Major Update to Full Text Search (OCR)

One of the most popular features in GovOS Cloud Search is Full Text Search.  This is the process where we run each document through our proprietary OCR engine.  (OCR stands for Optical Character Recognition.)  OCR converts the document images to text.  Then we load all that text into our search engine.  Using this feature you can search every word of every document.

It's not 100% accurate but it's pretty close.  And boy does it help find documents that you can't find any other way!  Because we all know that only a limited amount of data is indexed and is therefore searchable.  Full Text Search makes the entire document searchable.

We've been reviewing the searches you have been doing and we've been listening to your feedback.  We've taken everything we've learned to makes some really exciting improvements that will make Cloud Search even easier to use and much better and finding the results that are important to you.

Here are the changes we've made:

Phrase Search is the Default

If you are familiar with the advanced features available in Cloud Search, you know that two words entered in the search box perform an "OR" search.  So DAVID JOHNSON is converted to a search for DAVID OR JOHNSON.  We return all documents that contain EITHER DAVID or JOHNSON - as single words.  For example it would return both DAVID SMITH and a document with JOE JOHNSON.

But what if you actually want only David Johnson's documents?  For that we provide PHRASE SEARCH IDENTIFIERS: [double quotes].  So if you enter "DAVID JOHNSON" with the double quotes, you won't find DAVID SMITH.  You will only find DAVID JOHNSON.

Previously in order to get a Phrase Search, you had to use the double quotes.  We're changing Full Text Search to ALWAYS perform a Phrase Search, even without the double quotes.  We believe this will provide a better result.

Note that you still have full control.  You can still use quotes for advanced searches.  If you enter DAVID JOHNSON we'll make it a phrase search.  But if that's not what you are wanting, you can still be explicit by type DAVID AND JOHNSON or you can type DAVID OR JOHNSON.

This seems to more closely match industry standards, such as searching for text in Adobe Acrobat.

Hit Highlights in Card View

If you perform a Full Text Search, you'll see your result list with any matches that are found in the indexes.  If you switch to the Card View, you'll see a snippet of the document text with the matches highlighted.

Hit Highlights in Image Viewer

Once you click on the thumbnail and open the Image Viewer, we'll skip to the first match found in that document and we'll highlight the matching text.

Hit Highlights with Navigation Controls in Image Viewer

We're also providing Navigation Controls in the Image Viewer when the document contains more than one match.

With it you'll be able to move quickly through the document, seeing each match that exists.

Minutes Department Defaults to Full Text Search 

For the Minutes Department (or CCM Dept) Quick Search will automatically default to Full Text Search instead of Index Search. 

For this department, this option will provide the best results.  For other departments, the default will continue to be Search Index Only.  But you can select Full Text if desired.

More Improvements to Come

There are other changes we're making behind the scenes to improve the relevance of your searches.  And we'll continue to listen to your feedback for additional improvements to Cloud Search.