Skip to content
Objective-C full-text search engine.
C++ C Objective-C Objective-C++ Other
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
BRFullTextSearch Mac OS
BRFullTextSearch.xcodeproj
BRFullTextSearch.xcworkspace
BRFullTextSearch
BRFullTextSearchTests
Pods
SampleCocoaPodsProject
SampleOSXCocoaPodsProject
.gitignore
.gitmodules
BRFullTextSearch.podspec
CHANGELOG.md
LICENSE.txt
NOTICE.txt
Podfile
Podfile.lock
README.md

README.md

BRFullTextSearch

Objective-C full text search engine.

This project provides a way to integrate full-text search capabilities into your iOS or OS X project. First, it provides a protocol-based API for a simple text indexing and search framework. Second, it provides a CLucene based implementation of that framework.

Example Usage

The following snippet shows how the API works. The CLuceneSearchService reference is the only CLucene-specific portion of the code:

id<BRSearchService> service = [[CLuceneSearchService alloc] initWithIndexPath:@"/some/path"];

// add a document to the index
id<BRIndexable> doc = [[BRSimpleIndexable alloc] initWithIdentifier:@"1" data:@{
					   kBRSearchFieldNameTitle : @"Special document",
					   kBRSearchFieldNameValue : @"This is a long winded note with really important details in it."
					   }];
NSError *error = nil;
[service addObjectToIndexAndWait:doc error:&error];

// search for documents and log contents of each result
id<BRSearchResults> results = [service search:@"special"];
[results iterateWithBlock:^(NSUInteger index, id<BRSearchResult>result, BOOL *stop) {
	NSLog(@"Found result: %@", [result dictionaryRepresentation]);
}];

Sample projects

There are some sample projects included in the source distribution:

Predicate queries

The BRSearchService API supports NSPredicate based queries:

- (id<BRSearchResults>)searchWithPredicate:(NSPredicate *)predicate
                                    sortBy:(NSString *)sortFieldName
                                  sortType:(BRSearchSortType)sortType
                                 ascending:(BOOL)ascending;

This method of querying can be quite useful when constructing a query out of user-supplied query text. For example, you could support prefix based queries (for example, searching for ca* to match cat):

// get query as string, from text field for Example
NSString * query = ...;

static NSExpression *ValueExpression;
if ( ValueExpression == nil ) {
    ValueExpression = [NSExpression expressionForKeyPath:kBRSearchFieldNameValue];
}
NSArray *tokens = [[query stringByTrimmingCharactersInSet:[NSCharacterSet whitespaceAndNewlineCharacterSet]]
                   componentsSeparatedByCharactersInSet:[NSCharacterSet whitespaceAndNewlineCharacterSet]];
NSMutableArray *predicates = [NSMutableArray arrayWithCapacity:([tokens count] * 2)];
for ( NSString *token in tokens ) {
    [predicates addObject:[NSComparisonPredicate predicateWithLeftExpression:ValueExpression
                                                             rightExpression:[NSExpression expressionForConstantValue:token]
                                                                    modifier:NSDirectPredicateModifier
                                                                        type:NSLikePredicateOperatorType
                                                                     options:0]];
    [predicates addObject:[NSComparisonPredicate predicateWithLeftExpression:ValueExpression
                                                             rightExpression:[NSExpression expressionForConstantValue:token]
                                                                    modifier:NSDirectPredicateModifier
                                                                        type:NSBeginsWithPredicateOperatorType
                                                                     options:0]];
}
NSPredicate *predicateQuery = [NSCompoundPredicate orPredicateWithSubpredicates:predicates];
searchResults = [searchService searchWithPredicate:predicateQuery sortBy:nil sortType:0 ascending:NO];

Batch API

When indexing many documents at once, BRSearchService provides a set of methods specifically designed for efficient bulk operations:

// bulk block callback function.
typedef void (^BRSearchServiceIndexUpdateBlock)(id <BRIndexUpdateContext> updateContext);

