I hate to mess up this nice groups page with newbie issues but looks like I will.
The ApacheSolr module looks to be very cool but it's not what I need. It's one thing to backend all of Drupal with Solr but it's something altogether different (I think) to integrate Drupal with an existing Solr implementation.
I'm more interested in indexing independent data (already indexed) with Solr and then providing the wealth of Drupal features to view and manipulate that data. I suppose this is more like the OpenSearch Aggregation module? (going to download and study that now) Or more like the original Solr module - though I'm also struggling with figuring that one out as yet.
I'm not quite sure yet if it's even possible to get what I want. I don't know how returned data from Solr might integrate with existing nodes. Maybe a special type of node that's separate but equal to internal nodes and is abstracted to make Drupal think it's just a regular node?
It would be cool if I could just create a content type that has fields (CCK fields?) that correspond to my existing Solr schema and maintains the relationship between Solr and Drupal. Maybe if I ever get that far along and I'm not misunderstanding some core concept, that will be something I work on, someday.
I've even considered re-indexing my current schema into the ApacheSolr schema and converting my data to the new schema and just using it that way but it's not really what I need and it couples my data too tightly to Drupal. Though anything other than the ApacheSolr module way is daunting at this point. I'm beginning to wonder if I'm just barking up the wrong tree.
Any thoughts or comments? (or help?)
Comments
on my way
Well, I've used apachesolr to create a new module called solrclient. (Thank You). I toyed with the idea of having all of my solr docs as nodes but I saw that every node needs a mysql row. Well, I'm planning on having over a million docs and I'm familiar with the limitations of mysql and I don't even know how that might affect Drupal.
Since my solr data is "read only" to users, this works well for me. Maybe I'll implement a way for users to add docs directly to solr in the future but it's just not a need for me now.
So, my solrclient just queries solr and displays. I plan to integrate with nodeprofile so that a user can reference docs returned from solr by id in cck fields. Then, when access to the doc is required, it's just a quick call to solr. When I have some more of this functionality done, I'll submit it.