src/corpus/NEWS.md

In corpus: Text Corpus Analysis

corpus 0.7.0

• Added ability to search for specific terms.

• Added unicode character widths.

corpus 0.6.0

• Added the ability to iterate backwards over the characters in a text.

• Added n-gram frequency counter.

• Added termset for storing a set of terms.

• Added sentfitler for suppressing sentence breaks.

corpus 0.5.0

• Added new filter module with options for dropping, combining, and selecting terms from a token stream.

• Add the ability to exempt words from stemming.

corpus 0.4.0

• Prefix functions and data types with corpus_, to avoid namespace clashes.

• Add stopwords (from the Snowball library).

• Fix Linux portability errors and warnings; everything should compile smoothly on Linux now.

• Simplify word types to None, Number, Letter, Kana, and Ideo, using the same definitions as the ICU library.

corpus 0.3.0

• Add NFC Unicode normalization; use this instead of NFD in typemap.

• Add stemming (using the Snowball library).

corpus 0.2.0

• Remove syslog in favor of custom logging in error.h interface; log to stderr by default.

• Switch to simpler filebuf.h interface.

• Add Windows portability code (tested on mingw64).

corpus 0.1.0

• First milestone, with support for JSON decoding, text segmentation, and text normalization.

corpus documentation built on May 2, 2021, 9:06 a.m.