// perform a bulk operation, calling the passed on block
- (void)bulkUpdateIndex:(BRSearchServiceIndexUpdateBlock)updateBlock
                  queue:(dispatch_queue_t)finishedQueue
               finished:(BRSearchServiceUpdateCallbackBlock)finished;

// from within the block, the following methods can be used (notice the updateContext parameter):

- (void)addObjectToIndex:(id <BRIndexable> )object context:(id <BRIndexUpdateContext> )updateContext;

- (int)removeObjectFromIndex:(BRSearchObjectType)type withIdentifier:(NSString *)identifier
                     context:(id <BRIndexUpdateContext> )updateContext;

- (int)removeObjectsFromIndexMatchingPredicate:(NSPredicate *)predicate
                                       context:(id <BRIndexUpdateContext> )updateContext;

- (int)removeAllObjectsFromIndex:(id <BRIndexUpdateContext> )updateContext;

Here's an example of a bulk operation that adds 100,000 documents to the index; notice the strategic use of @autoreleasepool to keep a lid on memory use during the operation:

id<BRSearchService> service = ...;
[service bulkUpdateIndex:^(id<BRIndexUpdateContext> updateContext) {

    if ( [updateContext respondsToSelector:@selector(setOptimizeWhenDone:)] ) {
        updateContext.optimizeWhenDone = YES;
    }

    // add a bunch of documents to the index, in small autorelease batches
    for ( int i = 0; i < 100000; i+= 1000 ) {
        @autoreleasepool {
            for ( int j = 0; j < 1000; j++ ) {
                id<BRIndexable> doc = ...;
                [service addObjectToIndex:doc context:updateContext];
            }
        }
    }

} queue:dispatch_get_main_queue() finished:^(int updateCount, NSError *error) {
    // all finished here
}];

Core Data support

It's pretty easy to integrate BRFullTextSearch with Core Data, to maintain a search index while changes are persisted in Core Data. One way is to listen for the NSManagedObjectContextDidSaveNotification notification and process Core Data changes as index delete and update operations. The SampleCocoaPodsProject iOS project contains an example of this integration. The app allows you to create small sticky notes and search the text of those notes. See the CoreDataManager class in the sample project, whose maintainSearchIndexFromManagedObjectDidSave: method handles this.

The SampleOSXCocoaPodsProject OS X project also contains an example of this integration. See the CoreDataManager class in that project for details.

Project Integration

You can integrate BRFullTextSearch via CocoaPods, or manually as either a dependent project or static framework.

via CocoaPods

Install CocoaPods if not already available:

$ [sudo] gem install cocoapods
$ pod setup

Change to the directory of your Xcode project, and create a file named Podfile with contents similar to this:

platform :ios, '5.0'
pod 'BRFullTextSearch', '~> 2.0'

Install into your project:

$ pod install

Open your project in Xcode using the .xcworkspace file CocoaPods generated.

Note: the use_frameworks! option is not supported, see #4. Any pull requests to allow for building as a dynamic framework are very welcome!

via static framework

Using this approach you'll build a static library framework that you can manually integrate into your own project.

The BRFullTextSearch Xcode project includes a target called BRFullTextSearch.framework that builds a static library framework. Build that target, which will produce a Framework/Release/BRFullTextSearch.framework bundle at the root project directory. Copy that framework into your project and add it as a build dependency.

You must also add the following linker build dependencies, which you can do by clicking the + button in the Link Binary With Libraries section of the Build Phases tab in the project settings:

  • libz
  • libc++

Next, add -ObjC as an Other Linker Flags build setting. If you do not have any C++ sources in your project, you probably also need to add -stdlib=libc++ to this setting as well.

Finally, you'll need to add the path to the directory containing the BRFullTextSearch.framework bundle as a Framework Search Paths value in the Build Settings tab of the project settings.

The SampleStaticLibraryProject included in this repository is an example project set up using the static library framework integration approach. You must build BRFullTextSearch.framework first, then open this project. When you run the project, it will index a set of documents using some Latin text. You can then search for latin words using a simple UI.

You can’t perform that action at this time.