[
https://issues.apache.org/jira/browse/NUTCH-2168?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15083285#comment-15083285
]
Sebastian Nagel commented on NUTCH-2168:
----------------------------------------
Hi [~kalanya], looks like the indexed raw content of the JPEGs are causing the
invalid utf-8 character. The index-html plugin tries to treat any raw content
as readable content converting it to a String based on the platform-dependent
charset. What happens if the field content is specified as "binary" in
schema.xml (cf. patch for NUTCH-2130)?
Without the patch applied, non-HTML documents simply fail to parse and are
never indexed. That's probably the reason why the fix for this issue causes the
problem with index-html. I would suggest to open a separate issue to address
the indexer-solr problem with raw content from index-html.
> Parse-tika fails to retrieve parser
> -----------------------------------
>
> Key: NUTCH-2168
> URL: https://issues.apache.org/jira/browse/NUTCH-2168
> Project: Nutch
> Issue Type: Bug
> Components: parser
> Affects Versions: 2.3.1
> Reporter: Sebastian Nagel
> Fix For: 2.3.1
>
> Attachments: NUTCH-2168.patch
>
>
> The plugin parse-tika fails to parse most (all?) kinds of document types
> (PDF, xlsx, ...) when run via ParserChecker or ParserJob:
> {noformat}
> 2015-11-12 19:14:30,903 INFO parse.ParserJob - Parsing
> http://localhost/pdftest.pdf
> 2015-11-12 19:14:30,905 INFO parse.ParserFactory - ...
> 2015-11-12 19:14:30,907 ERROR tika.TikaParser - Can't retrieve Tika parser
> for mime-type application/pdf
> 2015-11-12 19:14:30,913 WARN parse.ParseUtil - Unable to successfully parse
> content http://localhost/pdftest.pdf of type application/pdf
> {noformat}
> The same document is successfully parsed by TestPdfParser.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
