manifoldcf-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Karl Wright <daddy...@gmail.com>
Subject Re: Mapping fields from Wiki to Solr
Date Tue, 30 Sep 2014 13:50:59 GMT
Hi Kambiz,

A great way to debug what fields are getting through with what names is to
index with Solr, but to look at the [INFO] messages in the solr log.  Every
document that is indexed includes a full URL which has the fields listed as
arguments.  This will help you figure out whether any metadata is making it
through, and what it is called, without needing to search the index.  If
you remove all field mappings in Solr and on the Tika extractor, and try
indexing a small number of documents, it will be very obvious what metadata
fields are in fact being sent downstream by the repository connector.

It's also not clear whether you are using the web connector to crawl your
wiki site, or the wiki connector.  If it's the wiki connector, it only
supports the following:

>>>>>>
                if (comment != null)
                  rd.addField("comment",comment);
                if (author != null)
                  rd.addField("author",author);
                if (title != null)
                  rd.addField("title",title);
                if (lastModified != null)
                {
                  rd.addField("last-modified",lastModified);

rd.setModifiedDate(DateParser.parseISO8601Date(lastModified));
                }
<<<<<<

Anything else that you want to support, you should create a ticket for.

Thanks!
Karl


On Tue, Sep 30, 2014 at 9:22 AM, Kambiz Niktabar <niktabar@yahoo.com> wrote:

> Hello,
>
> I tried to crawling a wiki site by Manifold CF and I wanted to map some
> fields to the Solr fields. I used both "Solr Field Mapping" tab in the job
> and also "Field Mapping" tab (added to the job by Tika content extractor
> connection) but none of them are working and I don't see any value for
> those fields in Solr.
> For example I did the following mapping:
>
>
> Metadata field name                         Solr field name
>
> fullurl                                                url
> user                                                  xpersonname
>
> Can you please advise?
>
> Regards
> Kambiz Niktabar
>

Mime
View raw message