SolrSearch - pagination and quoted terms

Hey folks, I'm continuing down the path of utilizing Omeka to make 6000+ scanned issues of local newspapers available for one of our libraries and while it hasn't necessarily been the smoothest trip, I'm getting there.

After solving an issue related to Solr and PDF Text (hint: change the 'PDF Text' 'Text' field type to 'Item' instead of 'File' and reindex!) I'm having a couple other interesting issues with Solr Search; pagination doesn't seem to work and when I search for something in quotes, the search query isn't displayed in the search box on the results page, as terms would be if they were not in quotes.

I wonder if I've done something horribly wrong... or maybe just crazy for trying to do this with Omeka instead of Drupal, or even a from-scratch rails app.

If you take a look at the search term is correctly displayed in the search box on the page, but the pagination links all link to the current page, not page=1, page=2 etc. SolrSearch includes a pagination.php in the public views, but it doesn't seem to actually be used.

And, if you look at you'll note that "kennedy" is not displayed in the search box on the results page. Odd. (It also seems facets where if the facet names have double-quotes in them break horribly, as well -- if I narrow by the facet "Issue of "The Leader and The Kalkaskian" Newspaper.(836)" the results page displays 0 results and is labeled "Description Issue of (remove)". It seems like there is widespread mishandling of special characters (whether it's a missing rawurlencode() somewhere or what, I do not know)).

Has anyone else run into any of these little buglets or should I go open more issues on the solrsearch github? Thanks! :)

We may be fortunate not to have any facets containing the string ' AND '.

( $fParam = implode(' AND ', $fParam); )

Not too sure about how to fix that (and we don't need to so I'm not going to worry about it today!)

We've also identified issues with the facet regex that makes it break on strings containing quotes, and not identify other strings as facets ('40_s:"1881-09-29"' doesn't match, because "[\w\s]+" doesn't match -. Changing the value regex to [-\w\s]+ seems to fix that issue, but not the quoted string issue.)

We're continuing to poke at things, I've added a couple unnecessary/silly facets just for testing purposes (format, when all the items in the database have the same format, and date, when each item has a different date. Those will be removed before I switch this site to production, clearly. *grin*)

I'll see if we can get it working then I'll submit an issue to the github project for SolrSearch with our findings. Our immediate fix might simply be disabling all facets.

The pagination issue(s) appear to be related to the neatscape theme. We're chasing it down, hopefully figure it out tomorrow! =)

May I ask how you modified the PDFText Plugin to write to the item instead of to the file record?

A description of the changes by made by @wjr can be found here :