Book Module Solr Facet

Sure, however we need to index that information, otherwise there is nothing to facet.

One initial suggestion - you might want to index a bid of 0 for nodes not in any book (see the OG integration) in order to facet/filter-out those nodes not in any book.

An alternative is here: http://www.nabble.com/How-to-know-if-a-field-is-null--td19114164.html

something like: &fq=-is_book_bid:[* TO *]

but I think there is still an outstanding issue to make the query class support negative queries.

Yes - we certainly intend to make refinements to the API to make this easier, there are a couple open issues.

Hmm, yes - just made an interesting observation: we can get a facet count for all those docs missing the field by adding to the params:

$params['f.is_book_bid.facet.missing'] = 'true';

This is better in many ways than indexing a 0, since it requires re-indexing only nodes in books, not all nodes.

This same technique should go in the OG module - the only trick is figureing out how to make a proper facet link.

doing this get results back like:

            [im_vid_1] => stdClass Object
                (
                    [_empty_] => 20
                )

So you can look for the key "_empty_" as we do already for language.

see: http://drupal.org/node/348029#comment-1458656

Given that book module is in core and the code here is pretty minimal, we should probably just make this part of the main module - especially if we can get the negative queries working right.

Also, did you just copy this over and modify from the OG module?

$params['fl'] .= ',is_book_bid';

I think I added that kind of code there since we were considering adding links to the groups to each search result. Based on the code here, there's no read to add the bid to search results.

hey, just realized that you're the fabulous Jody that AlexUA kept mentioning at our meetup!

Anyhow - I need to sort the minor APi changes for the Views integration people, then we can toss in the negative query support and this faceting will be like 10 lines in the main module (like it supports taxonomy now).

ok, this is basically working (includes some other misc improvements).

taxonomy doesn't properly handle missing facets, and I'm not sure about language.

stray !

plus breadcrumb theme function and better handling for $node->language.

Two remarks:

+  $face_missing = variable_get('apachesolr_facet_missing', array());

should probably be $facet_missing

+/**
+ * Enable facet missing for books.
+ */
+function apachesolr_search_update_6003() {
+  $ret = array();
+  $facet_missing = variable_get('apachesolr_facet_missing', array());
+  $facet_missing['apachesolr_search']['is_book_bid'] = 1;
+  variable_set('apachesolr_facet_missing', $facet_missing);
+  return $ret;
+}

and

+  // A book module facet.
+  $facets['is_book_bid'] = array(
+    'info' => t('Apache Solr Search: Filter by Book'),
+    'facet_field' => 'is_book_bid',
+  );

create the book facet unconditionnally, even if the book module is not enabled.

I think it's fine to set the facet_missing unconditionally, since it will have not effect absent the facet being used.

Patch with fixes for the other problems.

I'd like to find a way to use em tags or something for the missing facet - it's very non-obvious otherwise.

doxygen fix

Further improved - updated apachesolr_l() to handle $options['html']

document apachesolr_l() a little.

committed to 6.x