sc/source/filter/html/htmlimp.cxx

   1 /*************************************************************************
   2  *
   3  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   4  *
   5  * Copyright 2000, 2010 Oracle and/or its affiliates.
   6  *
   7  * OpenOffice.org - a multi-platform office productivity suite
   8  *
   9  * This file is part of OpenOffice.org.
  10  *
  11  * OpenOffice.org is free software: you can redistribute it and/or modify
  12  * it under the terms of the GNU Lesser General Public License version 3
  13  * only, as published by the Free Software Foundation.
  14  *
  15  * OpenOffice.org is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18  * GNU Lesser General Public License version 3 for more details
  19  * (a copy is included in the LICENSE file that accompanied this code).
  20  *
  21  * You should have received a copy of the GNU Lesser General Public License
  22  * version 3 along with OpenOffice.org.  If not, see
  23  * <http://www.openoffice.org/license.html>
  24  * for a copy of the LGPLv3 License.
  25  *
  26  ************************************************************************/
  27
  28 // MARKER(update_precomp.py): autogen include statement, do not remove
  29 #include "precompiled_sc.hxx"
  30
  31
  32
  33 //------------------------------------------------------------------------
  34
  35 #include "scitems.hxx"
  36 #include <editeng/eeitem.hxx>
  37
  38 #include <editeng/lrspitem.hxx>
  39 #include <editeng/paperinf.hxx>
  40 #include <editeng/sizeitem.hxx>
  41 #include <editeng/ulspitem.hxx>
  42 #include <editeng/boxitem.hxx>
  43 #include <vcl/svapp.hxx>
  44
  45 #include "htmlimp.hxx"
  46 #include "htmlpars.hxx"
  47 #include "filter.hxx"
  48 #include "global.hxx"
  49 #include "document.hxx"
  50 #include "editutil.hxx"
  51 #include "stlpool.hxx"
  52 #include "stlsheet.hxx"
  53 #include "compiler.hxx"
  54 #include "rangenam.hxx"
  55 #include "attrib.hxx"
  56 #include "ftools.hxx"
  57 #include "tokenarray.hxx"
  58
  59
  60 //------------------------------------------------------------------------
  61
  62 FltError ScFormatFilterPluginImpl::ScImportHTML( SvStream &rStream, const String& rBaseURL, ScDocument *pDoc,
  63         ScRange& rRange, double nOutputFactor, BOOL bCalcWidthHeight, SvNumberFormatter* pFormatter,
  64         bool bConvertDate )
  65 {
  66     ScHTMLImport aImp( pDoc, rBaseURL, rRange, bCalcWidthHeight );
  67     FltError nErr = (FltError) aImp.Read( rStream, rBaseURL );
  68     ScRange aR = aImp.GetRange();
  69     rRange.aEnd = aR.aEnd;
  70     aImp.WriteToDocument( TRUE, nOutputFactor, pFormatter, bConvertDate );
  71     return nErr;
  72 }
  73
  74 ScEEAbsImport *ScFormatFilterPluginImpl::CreateHTMLImport( ScDocument* pDocP, const String& rBaseURL, const ScRange& rRange, BOOL bCalcWidthHeight )
  75 {
  76     return new ScHTMLImport( pDocP, rBaseURL, rRange, bCalcWidthHeight );
  77 }
  78
  79 ScHTMLImport::ScHTMLImport( ScDocument* pDocP, const String& rBaseURL, const ScRange& rRange, BOOL bCalcWidthHeight ) :
  80     ScEEImport( pDocP, rRange )
  81 {
  82     Size aPageSize;
  83     OutputDevice* pDefaultDev = Application::GetDefaultDevice();
  84     const String& aPageStyle = mpDoc->GetPageStyle( rRange.aStart.Tab() );
  85     ScStyleSheet* pStyleSheet = (ScStyleSheet*)mpDoc->
  86         GetStyleSheetPool()->Find( aPageStyle, SFX_STYLE_FAMILY_PAGE );
  87     if ( pStyleSheet )
  88     {
  89         const SfxItemSet& rSet = pStyleSheet->GetItemSet();
  90         const SvxLRSpaceItem* pLRItem = (const SvxLRSpaceItem*) &rSet.Get( ATTR_LRSPACE );
  91         long nLeftMargin   = pLRItem->GetLeft();
  92         long nRightMargin  = pLRItem->GetRight();
