William M Conlon wrote on 4/4/08 5:09 PM:
> I have a list of documents to be indexed. In addition to the
> document path, the list includes other attributes that should be
> searchable, so they need to included in the index, although they may
> not be in the document itself.
> My first thought was to use -S prog, with my external program reading
> each document, generating HTML to feed swish-e, and inserting <meta
> name="lanuage" content="english"> for each attribute into the <head>
> section of the HTML.
That's what I would do.
> My second thought was that swish-e needs to accept attributes that
> are fed to the indexer with the document, perhaps in a *NEW*
> Attribute header, a la:
Would require hacking the source. And not really a good change, imo. It means
applying parsing and tokenization at the header-parsing stage, which just seems
unnecessary, especially when the MetaName feature already supports HTML or XML
tags in the document content
> And my last thought was to overload the Path-Name with the attributes
> and use ExtractPath to build metanames.
that's do-able too. But I would still use <meta> tags myself.
Peter Karman . http://peknet.com/ . peter(at)not-real.peknet.com
Users mailing list
Received on Sat Apr 5 23:40:04 2008