Use variant of XPath Accelerator to index nested PDF data