On Tue, Feb 7, 2012 at 15:33, Gary Holmes <[email protected]> wrote: > Hello all, > > I've setup full text indexing in RT4.0.4 as follows: > > - Vanilla install RT 4.0.4 on Ubuntu 10.04 LTS > - MySQL 5.1.61 built from source with mysqlse added > - Sphinx 2.0.3 added using the x86 package from the sphinxsearch.com site > - run /opt/rt4/sbin/rt-setup-fulltext-index etc > As the version of Sphinx had id64 enabled, I had to change the type of the > id field in the AttachmentsIndex to bigint to make it 64 bit. Other than > that, no issues (just a lot of time!!). RT works just fine. > > Here are my questions: > > - I can run manual searches (using search) and see hits that RT does not (0 > tickets found), others work OK. > - A lot of the data we need to search on is in entries that are of type html > not plain text. How can we index those too?
Take a look inside sbin/rt-fulltext-indexer. It's designed to process more than text and can be extended with additional functions to convert any content type to text or any other format that sphinx supports. > Any clues on how to proceed. If we can sort this, we want to index our own > RT instance and then offer this as a feature to other potential customers. > > Thanks, > Gary > > > -------- > RT Training Sessions (http://bestpractical.com/services/training.html) > * Boston — March 5 & 6, 2012 -- Best regards, Ruslan. -------- RT Training Sessions (http://bestpractical.com/services/training.html) * Boston March 5 & 6, 2012
