 static void LockLoadGrammar(void)
 {
   if( gblXMLGrammarMutex ) VXItrdMutexLock(gblXMLGrammarMutex);
 }
 
 static void UnlockLoadGrammar(void)
 {
   if( gblXMLGrammarMutex ) VXItrdMutexUnlock(gblXMLGrammarMutex);
 }  
 
 DocumentParser::DocumentParser()
   : parser(NULL), converter(NULL), loadedVXML20(false)
 {
   converter = new DocumentConverter();
   if (converter == NULL) throw VXIException::OutOfMemory();
 
   parser = XMLReaderFactory::createXMLReader();
   if (parser == NULL) {
     delete converter;
     throw VXIException::OutOfMemory();
   }
 
   DTDResolver * dtd = new DTDResolver();
   if (dtd == NULL) {
     delete converter;
     delete parser;
     throw VXIException::OutOfMemory();
   }
   parser->setEntityResolver(dtd);
 
   // These settings below should not change the Xerces defaults.  Their
   // presence makes the defaults explicit.
 
   parser->setFeature(XMLUni::fgSAX2CoreNameSpaces, true);
   parser->setFeature(XMLUni::fgSAX2CoreValidation, true);
   parser->setFeature(XMLUni::fgXercesDynamic, false);
   parser->setFeature(XMLUni::fgXercesSchema, true);
   parser->setFeature(XMLUni::fgXercesSchemaFullChecking, true);
   parser->setFeature(XMLUni::fgSAX2CoreNameSpacePrefixes, false);
 
   ErrorHandler *errReporter = new DOMTreeErrorReporter();
   parser->setErrorHandler(errReporter);
 
   parser->setContentHandler(converter);
 }
 
 
 DocumentParser::~DocumentParser()
 {
   if (parser != NULL) {
     const ErrorHandler * reporter = parser->getErrorHandler();
     delete reporter;
 
     const EntityResolver * resolver = parser->getEntityResolver();
     delete resolver;
 
     delete parser;
     delete converter;
     parser = NULL;
   }
 }
 
 
 //****************************************************************************
 // FetchBuffer
 //****************************************************************************
 
 // 1: Invalid parameter
 // 2: Unable to open URL
 // 3: Unable to read from URL
 int DocumentParser::FetchBuffer(const VXIchar * url,
                                 const VXIMapHolder & properties,
                                 VXIMapHolder & streamInfo,
                                 VXIinetInterface * inet,
                                 SimpleLogger & log,
                                 const VXIbyte * & result,
                                 VXIulong & read,
                                 vxistring & docURL)
 {
   if (log.IsLogging(2)) {
     log.StartDiagnostic(2) << L"DocumentParser::FetchBuffer(" << url
                            << L", " << properties.GetValue() << L")";
     log.EndDiagnostic();
   }
 
   if (inet == NULL || url == NULL || wcslen(url) == 0) return 1;
   
   // (1) Open URL
   VXIinetStream * stream;
 
   // VXIMapHolder streamInfo;
   if (streamInfo.GetValue() == NULL) 
   {
     return -1;
   }
   
   if (inet->Open(inet, L"vxi", url, INET_MODE_READ, 0, properties.GetValue(),
                  streamInfo.GetValue(), &stream) != 0)
   {
     if (log.IsLogging(0)) {
       log.StartDiagnostic(0) << L"DocumentParser::FetchBuffer - could not "
         L"open URL: " << url;
       log.EndDiagnostic();
     }
     return 2;
   }
   
   // (2) Determine document size & read into local memory buffer.
   const VXIValue * tempURL = NULL;
   tempURL = VXIMapGetProperty(streamInfo.GetValue(), INET_INFO_ABSOLUTE_NAME);
   if (tempURL == NULL || VXIValueGetType(tempURL) != VALUE_STRING) {
     inet->Close(inet, &stream);
     if (log.IsLogging(0)) {
       log.StartDiagnostic(0) << L"DocumentParser::FetchBuffer - could not "
         L"retrieve absolute path of document at URL: " << url;
       log.EndDiagnostic();
     }
     return 2;
   }
   docURL = VXIStringCStr(reinterpret_cast<const VXIString *>(tempURL));
 
   const VXIValue * tempSize = NULL;
   tempSize = VXIMapGetProperty(streamInfo.GetValue(), INET_INFO_SIZE_BYTES);
   if (tempSize == NULL || VXIValueGetType(tempSize) != VALUE_INTEGER) {
     inet->Close(inet, &stream);
     if (log.IsLogging(0)) {
       log.StartDiagnostic(0) << L"DocumentParser::FetchBuffer - could not "
         L"retrieve size of document at URL: " << url;
       log.EndDiagnostic();
     }
     return 2;
   }
 
   VXIint32 bufSize
     = VXIIntegerValue(reinterpret_cast<const VXIInteger *>(tempSize));