Index: tika-parsers/src/main/java/org/apache/tika/parser/microsoft/ooxml/XSLFPowerPointExtractorDecorator.java =================================================================== --- tika-parsers/src/main/java/org/apache/tika/parser/microsoft/ooxml/XSLFPowerPointExtractorDecorator.java (revision 1212405) +++ tika-parsers/src/main/java/org/apache/tika/parser/microsoft/ooxml/XSLFPowerPointExtractorDecorator.java (working copy) @@ -16,10 +16,6 @@ */ package org.apache.tika.parser.microsoft.ooxml; -import java.io.IOException; -import java.util.ArrayList; -import java.util.List; - import org.apache.poi.openxml4j.exceptions.InvalidFormatException; import org.apache.poi.openxml4j.opc.PackagePart; import org.apache.poi.openxml4j.opc.PackagePartName; @@ -28,22 +24,30 @@ import org.apache.poi.openxml4j.opc.TargetMode; import org.apache.poi.xslf.XSLFSlideShow; import org.apache.poi.xslf.extractor.XSLFPowerPointExtractor; -import org.apache.poi.xslf.usermodel.DrawingParagraph; +import org.apache.poi.xslf.usermodel.Placeholder; import org.apache.poi.xslf.usermodel.XMLSlideShow; -import org.apache.poi.xslf.usermodel.XSLFCommonSlideData; +import org.apache.poi.xslf.usermodel.XSLFComments; +import org.apache.poi.xslf.usermodel.XSLFGroupShape; import org.apache.poi.xslf.usermodel.XSLFRelation; +import org.apache.poi.xslf.usermodel.XSLFShape; +import org.apache.poi.xslf.usermodel.XSLFSheet; import org.apache.poi.xslf.usermodel.XSLFSlide; -import org.apache.poi.xslf.usermodel.XSLFSlideMaster; +import org.apache.poi.xslf.usermodel.XSLFTable; +import org.apache.poi.xslf.usermodel.XSLFTableCell; +import org.apache.poi.xslf.usermodel.XSLFTableRow; +import org.apache.poi.xslf.usermodel.XSLFTextShape; import org.apache.tika.exception.TikaException; import org.apache.tika.parser.ParseContext; import org.apache.tika.sax.XHTMLContentHandler; import org.apache.xmlbeans.XmlException; import org.openxmlformats.schemas.presentationml.x2006.main.CTComment; -import org.openxmlformats.schemas.presentationml.x2006.main.CTCommentList; -import org.openxmlformats.schemas.presentationml.x2006.main.CTNotesSlide; import org.openxmlformats.schemas.presentationml.x2006.main.CTSlideIdListEntry; import org.xml.sax.SAXException; +import java.io.IOException; +import java.util.ArrayList; +import java.util.List; + public class XSLFPowerPointExtractorDecorator extends AbstractOOXMLExtractor { public XSLFPowerPointExtractorDecorator(ParseContext context, XSLFPowerPointExtractor extractor) { @@ -54,110 +58,106 @@ * @see org.apache.poi.xslf.extractor.XSLFPowerPointExtractor#getText() */ @Override - protected void buildXHTML(XHTMLContentHandler xhtml) throws SAXException, - XmlException, IOException { + protected void buildXHTML(XHTMLContentHandler xhtml) throws SAXException, IOException { XMLSlideShow slideShow = (XMLSlideShow) extractor.getDocument(); - XSLFSlideShow rawSlideShow = null; - try { - rawSlideShow = slideShow._getXSLFSlideShow(); // TODO Avoid this in future - } catch(Exception e) { - throw new IOException(e.getMessage()); // Shouldn't happen - } XSLFSlide[] slides = slideShow.getSlides(); for (XSLFSlide slide : slides) { - // Find the ID, until we ditch the raw slideshow - CTSlideIdListEntry slideId = null; - for(CTSlideIdListEntry id : rawSlideShow.getSlideReferences().getSldIdList()) { - if(rawSlideShow.getSlidePart(id).getPartName().equals(slide.getPackagePart().getPartName())) { - slideId = id; - } - } - if(slideId == null) { - // This shouldn't normally happen - continue; - } - - XSLFSlideMaster master = slide.getMasterSheet(); - CTNotesSlide notes = rawSlideShow.getNotes(slideId); - CTCommentList comments = rawSlideShow.getSlideComments(slideId); + // slide + extractContent(slide.getShapes(), false, xhtml); - // TODO In POI 3.8 beta 5, improve how we get this - xhtml.startElement("div"); - XSLFCommonSlideData common = new XSLFCommonSlideData(slide.getXmlObject().getCSld()); - extractShapeContent(common, xhtml); + // slide layout which is the master sheet for this slide + XSLFSheet slideLayout = slide.getMasterSheet(); + extractContent(slideLayout.getShapes(), true, xhtml); - // If there are comments, extract them + // slide master which is the master sheet for all text layouts + XSLFSheet slideMaster = slideLayout.getMasterSheet(); + extractContent(slideMaster.getShapes(), true, xhtml); + + // notes (if present) + XSLFSheet slideNotes = slide.getNotes(); + if (slideNotes != null) { + extractContent(slideNotes.getShapes(), false, xhtml); + + // master sheet for this notes + XSLFSheet notesMaster = slideNotes.getMasterSheet(); + extractContent(notesMaster.getShapes(), true, xhtml); + } + + // comments (if present) + XSLFComments comments = slide.getComments(); if (comments != null) { - for (CTComment comment : comments.getCmArray()) { + for (CTComment comment : comments.getCTCommentsList().getCmList()) { xhtml.element("p", comment.getText()); } } - - // Get text from the master slide - // TODO: re-enable this once we fix TIKA-712 - /* - if(master != null) { - // TODO In POI 3.8 beta 5, improve how we get this - extractShapeContent(new XSLFCommonSlideData(master.getXmlObject().getCSld()), xhtml); - } - */ - - if (notes != null) { - // TODO In POI 3.8 beta 5, improve how we get this - extractShapeContent(new XSLFCommonSlideData(notes.getCSld()), xhtml); - } - xhtml.endElement("div"); } } - private void extractShapeContent(XSLFCommonSlideData data, XHTMLContentHandler xhtml) + private void extractContent(XSLFShape[] shapes, boolean skipPlaceholders, XHTMLContentHandler xhtml) throws SAXException { - for (DrawingParagraph p : data.getText()) { - xhtml.element("p", p.getText().toString()); + for (XSLFShape sh : shapes) { + if (sh instanceof XSLFTextShape) { + XSLFTextShape txt = (XSLFTextShape) sh; + Placeholder ph = txt.getTextType(); + if (skipPlaceholders && ph != null) { + continue; + } + xhtml.element("p", txt.getText()); + } else if (sh instanceof XSLFGroupShape){ + // recurse into groups of shapes + XSLFGroupShape group = (XSLFGroupShape)sh; + extractContent(group.getShapes(), skipPlaceholders, xhtml); + } else if (sh instanceof XSLFTable){ + XSLFTable tbl = (XSLFTable)sh; + for(XSLFTableRow row : tbl){ + List cells = row.getCells(); + extractContent(cells.toArray(new XSLFTableCell[cells.size()]), skipPlaceholders, xhtml); + } + } } } - + /** * In PowerPoint files, slides have things embedded in them, - * and slide drawings which have the images + * and slide drawings which have the images */ @Override protected List getMainDocumentParts() throws TikaException { - List parts = new ArrayList(); - XMLSlideShow slideShow = (XMLSlideShow) extractor.getDocument(); - XSLFSlideShow document = null; - try { - document = slideShow._getXSLFSlideShow(); // TODO Avoid this in future - } catch(Exception e) { - throw new TikaException(e.getMessage()); // Shouldn't happen - } - - for (CTSlideIdListEntry ctSlide : document.getSlideReferences().getSldIdList()) { - // Add the slide - PackagePart slidePart; - try { - slidePart = document.getSlidePart(ctSlide); - } catch(IOException e) { - throw new TikaException("Broken OOXML file", e); - } catch(XmlException xe) { - throw new TikaException("Broken OOXML file", xe); - } - parts.add(slidePart); - - // If it has drawings, return those too - try { - for(PackageRelationship rel : slidePart.getRelationshipsByType(XSLFRelation.VML_DRAWING.getRelation())) { - if(rel.getTargetMode() == TargetMode.INTERNAL) { - PackagePartName relName = PackagingURIHelper.createPartName(rel.getTargetURI()); - parts.add( rel.getPackage().getPart(relName) ); + List parts = new ArrayList(); + XMLSlideShow slideShow = (XMLSlideShow) extractor.getDocument(); + XSLFSlideShow document = null; + try { + document = slideShow._getXSLFSlideShow(); // TODO Avoid this in future + } catch (Exception e) { + throw new TikaException(e.getMessage()); // Shouldn't happen + } + + for (CTSlideIdListEntry ctSlide : document.getSlideReferences().getSldIdList()) { + // Add the slide + PackagePart slidePart; + try { + slidePart = document.getSlidePart(ctSlide); + } catch (IOException e) { + throw new TikaException("Broken OOXML file", e); + } catch (XmlException xe) { + throw new TikaException("Broken OOXML file", xe); + } + parts.add(slidePart); + + // If it has drawings, return those too + try { + for (PackageRelationship rel : slidePart.getRelationshipsByType(XSLFRelation.VML_DRAWING.getRelation())) { + if (rel.getTargetMode() == TargetMode.INTERNAL) { + PackagePartName relName = PackagingURIHelper.createPartName(rel.getTargetURI()); + parts.add(rel.getPackage().getPart(relName)); + } } - } - } catch(InvalidFormatException e) { - throw new TikaException("Broken OOXML file", e); - } - } + } catch (InvalidFormatException e) { + throw new TikaException("Broken OOXML file", e); + } + } - return parts; + return parts; } }