The 3.4.0 release introduces a native C# / .NET API to Carrot2
as well as a number of smaller improvements.
|Carrot2 C# / .NET API|
Starting with release 3.4.0, Carrot2 provides a native C# API for calling clustering
from C# / .NET software:
A similar C# API is available for Lingo3G,
a real-time text clustering engine from Carrot Search.
- All required DLL files, C# code examples and API reference included
- No need to install Java JRE/SDK
- Clustering engine feature set identical to Java version
- Clustering results identical to Java version
Clustering performance comparable to Java version in case of the STC algorithm.
Due to highly computational nature of Lingo, its C# version is slower than the Java counterpart.
- Works with .NET Framework 3.5 or later
Java 1.6.0 required
As of version 3.4.0, Carrot2 stops supporting Java 1.5, which
reached its End of Service Life in October 2009.
Lucene API removed
Versions prior to 3.4.0 depended on
Lucene 3.x API, which could cause integration problems in systems
relying on earlier versions of Lucene. Version 3.4.0 completely removes the
dependency on Lucene API and as a result, Carrot2 will not interfere
with earlier and future releases of Lucene.
Bing document source
MSN Live document source has been replaced with Bing.
OpenSearch document source in Workbench
The OpenSearch document source is now available in Document Clustering Workbench.
A number of dependencies have been updated:
- Mahout Math and Collections to version 0.3
- High Performance Primitive Collections to version 0.3.1
- Google Collections replaced with Guava
- SimpleXML to version 2.3.5
- Jackson JSON to version 1.5.2
For a complete list of improvements and bug fixes, see JIRA
issues fixed in version 3.4.0. For more new features, please see release 3.3.0 and release 3.2.0 notes.