Author: tallison
Date: Fri Sep 27 15:45:55 2013
New Revision: 1526960
URL: http://svn.apache.org/r1526960
Log:
POI-54722 table text in ppt files
Added:
poi/trunk/test-data/slideshow/54722.ppt (with props)
Modified:
poi/trunk/src/scratchpad/src/org/apache/poi/hslf/extractor/PowerPointExtractor.java
poi/trunk/src/scratchpad/testcases/org/apache/poi/hslf/extractor/TestExtractor.java
Modified:
poi/trunk/src/scratchpad/src/org/apache/poi/hslf/extractor/PowerPointExtractor.java
URL:
http://svn.apache.org/viewvc/poi/trunk/src/scratchpad/src/org/apache/poi/hslf/extractor/PowerPointExtractor.java?rev=1526960&r1=1526959&r2=1526960&view=diff
==============================================================================
---
poi/trunk/src/scratchpad/src/org/apache/poi/hslf/extractor/PowerPointExtractor.java
(original)
+++
poi/trunk/src/scratchpad/src/org/apache/poi/hslf/extractor/PowerPointExtractor.java
Fri Sep 27 15:45:55 2013
@@ -252,6 +252,12 @@ public final class PowerPointExtractor e
// Slide text
textRunsToText(ret, slide.getTextRuns());
+ // Table text
+ for (Shape shape : slide.getShapes()){
+ if (shape instanceof Table){
+ extractTableText(ret, (Table)shape);
+ }
+ }
// Slide footer, if set
if (hf != null && hf.isFooterVisible() &&
hf.getFooterText() != null) {
ret.append(hf.getFooterText() + "\n");
@@ -306,6 +312,23 @@ public final class PowerPointExtractor e
return ret.toString();
}
+ private void extractTableText(StringBuffer ret, Table table) {
+ for (int row = 0; row < table.getNumberOfRows(); row++){
+ for (int col = 0; col < table.getNumberOfColumns(); col++){
+ TableCell cell = table.getCell(row, col);
+ //defensive null checks; don't know if they're necessary
+ if (cell != null){
+ String txt = cell.getText();
+ txt = (txt == null) ? "" : txt;
+ ret.append(txt);
+ if (col < table.getNumberOfColumns()-1){
+ ret.append("\t");
+ }
+ }
+ }
+ ret.append('\n');
+ }
+ }
private void textRunsToText(StringBuffer ret, TextRun[] runs) {
if (runs==null) {
return;
Modified:
poi/trunk/src/scratchpad/testcases/org/apache/poi/hslf/extractor/TestExtractor.java
URL:
http://svn.apache.org/viewvc/poi/trunk/src/scratchpad/testcases/org/apache/poi/hslf/extractor/TestExtractor.java?rev=1526960&r1=1526959&r2=1526960&view=diff
==============================================================================
---
poi/trunk/src/scratchpad/testcases/org/apache/poi/hslf/extractor/TestExtractor.java
(original)
+++
poi/trunk/src/scratchpad/testcases/org/apache/poi/hslf/extractor/TestExtractor.java
Fri Sep 27 15:45:55 2013
@@ -367,4 +367,23 @@ public final class TestExtractor extends
assertEquals(expectText, extractor.getText());
}
}
+
+ public void testTable() throws Exception{
+ ppe = new
PowerPointExtractor(slTests.openResourceAsStream("54111.ppt"));
+ String text = ppe.getText();
+ String target = "TH Cell 1\tTH Cell 2\tTH Cell 3\tTH Cell 4\n"+
+ "Row 1, Cell 1\tRow 1, Cell 2\tRow 1, Cell 3\tRow 1,
Cell 4\n"+
+ "Row 2, Cell 1\tRow 2, Cell 2\tRow 2, Cell 3\tRow 2,
Cell 4\n"+
+ "Row 3, Cell 1\tRow 3, Cell 2\tRow 3, Cell 3\tRow 3,
Cell 4\n"+
+ "Row 4, Cell 1\tRow 4, Cell 2\tRow 4, Cell 3\tRow 4,
Cell 4\n"+
+ "Row 5, Cell 1\tRow 5, Cell 2\tRow 5, Cell 3\tRow 5,
Cell 4\n";
+ assertTrue(text.contains(target));
+
+ ppe = new
PowerPointExtractor(slTests.openResourceAsStream("54722.ppt"));
+ text = ppe.getText();
+
+ target = "this\tText\tis\twithin\ta\n"+
+ "table\t1\t2\t3\t4";
+ assertTrue(text.contains(target));
+ }
}
Added: poi/trunk/test-data/slideshow/54722.ppt
URL:
http://svn.apache.org/viewvc/poi/trunk/test-data/slideshow/54722.ppt?rev=1526960&view=auto
==============================================================================
Binary file - no diff available.
Propchange: poi/trunk/test-data/slideshow/54722.ppt
------------------------------------------------------------------------------
svn:mime-type = application/octet-stream
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
