-
Construct Ferret Token directly.
0.1.6 / 2008-03-16¶ ↑
-
Optimize for simple algorithm. One time faster than before. And less memory usage.
0.1.5 / 2008-03-03¶ ↑
-
Bug fix: Ferret Token is not Duck-Typing. We need to construct Ferret token instead of reuse
RMMSeg
Token.
0.1.4 / 2008-03-02¶ ↑
-
Let user store their customized word to Dictionary after loaded.
-
Improved performance of SimpleAlgorithm.
0.1.3 / 2008-02-28¶ ↑
-
Make
RMMSeg
Token campatible to Ferret Token. -
Use while instead of loop for performance improvement.
-
Avoid many costly String#jlength call for performance improvement (use only 70% time and 40% memory as before).
0.1.2 / 2008-02-25¶ ↑
-
Add cache to find_match_words: performance improved.
-
Implement Chunk as a module instead of a class: performance improved.
-
Don’t store unnecessary data in dictionary: memory usage reduced.
0.1.1 / 2008-02-05¶ ↑
-
Use bones instead of hoe as building tool.
0.1.0 / 2008-02-01¶ ↑
-
Add filter to filter out Chinese punctuations.
0.0.1 / 2008-01-31¶ ↑
-
Analyzer integration with Ferret.
-
rdoc added
-
Lazily init the
Word
objects inside theDictionary
. -
Handle English punctuation correctly.