XSLF: text extraction from tables

git-svn-id: https://svn.apache.org/repos/asf/poi/trunk@897875 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Maxim Valyanskiy 2010-01-11 14:27:54 +00:00
parent 5621bb0800
commit f1a1f4a6ff
10 changed files with 219 additions and 44 deletions

View File

@ -16,28 +16,18 @@
==================================================================== */ ==================================================================== */
package org.apache.poi.xslf.extractor; package org.apache.poi.xslf.extractor;
import java.io.IOException;
import org.apache.poi.POIXMLTextExtractor; import org.apache.poi.POIXMLTextExtractor;
import org.apache.poi.openxml4j.exceptions.OpenXML4JException; import org.apache.poi.openxml4j.exceptions.OpenXML4JException;
import org.apache.poi.openxml4j.opc.OPCPackage; import org.apache.poi.openxml4j.opc.OPCPackage;
import org.apache.poi.xslf.XSLFSlideShow; import org.apache.poi.xslf.XSLFSlideShow;
import org.apache.poi.xslf.usermodel.DrawingParagraph;
import org.apache.poi.xslf.usermodel.XMLSlideShow; import org.apache.poi.xslf.usermodel.XMLSlideShow;
import org.apache.poi.xslf.usermodel.XSLFCommonSlideData;
import org.apache.poi.xslf.usermodel.XSLFSlide; import org.apache.poi.xslf.usermodel.XSLFSlide;
import org.apache.xmlbeans.XmlException; import org.apache.xmlbeans.XmlException;
import org.apache.xmlbeans.XmlObject; import org.openxmlformats.schemas.presentationml.x2006.main.*;
import org.apache.xmlbeans.XmlCursor;
import org.openxmlformats.schemas.drawingml.x2006.main.CTRegularTextRun; import java.io.IOException;
import org.openxmlformats.schemas.drawingml.x2006.main.CTTextBody;
import org.openxmlformats.schemas.drawingml.x2006.main.CTTextParagraph;
import org.openxmlformats.schemas.drawingml.x2006.main.CTTextLineBreak;
import org.openxmlformats.schemas.presentationml.x2006.main.CTComment;
import org.openxmlformats.schemas.presentationml.x2006.main.CTCommentList;
import org.openxmlformats.schemas.presentationml.x2006.main.CTGroupShape;
import org.openxmlformats.schemas.presentationml.x2006.main.CTNotesSlide;
import org.openxmlformats.schemas.presentationml.x2006.main.CTShape;
import org.openxmlformats.schemas.presentationml.x2006.main.CTSlide;
import org.openxmlformats.schemas.presentationml.x2006.main.CTSlideIdListEntry;
public class XSLFPowerPointExtractor extends POIXMLTextExtractor { public class XSLFPowerPointExtractor extends POIXMLTextExtractor {
private XMLSlideShow slideshow; private XMLSlideShow slideshow;
@ -110,7 +100,7 @@ public class XSLFPowerPointExtractor extends POIXMLTextExtractor {
slideshow._getXSLFSlideShow().getSlideComments(slideId); slideshow._getXSLFSlideShow().getSlideComments(slideId);
if(slideText) { if(slideText) {
extractText(rawSlide.getCSld().getSpTree(), text); extractText(slides[i].getCommonSlideData(), text);
// Comments too for the slide // Comments too for the slide
if(comments != null) { if(comments != null) {
@ -123,8 +113,9 @@ public class XSLFPowerPointExtractor extends POIXMLTextExtractor {
} }
} }
} }
if(notesText && notes != null) { if(notesText && notes != null) {
extractText(notes.getCSld().getSpTree(), text); extractText(new XSLFCommonSlideData(notes.getCSld()), text);
} }
} catch(Exception e) { } catch(Exception e) {
throw new RuntimeException(e); throw new RuntimeException(e);
@ -134,31 +125,10 @@ public class XSLFPowerPointExtractor extends POIXMLTextExtractor {
return text.toString(); return text.toString();
} }
private void extractText(CTGroupShape gs, StringBuffer text) { private void extractText(XSLFCommonSlideData data, StringBuffer text) {
CTShape[] shapes = gs.getSpArray(); for (DrawingParagraph p : data.getText()) {
for (int i = 0; i < shapes.length; i++) { text.append(p.getText());
CTTextBody textBody = text.append("\n");
shapes[i].getTxBody(); }
if(textBody != null) { }
CTTextParagraph[] paras =
textBody.getPArray();
for (int j = 0; j < paras.length; j++) {
XmlCursor c = paras[j].newCursor();
c.selectPath("./*");
while (c.toNextSelection()) {
XmlObject o = c.getObject();
if(o instanceof CTRegularTextRun){
CTRegularTextRun txrun = (CTRegularTextRun)o;
text.append( txrun.getT() );
} else if (o instanceof CTTextLineBreak){
text.append('\n');
}
}
// End each paragraph with a new line
text.append("\n");
}
}
}
}
} }

