Operations on dictionary. More...

#include <sphinxbase/hash_table.h>
#include "s3types.h"
#include "bin_mdef.h"
#include "pocketsphinx_export.h"

Data Structures
struct	dictword_t
	a structure for one dictionary word. More...

struct	dict_t
	a structure for a dictionary. More...

Macros
#define	S3DICT_INC_SZ 4096

#define	dict_size(d) ((d)->n_word)
	Packaged macro access to dictionary members.

#define	dict_num_fillers(d) (dict_filler_end(d) - dict_filler_start(d))

#define	dict_num_real_words(d) (dict_size(d) - (dict_filler_end(d) - dict_filler_start(d)) - 2)
	Number of "real words" in the dictionary.

#define	dict_basewid(d, w) ((d)->word[w].basewid)

#define	dict_wordstr(d, w) ((w) < 0 ? NULL : (d)->word[w].word)

#define	dict_basestr(d, w) ((d)->word[dict_basewid(d,w)].word)

#define	dict_nextalt(d, w) ((d)->word[w].alt)

#define	dict_pronlen(d, w) ((d)->word[w].pronlen)

#define	dict_pron(d, w, p) ((d)->word[w].ciphone[p])
	The CI phones of the word w at position p.

#define	dict_filler_start(d) ((d)->filler_start)

#define	dict_filler_end(d) ((d)->filler_end)

#define	dict_startwid(d) ((d)->startwid)

#define	dict_finishwid(d) ((d)->finishwid)

#define	dict_silwid(d) ((d)->silwid)

#define	dict_is_single_phone(d, w) ((d)->word[w].pronlen == 1)

#define	dict_first_phone(d, w) ((d)->word[w].ciphone[0])

#define	dict_second_phone(d, w) ((d)->word[w].ciphone[1])

#define	dict_second_last_phone(d, w) ((d)->word[w].ciphone[(d)->word[w].pronlen - 2])

#define	dict_last_phone(d, w) ((d)->word[w].ciphone[(d)->word[w].pronlen - 1])

#define	S3_START_WORD "<s>"

#define	S3_FINISH_WORD "</s>"

#define	S3_SILENCE_WORD "<sil>"

#define	S3_UNKNOWN_WORD "<UNK>"

Functions
dict_t *	dict_init (cmd_ln_t config, bin_mdef_t mdef)
	Initialize a new dictionary.

int	dict_write (dict_t dict, char const filename, char const *format)
	Write dictionary to a file.

POCKETSPHINX_EXPORT s3wid_t	dict_wordid (dict_t d, const char word)
	Return word id for given word string if present.

int	dict_filler_word (dict_t *d, s3wid_t w)
	Return 1 if w is a filler word, 0 if not.

POCKETSPHINX_EXPORT int	dict_real_word (dict_t *d, s3wid_t w)
	Test if w is a "real" word, i.e.

s3wid_t	dict_add_word (dict_t d, char const word, s3cipid_t const *p, int32 np)
	Add a word with the given ciphone pronunciation list to the dictionary.

const char *	dict_ciphone_str (dict_t *d, s3wid_t wid, int32 pos)
	Return value: CI phone string for the given word, phone position.

int32	dict_word2basestr (char *word)
	If the given word contains a trailing "(....)" (i.e., a Sphinx-II style alternative pronunciation specification), strip that trailing portion from it.

dict_t *	dict_retain (dict_t *d)
	Retain a pointer to an dict_t.

int	dict_free (dict_t *d)
	Release a pointer to a dictionary.

void	dict_report (dict_t *d)
	Report a dictionary structure.

Detailed Description

Operations on dictionary.

Definition in file dict.h.

Macro Definition Documentation

◆ dict_basestr

#define dict_basestr	(	d,
		w
	)	((d)->word[dict_basewid(d,w)].word)

Definition at line 162 of file dict.h.

◆ dict_basewid

#define dict_basewid	(	d,
		w
	)	((d)->word[w].basewid)

Definition at line 160 of file dict.h.

◆ dict_filler_end

#define dict_filler_end ( d ) ((d)->filler_end)

Definition at line 167 of file dict.h.

◆ dict_filler_start

#define dict_filler_start ( d ) ((d)->filler_start)

Definition at line 166 of file dict.h.

