[ https://issues.apache.org/jira/browse/TIKA-2310?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17012593#comment-17012593 ]
Alexey Zhukov commented on TIKA-2310: ------------------------------------- Opf file does correctly processed, but EpubParser implementation presumes that spine contents are to be placed into htm and html files only (see EpubParser.java:282) and ignores those with different type. But looks like EPUB specification ([link|[https://www.w3.org/publishing/epub3/epub-spec.html#dfn-epub-content-document]]) does allow file extension that are differ from htm/html and there may exist epub files (see attached) that can't be correctly parsed [^Dzhordzh_Oruell_1984_en_.epub] > Try to order chapters in epub correctly > --------------------------------------- > > Key: TIKA-2310 > URL: https://issues.apache.org/jira/browse/TIKA-2310 > Project: Tika > Issue Type: Bug > Reporter: Tim Allison > Assignee: Tim Allison > Priority: Minor > Fix For: 1.21 > > Attachments: Dzhordzh_Oruell_1984_en_.epub > > > [~johanvanderknijff] recently pointed out on twitter that our Epub parser > doesn't handle chapters in the right order. We should try to fix our parser > so that the output is in the correct order. > Epub is new to me, but it looks like we can scrape the order out of > content.opf(?). > This would require dumping the stream to a ZipFile for direct access to zip > entries, but we require that of ooxml... -- This message was sent by Atlassian Jira (v8.3.4#803005)