View File

@ -0,0 +1,33 @@
package org.apache.poi.xslf.usermodel;
import org.openxmlformats.schemas.drawingml.x2006.main.CTTextParagraph;
import org.openxmlformats.schemas.drawingml.x2006.main.CTRegularTextRun;
import org.openxmlformats.schemas.drawingml.x2006.main.CTTextLineBreak;
import org.apache.xmlbeans.XmlCursor;
import org.apache.xmlbeans.XmlObject;
public class DrawingParagraph {
private final CTTextParagraph p;
public DrawingParagraph(CTTextParagraph p) {
this.p = p;
}
public CharSequence getText() {
StringBuilder text = new StringBuilder();
XmlCursor c = p.newCursor();
c.selectPath("./*");
while (c.toNextSelection()) {
XmlObject o = c.getObject();
if (o instanceof CTRegularTextRun) {
CTRegularTextRun txrun = (CTRegularTextRun) o;
text.append(txrun.getT());
} else if (o instanceof CTTextLineBreak) {
text.append('\n');
}
}
return text;
}
}

View File

@ -0,0 +1,23 @@
package org.apache.poi.xslf.usermodel;
import org.openxmlformats.schemas.drawingml.x2006.main.CTTable;
import org.openxmlformats.schemas.drawingml.x2006.main.CTTableRow;
public class DrawingTable {
private final CTTable table;
public DrawingTable(CTTable table) {
this.table = table;
}
public DrawingTableRow[] getRows() {
CTTableRow[] ctTableRows = table.getTrArray();
DrawingTableRow[] o = new DrawingTableRow[ctTableRows.length];
for (int i=0; i<o.length; i++) {
o[i] = new DrawingTableRow(ctTableRows[i]);
}
return o;
}
}

View File

@ -0,0 +1,17 @@
package org.apache.poi.xslf.usermodel;
import org.openxmlformats.schemas.drawingml.x2006.main.CTTableCell;
public class DrawingTableCell {
private final CTTableCell cell;
private final DrawingTextBody drawingTextBody;
public DrawingTableCell(CTTableCell cell) {
this.cell = cell;
drawingTextBody = new DrawingTextBody(this.cell.getTxBody());
}
public DrawingTextBody getTextBody() {
return drawingTextBody;
}
}

View File

@ -0,0 +1,23 @@
package org.apache.poi.xslf.usermodel;
import org.openxmlformats.schemas.drawingml.x2006.main.CTTableRow;
import org.openxmlformats.schemas.drawingml.x2006.main.CTTableCell;
public class DrawingTableRow {
private final CTTableRow row;
public DrawingTableRow(CTTableRow row) {
this.row = row;
}
public DrawingTableCell[] getCells() {
CTTableCell[] ctTableCells = row.getTcArray();
DrawingTableCell[] o = new DrawingTableCell[ctTableCells.length];
for (int i=0; i<o.length; i++) {
o[i] = new DrawingTableCell(ctTableCells[i]);
}
return o;
}
}

View File

@ -0,0 +1,23 @@
package org.apache.poi.xslf.usermodel;
import org.openxmlformats.schemas.drawingml.x2006.main.CTTextBody;
import org.openxmlformats.schemas.drawingml.x2006.main.CTTextParagraph;
public class DrawingTextBody {
private final CTTextBody textBody;
public DrawingTextBody(CTTextBody textBody) {
this.textBody = textBody;
}
public DrawingParagraph[] getParagraphs() {
CTTextParagraph[] pArray = textBody.getPArray();
DrawingParagraph[] o = new DrawingParagraph[pArray.length];
for (int i=0; i<o.length; i++) {
o[i] = new DrawingParagraph(pArray[i]);
}
return o;
}
}

