phonetisaurus-align.cc File Reference

#include <include/M2MFstAligner.h>
#include <include/LatticePruner.h>
#include <include/util.h>
#include <include/PhonetisaurusRex.h>

Go to the source code of this file.

Functions
int	load_input_file (M2MFstAligner *aligner, string input_file, string delim, string s1_char_delim, string s2_char_delim, bool init=false)
void	write_alignments (M2MFstAligner *aligner, string ofile_name, StdArc::Weight threshold, int nbest, bool fb, bool penalize)
void	compileNBestFarArchive (M2MFstAligner *aligner, vector< VectorFst< LogArc > > *fsts, string far_name, StdArc::Weight threshold, int nbest, bool fb, bool penalize)
	DEFINE_bool (seq1_del, true,"Allow deletions in sequence one.")
	DEFINE_bool (seq2_del, true,"Allow deletions in sequence two.")
	DEFINE_bool (penalize, true,"Penalize scores.")
	DEFINE_bool (penalize_em, false,"Penalize links during EM training.")
	DEFINE_bool (load_model, false,"Load a pre-trained model for use.")
	DEFINE_bool (lattice, false,"Write out the alignment lattices as an fst archive (.far).")
	DEFINE_bool (restrict, true,"Restrict links to M-1, 1-N during initialization.")
	DEFINE_bool (mbr, false,"Use the LMBR decoder (not yet implemented).")
	DEFINE_bool (fb, false,"Use forward-backward pruning for the alignment lattices.")
	DEFINE_int32 (seq1_max, 2,"Maximum subsequence length for sequence one.")
	DEFINE_int32 (seq2_max, 2,"Maximum subsequence length for sequence two.")
	DEFINE_int32 (iter, 11,"Maximum number of EM iterations to perform.")
	DEFINE_int32 (nbest, 1,"Output the N-best alignments given the model.")
	DEFINE_string (input,"","Two-column input file to align.")
	DEFINE_string (seq1_sep,"|","Multi-token separator for input tokens.")
	DEFINE_string (seq2_sep,"|","Multi-token separator for output tokens.")
	DEFINE_string (s1s2_sep,"}","Token used to separate input-output subsequences in the g2p model.")
	DEFINE_string (delim,"\t","Delimiter separating entry one and entry two in the input file.")
	DEFINE_string (eps,"<eps>","Epsilon symbol.")
	DEFINE_string (skip,"_","Skip token used to represent null transitions. Distinct from epsilon.")
	DEFINE_string (ofile,"","Output file to write the aligned dictionary to.")
	DEFINE_string (s1_char_delim,"","Sequence one input delimeter.")
	DEFINE_string (s2_char_delim," ","Sequence two input delimeter.")
	DEFINE_string (model_file,"","FST-format alignment model to load.")
	DEFINE_string (write_model,"","Write out the alignment model in OpenFst format to filename.")
	DEFINE_double (thresh, 1e-10,"Delta threshold for EM training termination.")
	DEFINE_double (pthresh,-99,"Pruning threshold. Use to prune unlikely N-best candidates when using multiple alignments.")
int	main (int argc, char *argv[])

Function Documentation

void compileNBestFarArchive	(	M2MFstAligner *	aligner,
		vector< VectorFst< LogArc > > *	fsts,
		string	far_name,
		StdArc::Weight	threshold,
		int	nbest,
		bool	fb,
		bool	penalize
	)

Definition at line 146 of file phonetisaurus-align.cc.

DEFINE_bool	(	seq1_del	,
		true	,
		"Allow deletions in sequence one."
	)

DEFINE_bool	(	seq2_del	,
		true	,
		"Allow deletions in sequence two."
	)

DEFINE_bool	(	penalize	,
		true	,
		"Penalize scores."
	)

DEFINE_bool	(	penalize_em	,
		false	,
		"Penalize links during EM training."
	)

DEFINE_bool	(	load_model	,
		false	,
		"Load a pre-trained model for use."
	)

DEFINE_bool	(	lattice	,
		false	,
		"Write out the alignment lattices as an fst archive (.far)."
	)

DEFINE_bool	(	restrict	,
		true	,
		"Restrict links to M-	1,
		1-N during initialization."
	)

DEFINE_bool	(	mbr	,
		false	,
		"Use the LMBR decoder (not yet implemented)."
	)

DEFINE_bool	(	fb	,
		false	,
		"Use forward-backward pruning for the alignment lattices."
	)

DEFINE_double	(	thresh	,
		1e-	10,
		"Delta threshold for EM training termination."
	)

DEFINE_double	(	pthresh	,
		-	99,
		"Pruning threshold. Use to prune unlikely N-best candidates when using multiple alignments."
	)

DEFINE_int32	(	seq1_max	,
		2	,
		"Maximum subsequence length for sequence one."
	)

DEFINE_int32	(	seq2_max	,
		2	,
		"Maximum subsequence length for sequence two."
	)

DEFINE_int32	(	iter	,
		11	,
		"Maximum number of EM iterations to perform."
	)

DEFINE_int32	(	nbest	,
		1	,
		"Output the N-best alignments given the model."
	)

DEFINE_string	(	input	,
		""	,
		"Two-column input file to align."
	)

DEFINE_string	(	seq1_sep	,
		"|"	,
		"Multi-token separator for input tokens."
	)

DEFINE_string	(	seq2_sep	,
		"|"	,
		"Multi-token separator for output tokens."
	)

DEFINE_string	(	s1s2_sep	,
		"}"	,
		"Token used to separate input-output subsequences in the g2p model."
	)

DEFINE_string	(	delim	,
		"\t"	,
		"Delimiter separating entry one and entry two in the input file."
	)

DEFINE_string	(	eps	,
		"<eps>"	,
		"Epsilon symbol."
	)

DEFINE_string	(	skip	,
		"_"	,
		"Skip token used to represent null transitions. Distinct from epsilon."
	)

DEFINE_string	(	ofile	,
		""	,
		"Output file to write the aligned dictionary to."
	)

DEFINE_string	(	s1_char_delim	,
		""	,
		"Sequence one input delimeter."
	)

DEFINE_string	(	s2_char_delim	,
		" "	,
		"Sequence two input delimeter."
	)

DEFINE_string	(	model_file	,
		""	,
		"FST-format alignment model to load."
	)

DEFINE_string	(	write_model	,
		""	,
		"Write out the alignment model in OpenFst format to filename."
	)

int load_input_file	(	M2MFstAligner *	aligner,
		string	input_file,
		string	delim,
		string	s1_char_delim,
		string	s2_char_delim,
		bool	init = false
	)

Definition at line 39 of file phonetisaurus-align.cc.

int main	(	int	argc,
		char *	argv[]
	)

Definition at line 263 of file phonetisaurus-align.cc.

void write_alignments	(	M2MFstAligner *	aligner,
		string	ofile_name,
		StdArc::Weight	threshold,
		int	nbest,
		bool	fb,
		bool	penalize
	)

Definition at line 75 of file phonetisaurus-align.cc.