  93         const SvxULSpaceItem* pULItem = (const SvxULSpaceItem*) &rSet.Get( ATTR_ULSPACE );
  94         long nTopMargin    = pULItem->GetUpper();
  95         long nBottomMargin = pULItem->GetLower();
  96         aPageSize = ((const SvxSizeItem&) rSet.Get(ATTR_PAGE_SIZE)).GetSize();
  97         if ( !aPageSize.Width() || !aPageSize.Height() )
  98         {
  99             DBG_ERRORFILE("PageSize Null ?!?!?");
 100             aPageSize = SvxPaperInfo::GetPaperSize( PAPER_A4 );
 101         }
 102         aPageSize.Width() -= nLeftMargin + nRightMargin;
 103         aPageSize.Height() -= nTopMargin + nBottomMargin;
 104         aPageSize = pDefaultDev->LogicToPixel( aPageSize, MapMode( MAP_TWIP ) );
 105     }
 106     else
 107     {
 108         DBG_ERRORFILE("kein StyleSheet?!?");
 109         aPageSize = pDefaultDev->LogicToPixel(
 110             SvxPaperInfo::GetPaperSize( PAPER_A4 ), MapMode( MAP_TWIP ) );
 111     }
 112     if( bCalcWidthHeight )
 113         mpParser = new ScHTMLLayoutParser( mpEngine, rBaseURL, aPageSize, pDocP );
 114     else
 115         mpParser = new ScHTMLQueryParser( mpEngine, pDocP );
 116 }
 117
 118
 119 ScHTMLImport::~ScHTMLImport()
 120 {
 121     // Reihenfolge wichtig, sonst knallt's irgendwann irgendwo in irgendeinem Dtor!
 122     // Ist gewaehrleistet, da ScEEImport Basisklasse ist
 123     delete (ScHTMLParser*) mpParser;        // vor EditEngine!
 124 }
 125
 126
 127 void ScHTMLImport::InsertRangeName( ScDocument* pDoc, const String& rName, const ScRange& rRange )
 128 {
 129     ScComplexRefData aRefData;
 130     aRefData.InitRange( rRange );
 131     ScTokenArray aTokArray;
 132     aTokArray.AddDoubleReference( aRefData );
 133     ScRangeData* pRangeData = new ScRangeData( pDoc, rName, aTokArray );
 134     if( !pDoc->GetRangeName()->Insert( pRangeData ) )
 135         delete pRangeData;
 136 }
 137
 138 void ScHTMLImport::WriteToDocument(
 139     BOOL bSizeColsRows, double nOutputFactor, SvNumberFormatter* pFormatter, bool bConvertDate )
 140 {
 141     ScEEImport::WriteToDocument( bSizeColsRows, nOutputFactor, pFormatter, bConvertDate );
 142
 143     const ScHTMLParser* pParser = GetParser();
 144     const ScHTMLTable* pGlobTable = pParser->GetGlobalTable();
 145     if( !pGlobTable )
 146         return;
 147
 148     // set cell borders for HTML table cells
 149     pGlobTable->ApplyCellBorders( mpDoc, maRange.aStart );
 150
 151     // correct cell borders for merged cells
 152     for ( ScEEParseEntry* pEntry = pParser->First(); pEntry; pEntry = pParser->Next() )
 153     {
 154         if( (pEntry->nColOverlap > 1) || (pEntry->nRowOverlap > 1) )
 155         {
 156             SCTAB nTab = maRange.aStart.Tab();
 157             const ScMergeAttr* pItem = (ScMergeAttr*) mpDoc->GetAttr( pEntry->nCol, pEntry->nRow, nTab, ATTR_MERGE );
 158             if( pItem->IsMerged() )
 159             {
 160                 SCCOL nColMerge = pItem->GetColMerge();
 161                 SCROW nRowMerge = pItem->GetRowMerge();
 162
 163                 const SvxBoxItem* pToItem = (const SvxBoxItem*)
 164                     mpDoc->GetAttr( pEntry->nCol, pEntry->nRow, nTab, ATTR_BORDER );
 165                 SvxBoxItem aNewItem( *pToItem );
 166                 if( nColMerge > 1 )
 167                 {
 168                     const SvxBoxItem* pFromItem = (const SvxBoxItem*)
 169                         mpDoc->GetAttr( pEntry->nCol + nColMerge - 1, pEntry->nRow, nTab, ATTR_BORDER );
 170                     aNewItem.SetLine( pFromItem->GetLine( BOX_LINE_RIGHT ), BOX_LINE_RIGHT );
 171                 }
 172                 if( nRowMerge > 1 )
 173                 {
 174                     const SvxBoxItem* pFromItem = (const SvxBoxItem*)
