[PDFBOX-4003] Can't retrieve number tree from structure tree

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.0.8
• Fix Version/s: 2.0.14, 3.0.0 PDFBox
• Component/s: PDModel
• Labels: StructureTree
• Assignee: Tilman Hausherr
• Reporter: Tilman Hausherr
• Created: 09/Nov/17 16:50
• Updated: 28/Feb/19 17:44
• Resolved: 01/Jan/19 07:49

Description

PDDocument doc = PDDocument.load(new File("GeneralForbearance.pdf"));
Object value = doc.getDocumentCatalog().getStructureTreeRoot().getParentTree().getValue(0);

The code above always fails when used on a PDF with a structure tree:

Exception in thread "main" java.io.IOException: Error while trying to create value in number tree:org.apache.pdfbox.cos.COSBase.<init>(org.apache.pdfbox.cos.COSArray)
	at org.apache.pdfbox.pdmodel.common.PDNumberTreeNode.convertCOSToPD(PDNumberTreeNode.java:212)
	at org.apache.pdfbox.pdmodel.common.PDNumberTreeNode.getNumbers(PDNumberTreeNode.java:185)
	at org.apache.pdfbox.pdmodel.common.PDNumberTreeNode.getValue(PDNumberTreeNode.java:139)
	at pdfboxpageimageextraction.MergeTest.main(MergeTest.java:29)
Caused by: java.lang.NoSuchMethodException: org.apache.pdfbox.cos.COSBase.<init>(org.apache.pdfbox.cos.COSArray)
	at java.lang.Classj.getConstructor0(Class.java:3082)
	at java.lang.Class.getDeclaredConstructor(Class.java:2178)
	at org.apache.pdfbox.pdmodel.common.PDNumberTreeNode.convertCOSToPD(PDNumberTreeNode.java:206)
	... 3 more

I suspect that it is related to the PDNumberTreeNode having been called with a COSBase class parameter in getParentTree().

That one doesn't have a constructor with a parameter.

The structure tree numbers tree has mixed contents, these can be arrays or dictionaries.

What we need is some PD wrapper that can have both inside.

Issue Links

• is depended upon by: PDFBOX-4418 ParentTree can't be merged if it has several levels