[
https://issues.apache.org/jira/browse/TIKA-1288?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16468483#comment-16468483
]
Manuel Iglesias commented on TIKA-1288:
---------------------------------------
I checked Tika 1.18 and the bug is still there. Most .epub documents are read
incorrectly (either no content or only partial content). I looked inside some
of those documents and I saw that the text is contained in several '.htm'
files. The relevant Tika parser 'org.apache.tika.parser.epub.EpubParser'
doesn't detect '.htm' files. I think the bug could be solved by changing just
one line:
'else if (entry.getName().endsWith(".html") ||
entry.getName().endsWith(".xhtml"))' ->
'else if (entry.getName().endsWith(".html") ||
(entry.getName().endsWith(".htm") || entry.getName().endsWith(".xhtml"))'
> Epub's content extracted partially
> ----------------------------------
>
> Key: TIKA-1288
> URL: https://issues.apache.org/jira/browse/TIKA-1288
> Project: Tika
> Issue Type: Bug
> Components: parser
> Affects Versions: 1.5
> Environment: win 8, jre 1.5
> Reporter: Alex Andrushchak
> Priority: Major
> Attachments: bad-parsed.epub, epub30-spec-20121128.epub
>
>
> There are 11 parts (xhtml files) in epub document, but tika extracted content
> from single file only. i extracted all files from epub's zip and all of them
> are valid xhtml files.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
