[
https://issues.apache.org/jira/browse/NUTCH-3032?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17832463#comment-17832463
]
ASF GitHub Bot commented on NUTCH-3032:
---------------------------------------
CatChullain commented on PR #810:
URL: https://github.com/apache/nutch/pull/810#issuecomment-2028122038
Thanks, Lewis! I moved all four to DEBUG, but I wonder where might be good
spots for INFO level messages. I'm thinking of the operator or tech who doesn't
dig into code and has an issue in the config. During dev & test myself, I
sometimes forgot to increment the index.arbitrary.function.count and the plugin
ignored the later fields. Just outputting that count value, and maybe something
when overwrite is true, might be helpful for alerting someone that the config
might not be what they'd believed.
Do either of those (or something else) seem worthwhile, or does it make more
sense to let people use it and see what issues they raise?
> Indexing plugin as an adapter for end user's own POJO instances
> ---------------------------------------------------------------
>
> Key: NUTCH-3032
> URL: https://issues.apache.org/jira/browse/NUTCH-3032
> Project: Nutch
> Issue Type: Improvement
> Components: indexer
> Reporter: Joe Gilvary
> Priority: Major
> Labels: indexing
> Attachments: NUTCH-3032.patch
>
>
> It could be helpful to let end users manipulate information at indexing time
> with their own code without the need for writing their own indexing plugin. I
> mentioned this on the dev mailing list
> (https://www.mail-archive.com/dev@nutch.apache.org/msg31190.html) with some
> description of my work in progress.
> One potential use is to address some of the same concerns that NUTCH-585
> discusses regarding an alternative approach to picking and choosing which
> content to index, but this approach would allow making index time decisions,
> rather than setting the configuration for all content at the start of the
> indexing run.
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
