Drupal.org Names Achieve Lead Solr Developer

Achieve Internet has been leading Solr development for the Drupal.org redesign for a while, and we were officially named lead Solr developer last week.
 
One general goal for the new site is to efficiently deliver useful and relevant content to the user. Technically, this involves several important tasks:

1. Defining how content will appear on a given search page
2. Making sure content is tagged appropriately
3. Indexing the content properly in Solr
4. Writing flexible code that can generate dynamic and complex queries

As far as the appearance of content on the page goes, the general layout for search will be similar to the existing site but a search block will allow users to add facets and sorting to their terms. The results will display below the search block, and the right side of the page will feature related content. The diagram to the right demonstrates the general layout, with examples of what the right-hand blocks might display.

 
Items 2-4 above enable more robust faceted searching. For just one example, on the current Drupal.org, we can filter themes by Drupal version. The new site will allow us to filter on Drupal version, screen resolution, and number of columns.
 
One additional trick is that we’re using checkboxes so users can do “or” searches—for example, for themes with 2 OR 3 columns. The consequence of this is that we can’t just add the tid that corresponds to “column” to the Solr query. Instead, we have to create a new id for the field and pass it with all possible values.
 
This week, a few members of the redesign team (including Achieve’s CTO Bill O’Connor) will be meeting and sprinting in Portland to move the effort forward. Some issues we’ll be discussing this week include:

• Views integration– query customization for Solr has required that we do a fair amount of work outside the views module (e.g. the various blocks on the right-hand side of the Download & Extend search area). The d.o. team will draw some lines in the sand for where it makes sense to work outside views and when it makes sense to work inside.
• Multi-site searching capability– can we use Solr to index all content on the various Drupal sites (association, groups sites, etc.)?
• Attachments – should documents be indexed on Solr for searching?

 
For these sorts of issues, we’ll try to determine how much time and effort would be required to complete them, which ones we can include in the launch, and how we can plan to tackle the others down the road.
 
More updates to come when Bill returns.