Google’s enterprise search technology is becoming an increasingly popular choice for IT managers to manage their intranets and pool data from multiple sources.
It provides:
- Excellent keyword searching (obviously) based on pagerank as well as customizable weighting factors
- Recommended links
- Source and Date Biasing
- Good support for different char sets
- Incredible speed and performance
- Support for multiple collections and frontends
- An easy REST interface for API integration
Leveraging Google’s technology on your CMS (Drupal) gives us the following benefits:
- Potential for more relevant results
- Integration with 3rd party databases and sites
- Advanced search features like synonyms, stemming and language detection
How the module works
See the Google Appliance for Drupal module page for more details on implementation.
First thing you need to do is to setup the module so it knows where to connect to:

At the minimum, you need:
- Search name (this will appear on a new tab on the search screen).
- Host Name ( the URL or IP where your GSA or Mini is located )
- Collection (which collection you wish to search ).
- Client ( This doesn’t matter much, it just has to be valid. This is equivilent to the “frontend” in the GSA ).
Okay, done? If you want to, enable caching (which will cache results so you don’t need to re-query for the same search within the timeout period) and set the debug level.
Now, you’ll need to tell the mini where to crawl. For this, just go to your GSA administration screen, and punch in the url of your site. For node pages, the module will add meta-tags for the following:
- Taxonomy (Advanced search filter coming soon!)
- Date Modified and Created (Date sorting coming soon!)
- Author
- Status (pub/unpub)
- Language (if using i18n)
After installing the module, you will see a new tab on the search screen:

Fire off a search and see your results (drupalified).

In addition, you can enable the recommended links block, and if you have configured key matches in drupal, they will show up in this block.
This module is still in Beta, so expect some issues. Here are a few I know of:
- No meta tags on non-node pages, means they will be found, but won’t have the type / author, etc fields in the results
- Does not use url() on incoming links, which means if the mini finds node/123 pages, they won’t get aliases
There are probably lots more, but hopefully, this will get people who are interested in this going and we can work on making it better.
I am available for GSA and mini consulting and custom integrations, just contact me