Unreadable PDF file error should carry information to be able to identify the source

[ediweissmann]

When processing multiple files in a batch, if the task fails because one of the files cannot be opened or properly parsed, it's hard to identify which file caused the problem.

IllegalStateException: Catalog cannot be found
        at org.sejda.sambox.cos.COSDocument.lambda$getCatalog$1(COSDocument.java:130)
        at java.util.Optional.orElseThrow(Optional.java:290)
        at org.sejda.sambox.cos.COSDocument.getCatalog(COSDocument.java:130)
        at org.sejda.sambox.pdmodel.PDDocument.getDocumentCatalog(PDDocument.java:259)
        at org.sejda.sambox.pdmodel.PDDocument.getPages(PDDocument.java:305)
        at org.sejda.impl.sambox.component.PageTreeRebuilder.rebuild(PageTreeRebuilder.java:53)
        at org.sejda.impl.sambox.RepairTask.execute(RepairTask.java:79)
        at org.sejda.impl.sambox.RepairTask.execute(RepairTask.java:48)

Suggested solution:
Bubble up an exception that also carries information about the source PDF being in context.