Some engines, including Solr, have support for indexing file contents.
We should do 3 things:
- Update search structure to support that concept (each engine would need to declare file indexing as a feature)
- Update the solr engine to do this
- Figure out what files we index
For number 2, the limitation is that the pecl-solr plugin doesn't support file indexing from what I can tell, but solr itself does. The way around that is to setup your own curl object make the connection, and send the file as a post.
Number 3 is also a little less clear. For some things, like assign attachments and files in a multi-file resource, you can index all the files, as anybody who can see the item can see all the files in those areas. For single file resources, I think we can only safely index the marked "main" file.