00001
00002
00003
00004
00005
00006
00007
00008
00009
00010
00011
00012
00013
00014
00015
00016
00017
00018
00019 #ifndef INDRI_POWERPOINTDOCUMENTEXTRACTOR_HPP
00020 #define INDRI_POWERPOINTDOCUMENTEXTRACTOR_HPP
00021 #ifdef WIN32
00022 #include <string>
00023 #include "indri/Buffer.hpp"
00024 #include "indri/UnparsedDocument.hpp"
00025 #undef WIN32_LEAN_AND_MEAN
00026 #undef NOGDI
00027 #include <windows.h>
00028 #include <unknwn.h>
00029 #include <oaidl.h>
00030 #include "indri/DocumentIterator.hpp"
00031
00032 class PowerPointDocumentExtractor : public DocumentIterator {
00033 private:
00034 IUnknown* _powerPointUnknown;
00035 IDispatch* _powerPointDispatch;
00036 IDispatch* _presentationsDispatch;
00037 UnparsedDocument _unparsedDocument;
00038
00039 std::string _documentPath;
00040 Buffer _documentBuffer;
00041 bool _documentWaiting;
00042
00043 public:
00044 PowerPointDocumentExtractor();
00045 ~PowerPointDocumentExtractor();
00046
00047 void open( const std::string& filename );
00048 UnparsedDocument* nextDocument();
00049 void close();
00050 };
00051 #endif // WIN32
00052 #endif // INDRI_POWERPOINTDOCUMENTEXTRACTOR_HPP
00053