[ https://issues.apache.org/jira/browse/PDFBOX-2252?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14933277#comment-14933277 ]
Andreas Meier commented on PDFBOX-2252: --------------------------------------- I have already studied one of the two papers you posted and did some tests with lapdftext a week ago. (using the PDFTextStripperByArea.java and lapdftext 1.7.2-SNAPSHOT) The results were pretty good and lapdftext would be the tool of choice in my opinion Unfortunately I also discovered that it relies on Java 1.6, any other Java Version breaks the dependencies and will not work. That's the reason, why I mentioned the "specific Java Version" in my post. The layout analysis algorithm of lapdftext combined with pdfbox would be amazing. I have not taken an in-depth look at papers or the lapdftext-code, but I guess it would be hard work to re-code the algorithm. I have created https://issues.apache.org/jira/browse/PDFBOX-2998 > PDFTextStripper has problem with documents with mixed language directions > ------------------------------------------------------------------------- > > Key: PDFBOX-2252 > URL: https://issues.apache.org/jira/browse/PDFBOX-2252 > Project: PDFBox > Issue Type: Bug > Components: Text extraction > Affects Versions: 1.8.6, 2.0.0 > Reporter: Amir > Assignee: Maruan Sahyoun > Priority: Critical > Fix For: 2.1.0 > > Attachments: BidiMirroring.txt, IsMirroredDeviations.txt, > PDFTextStripper-201709271718.patch, PDFTextStripper-201709272018.patch, > PDFTextStripper.java.patch, PDFTextStripper.java.patch, atest.pdf, > bugzilla867751.pdf, overlap.jpg, test.pdf, wikipedia_dl_lyric_test.pdf > > > When the input document of PDFTextStripper is a combination of right-to-left > and left-to-right languages, the output characters of one language is > reversed. > A sample bilingual pdf document is attached. > PDFTextStripper has a variable "isRtlDominant" in "writePage" function, which > is defined as follows: boolean isRtlDominant = rtlCount > ltrCount; > This class clearly count the number of rtl characters and decide if the whole > content should be revered or not. It's not true, it must operate on each > word, not the whole document. -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@pdfbox.apache.org For additional commands, e-mail: dev-h...@pdfbox.apache.org