 175                         mpDoc->GetAttr( pEntry->nCol, pEntry->nRow + nRowMerge - 1, nTab, ATTR_BORDER );
 176                     aNewItem.SetLine( pFromItem->GetLine( BOX_LINE_BOTTOM ), BOX_LINE_BOTTOM );
 177                 }
 178                 mpDoc->ApplyAttr( pEntry->nCol, pEntry->nRow, nTab, aNewItem );
 179             }
 180         }
 181     }
 182
 183     // create ranges for HTML tables
 184      // 1 - entire document
 185     ScRange aNewRange( maRange.aStart );
 186     aNewRange.aEnd.IncCol( static_cast<SCsCOL>(pGlobTable->GetDocSize( tdCol )) - 1 );
 187     aNewRange.aEnd.IncRow( pGlobTable->GetDocSize( tdRow ) - 1 );
 188     InsertRangeName( mpDoc, ScfTools::GetHTMLDocName(), aNewRange );
 189
 190     // 2 - all tables
 191     InsertRangeName( mpDoc, ScfTools::GetHTMLTablesName(), ScRange( maRange.aStart ) );
 192
 193     // 3 - single tables
 194     SCsCOL nColDiff = (SCsCOL)maRange.aStart.Col();
 195     SCsROW nRowDiff = (SCsROW)maRange.aStart.Row();
 196     SCsTAB nTabDiff = (SCsTAB)maRange.aStart.Tab();
 197
 198     ScHTMLTable* pTable = NULL;
 199     ScHTMLTableId nTableId = SC_HTML_GLOBAL_TABLE;
 200     while( (pTable = pGlobTable->FindNestedTable( ++nTableId )) != 0 )
 201     {
 202         pTable->GetDocRange( aNewRange );
 203         aNewRange.Move( nColDiff, nRowDiff, nTabDiff );
 204         // insert table number as name
 205         InsertRangeName( mpDoc, ScfTools::GetNameFromHTMLIndex( nTableId ), aNewRange );
 206         // insert table id as name
 207         if( pTable->GetTableName().Len() )
 208         {
 209             String aName( ScfTools::GetNameFromHTMLName( pTable->GetTableName() ) );
 210             USHORT nPos;
 211             if( !mpDoc->GetRangeName()->SearchName( aName, nPos ) )
 212                 InsertRangeName( mpDoc, aName, aNewRange );
 213         }
 214     }
 215 }
 216
 217 String ScFormatFilterPluginImpl::GetHTMLRangeNameList( ScDocument* pDoc, const String& rOrigName )
 218 {
 219     return ScHTMLImport::GetHTMLRangeNameList( pDoc, rOrigName );
 220 }
 221
 222 String ScHTMLImport::GetHTMLRangeNameList( ScDocument* pDoc, const String& rOrigName )
 223 {
 224     DBG_ASSERT( pDoc, "ScHTMLImport::GetHTMLRangeNameList - missing document" );
 225
 226     String aNewName;
 227     ScRangeName* pRangeNames = pDoc->GetRangeName();
 228     ScRangeList aRangeList;
 229     xub_StrLen nTokenCnt = rOrigName.GetTokenCount( ';' );
 230     xub_StrLen nStringIx = 0;
 231     for( xub_StrLen nToken = 0; nToken < nTokenCnt; nToken++ )
 232     {
 233         String aToken( rOrigName.GetToken( 0, ';', nStringIx ) );
 234         if( pRangeNames && ScfTools::IsHTMLTablesName( aToken ) )
 235         {   // build list with all HTML tables
 236             ULONG nIndex = 1;
 237             USHORT nPos;
 238             BOOL bLoop = TRUE;
 239             while( bLoop )
 240             {
 241                 aToken = ScfTools::GetNameFromHTMLIndex( nIndex++ );
 242                 bLoop = pRangeNames->SearchName( aToken, nPos );
 243                 if( bLoop )
 244                 {
 245                     const ScRangeData* pRangeData = (*pRangeNames)[ nPos ];
 246                     ScRange aRange;
 247                     if( pRangeData && pRangeData->IsReference( aRange ) && !aRangeList.In( aRange ) )
 248                     {
 249                         ScGlobal::AddToken( aNewName, aToken, ';' );
 250                         aRangeList.Append( aRange );
 251                     }
 252                 }
 253             }
 254         }
 255         else
 256             ScGlobal::AddToken( aNewName, aToken, ';' );
 257     }
 258     return aNewName;
 259 }
 260