Age | Commit message (Collapse) | Author |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
(default is still tight)
use --loose when compiling corpus
or tight_phrases = False in config
|
|
|
|
|
|
|
|
|
|
|
|
|
|
- use bytes instead of char*
- add some basic docstrings to functions/constructors
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
+ API naming fixes
+ Multiple feature definition files can be passed to the extractor
|
|
|
|
|
|
|
|
|
|
|
|
|
|
+ various surface fixes
|
|
|
|
+ sparse features in extractor
+ hg.intersect(string)
+ basestring = str|unicode
|
|
- TDConvert returns a string
- various c_str fixes (make copies)
- cleanup .gitignore
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|