View File

@ -0,0 +1,67 @@
package org.apache.poi.xslf.usermodel;
import org.apache.xmlbeans.XmlCursor;
import org.apache.xmlbeans.XmlObject;
import org.openxmlformats.schemas.drawingml.x2006.main.CTGraphicalObjectData;
import org.openxmlformats.schemas.drawingml.x2006.main.CTTable;
import org.openxmlformats.schemas.drawingml.x2006.main.CTTextBody;
import org.openxmlformats.schemas.presentationml.x2006.main.CTCommonSlideData;
import org.openxmlformats.schemas.presentationml.x2006.main.CTGraphicalObjectFrame;
import org.openxmlformats.schemas.presentationml.x2006.main.CTGroupShape;
import org.openxmlformats.schemas.presentationml.x2006.main.CTShape;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;
public class XSLFCommonSlideData {
private final CTCommonSlideData data;
public XSLFCommonSlideData(CTCommonSlideData data) {
this.data = data;
}
public List<DrawingParagraph> getText() {
CTGroupShape gs = data.getSpTree();
List<DrawingParagraph> out = new ArrayList<DrawingParagraph>();
CTShape[] shapes = gs.getSpArray();
for (int i = 0; i < shapes.length; i++) {
CTTextBody ctTextBody = shapes[i].getTxBody();
if (ctTextBody==null) {
continue;
}
DrawingTextBody textBody = new DrawingTextBody(ctTextBody);
out.addAll(Arrays.asList(textBody.getParagraphs()));
}
CTGraphicalObjectFrame[] graphicFrames = gs.getGraphicFrameArray();
for (CTGraphicalObjectFrame frame: graphicFrames) {
CTGraphicalObjectData data = frame.getGraphic().getGraphicData();
XmlCursor c = data.newCursor();
c.selectPath("./*");
while (c.toNextSelection()) {
XmlObject o = c.getObject();
if (o instanceof CTTable) {
DrawingTable table = new DrawingTable((CTTable) o);
for (DrawingTableRow row : table.getRows()) {
for (DrawingTableCell cell : row.getCells()) {
DrawingTextBody textBody = cell.getTextBody();
out.addAll(Arrays.asList(textBody.getParagraphs()));
}
}
}
}
}
return out;
}
}

View File

@ -26,11 +26,13 @@ import org.openxmlformats.schemas.presentationml.x2006.main.CTSlideIdListEntry;
public class XSLFSlide extends XSLFSheet implements Slide { public class XSLFSlide extends XSLFSheet implements Slide {
private CTSlide slide; private CTSlide slide;
private CTSlideIdListEntry slideId; private CTSlideIdListEntry slideId;
private XSLFCommonSlideData data;
public XSLFSlide(CTSlide slide, CTSlideIdListEntry slideId, SlideShow parent) { public XSLFSlide(CTSlide slide, CTSlideIdListEntry slideId, SlideShow parent) {
super(parent); super(parent);
this.slide = slide; this.slide = slide;
this.slideId = slideId; this.slideId = slideId;
this.data = new XSLFCommonSlideData(slide.getCSld());
} }
/** /**
@ -88,4 +90,8 @@ public class XSLFSlide extends XSLFSheet implements Slide {
// TODO Auto-generated method stub // TODO Auto-generated method stub
} }
public XSLFCommonSlideData getCommonSlideData() {
return data;
}
} }

View File

@ -113,4 +113,17 @@ public class TestXSLFPowerPointExtractor extends TestCase {
// Check comments are there // Check comments are there
assertTrue("Unable to find expected word in text\n" + text, text.contains("testdoc")); assertTrue("Unable to find expected word in text\n" + text, text.contains("testdoc"));
} }
public void testTable() throws Exception {
POIDataSamples slTests = POIDataSamples.getSlideShowInstance();
xmlA = new XSLFSlideShow(OPCPackage.open(slTests.openResourceAsStream("present1.pptx")));
XSLFPowerPointExtractor extractor =
new XSLFPowerPointExtractor(xmlA);
String text = extractor.getText();
assertTrue(text.length() > 0);
// Check comments are there
assertTrue("Unable to find expected word in text\n" + text, text.contains("TEST"));
}
} }

Binary file not shown.