[ https://issues.apache.org/jira/browse/TIKA-2845?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16808708#comment-16808708 ]
Tim Allison edited comment on TIKA-2845 at 4/3/19 1:17 PM: ----------------------------------------------------------- The attached file opens in MS Edge, has no "contents" element but does have an embedded file in an annotation that can be downloaded/exported w Edge. We are not currently extracting this embedded file. NOTE: Adobe DC Reader is not able to open this file. was (Author: talli...@mitre.org): The attached file opens in Adobe, has no "contents" element but does have an embedded file in an annotation that can be downloaded/exported w Adobe. We are not currently extracting this embedded file. > Override ProcessPages in PDFTextStripper > ---------------------------------------- > > Key: TIKA-2845 > URL: https://issues.apache.org/jira/browse/TIKA-2845 > Project: Tika > Issue Type: Task > Reporter: Tim Allison > Priority: Major > Attachments: testPDFFileEmbInAnnotation_noContents.pdf > > > On the PDFBox user list, [~lehmi] confirmed (and [~tilman] clarified) that > PDFTextStripper's {{processPages}} skips pages that lack a "Contents" > element. Inline images are part of the "Contents" element and would still be > processed (e.g. in OCR). > > However, there are other elements that might be on a page that does not have > a "Contents" element, such as an annotation with an embedded file. > > We should override {{processPages()}} to process all pages. -- This message was sent by Atlassian JIRA (v7.6.3#76005)