◆ dict_finishwid

#define dict_finishwid ( d ) ((d)->finishwid)

Definition at line 169 of file dict.h.

◆ dict_first_phone

#define dict_first_phone	(	d,
		w
	)	((d)->word[w].ciphone[0])

Definition at line 172 of file dict.h.

◆ dict_is_single_phone

#define dict_is_single_phone	(	d,
		w
	)	((d)->word[w].pronlen == 1)

Definition at line 171 of file dict.h.

◆ dict_last_phone

#define dict_last_phone	(	d,
		w
	)	((d)->word[w].ciphone[(d)->word[w].pronlen - 1])

Definition at line 175 of file dict.h.

◆ dict_nextalt

#define dict_nextalt	(	d,
		w
	)	((d)->word[w].alt)

Definition at line 163 of file dict.h.

◆ dict_num_fillers

#define dict_num_fillers ( d ) (dict_filler_end(d) - dict_filler_start(d))

Definition at line 152 of file dict.h.

◆ dict_num_real_words

#define dict_num_real_words ( d ) (dict_size(d) - (dict_filler_end(d) - dict_filler_start(d)) - 2)

Number of "real words" in the dictionary.

This is the number of words that are not fillers, ~~, or~~ .

Definition at line 158 of file dict.h.

◆ dict_pron

#define dict_pron	(	d,
		w,
		p
	)	((d)->word[w].ciphone[p])

The CI phones of the word w at position p.

Definition at line 165 of file dict.h.

◆ dict_pronlen

#define dict_pronlen	(	d,
		w
	)	((d)->word[w].pronlen)

Definition at line 164 of file dict.h.

◆ dict_second_last_phone

#define dict_second_last_phone	(	d,
		w
	)	((d)->word[w].ciphone[(d)->word[w].pronlen - 2])

Definition at line 174 of file dict.h.

◆ dict_second_phone

#define dict_second_phone	(	d,
		w
	)	((d)->word[w].ciphone[1])

Definition at line 173 of file dict.h.

◆ dict_silwid

#define dict_silwid ( d ) ((d)->silwid)

Definition at line 170 of file dict.h.

◆ dict_size

#define dict_size ( d ) ((d)->n_word)

Packaged macro access to dictionary members.

Definition at line 151 of file dict.h.

◆ dict_startwid

#define dict_startwid ( d ) ((d)->startwid)

Definition at line 168 of file dict.h.

◆ dict_wordstr

#define dict_wordstr	(	d,
		w
	)	((w) < 0 ? NULL : (d)->word[w].word)

Definition at line 161 of file dict.h.

◆ S3_FINISH_WORD

#define S3_FINISH_WORD "</s>"

Definition at line 179 of file dict.h.

◆ S3_SILENCE_WORD

#define S3_SILENCE_WORD "<sil>"

Definition at line 180 of file dict.h.

◆ S3_START_WORD

#define S3_START_WORD "<s>"

Definition at line 178 of file dict.h.

◆ S3_UNKNOWN_WORD

#define S3_UNKNOWN_WORD "<UNK>"

Definition at line 181 of file dict.h.

◆ S3DICT_INC_SZ

#define S3DICT_INC_SZ 4096

Definition at line 53 of file dict.h.

Function Documentation

◆ dict_add_word()

s3wid_t dict_add_word	(	dict_t *	d,
		char const *	word,
		s3cipid_t const *	p,
		int32	np
	)

Add a word with the given ciphone pronunciation list to the dictionary.

Return value: Result word id if successful, BAD_S3WID otherwise

Parameters

d	The dictionary structure.
word	The word.
p	The pronunciation.
np	Number of phones.

Definition at line 80 of file dict.c.

References dictword_t::alt, BAD_S3WID, dictword_t::basewid, dictword_t::ciphone, dict_t::ht, dict_t::max_words, dict_t::n_word, dictword_t::pronlen, dictword_t::word, and dict_t::word.

◆ dict_ciphone_str()

const char * dict_ciphone_str	(	dict_t *	d,
		s3wid_t	wid,
		int32	pos
	)

Return value: CI phone string for the given word, phone position.

Parameters

d	In: Dictionary to look up
wid	In: Component word being looked up
pos	In: Pronunciation phone position

Definition at line 69 of file dict.c.

References dictword_t::ciphone, dict_t::mdef, and dict_t::word.

