From 851e389dffdd6996ea32d70defb8906de80b9edc Mon Sep 17 00:00:00 2001 From: Chris Dyer Date: Mon, 14 Dec 2009 20:35:11 -0500 Subject: few small fixes of alignment tools, add new orthographic similarity feature for word aligner, final naming of directories, libraries in cdec --- src/csplit.h | 30 ------------------------------ 1 file changed, 30 deletions(-) delete mode 100644 src/csplit.h (limited to 'src/csplit.h') diff --git a/src/csplit.h b/src/csplit.h deleted file mode 100644 index ce6295c1..00000000 --- a/src/csplit.h +++ /dev/null @@ -1,30 +0,0 @@ -#ifndef _CSPLIT_H_ -#define _CSPLIT_H_ - -#include "translator.h" -#include "lattice.h" - -// this "translator" takes single words (with NO SPACES) and segments -// them using the approach described in: -// -// C. Dyer. (2009) Using a maximum entropy model to build segmentation -// lattices for MT. In Proceedings of NAACL HLT 2009. -// note, an extra word space marker # is inserted at the left edge of -// the forest! -struct CompoundSplitImpl; -struct CompoundSplit : public Translator { - CompoundSplit(const boost::program_options::variables_map& conf); - bool Translate(const std::string& input, - SentenceMetadata* smeta, - const std::vector& weights, - Hypergraph* forest); - - // given a forest generated by CompoundSplit::Translate, - // find the edge representing the unsegmented form - static int GetFullWordEdgeIndex(const Hypergraph& forest); - - private: - boost::shared_ptr pimpl_; -}; - -#endif -- cgit v1.2.3