◆ dict_filler_word()

int dict_filler_word	(	dict_t *	d,
		s3wid_t	w
	)

Return 1 if w is a filler word, 0 if not.

A filler word is one that was read in from the filler dictionary; however, sentence START and FINISH words are not filler words.

Parameters

d	The dictionary structure
w	The word ID

Definition at line 413 of file dict.c.

References dict_t::filler_start, dict_t::finishwid, and dict_t::startwid.

◆ dict_free()

int dict_free ( dict_t * d )

Release a pointer to a dictionary.

Definition at line 468 of file dict.c.

References dictword_t::ciphone, dict_t::ht, dict_t::mdef, dict_t::n_word, dictword_t::word, and dict_t::word.

◆ dict_init()

dict_t * dict_init	(	cmd_ln_t *	config,
		bin_mdef_t *	mdef
	)

Initialize a new dictionary.

If config and mdef are supplied, then the dictionary will be read from the files specified by the -dict and -fdict options in config, with case sensitivity determined by the -dictcase option.

Otherwise an empty case-sensitive dictionary will be created.

Return ptr to dict_t if successful, NULL otherwise.

Parameters

config	Configuration (-dict, -fdict, -dictcase) or NULL
mdef	For looking up CI phone IDs (or NULL)

Definition at line 252 of file dict.c.

References BAD_S3WID, dict_t::filler_end, dict_t::filler_start, dict_t::finishwid, dict_t::ht, dict_t::max_words, dict_t::mdef, dict_t::n_word, dict_t::silwid, dict_t::startwid, and dict_t::word.

◆ dict_real_word()

POCKETSPHINX_EXPORT int dict_real_word	(	dict_t *	d,
		s3wid_t	w
	)

Test if w is a "real" word, i.e.

neither a filler word nor START/FINISH.

Parameters

d	The dictionary structure
w	The word ID

Definition at line 427 of file dict.c.

References dict_t::filler_start, dict_t::finishwid, and dict_t::startwid.

◆ dict_report()

void dict_report ( dict_t * d )

Report a dictionary structure.

Parameters

d	A dictionary structure

Definition at line 499 of file dict.c.

References dict_t::max_words, and dict_t::n_word.

◆ dict_retain()

dict_t * dict_retain ( dict_t * d )

Retain a pointer to an dict_t.

Definition at line 461 of file dict.c.

◆ dict_word2basestr()

int32 dict_word2basestr ( char * word )

If the given word contains a trailing "(....)" (i.e., a Sphinx-II style alternative pronunciation specification), strip that trailing portion from it.

Note that the given string is modified. Return value: If string was modified, the character position at which the original string was truncated; otherwise -1.

Definition at line 442 of file dict.c.

◆ dict_wordid()

POCKETSPHINX_EXPORT s3wid_t dict_wordid	(	dict_t *	d,
		const char *	word
	)

Return word id for given word string if present.

Otherwise return BAD_S3WID

Definition at line 399 of file dict.c.

References BAD_S3WID, and dict_t::ht.

◆ dict_write()

int dict_write	(	dict_t *	dict,
		char const *	filename,
		char const *	format
	)

Write dictionary to a file.

Definition at line 221 of file dict.c.

References dict_t::n_word.

Data Structures

Macros

Functions

Detailed Description

Macro Definition Documentation

◆ dict_basestr

◆ dict_basewid

◆ dict_filler_end

◆ dict_filler_start

◆ dict_finishwid

◆ dict_first_phone

◆ dict_is_single_phone

◆ dict_last_phone

◆ dict_nextalt

◆ dict_num_fillers

◆ dict_num_real_words

◆ dict_pron

◆ dict_pronlen

◆ dict_second_last_phone

◆ dict_second_phone

◆ dict_silwid

◆ dict_size

◆ dict_startwid

◆ dict_wordstr

◆ S3_FINISH_WORD

◆ S3_SILENCE_WORD

◆ S3_START_WORD

◆ S3_UNKNOWN_WORD

◆ S3DICT_INC_SZ

Function Documentation

◆ dict_add_word()

◆ dict_ciphone_str()

◆ dict_filler_word()

◆ dict_free()

◆ dict_init()

◆ dict_real_word()

◆ dict_report()

◆ dict_retain()

◆ dict_word2basestr()

◆ dict_wordid()

◆ dict_write()