bit::ArpaReader Class Reference

Reads ARPA language model file. More...

#include <ArpaReader.hh>

List of all members.

Public Member Functions
	ArpaReader (FILE *file=NULL)
	Construct a reader associated with a possible file.
void	reset (FILE *file)
	Reset the structure and prepare reading from file.
void	read_header ()
	Read the header of the ARPA file.
bool	read_ngram ()
	Read next ngram from the ARPA file.
bool	read_order_ngrams (bool sort=false)
	Read all ngrams of the next order.
Public Attributes
bit::ArpaReader::Options	opt
	Configurable options.
bit::ArpaReader::Header	header
	Information in the ARPA header.
bit::ArpaReader::Ngram	ngram
	The ngram read at the last call of read_ngram().
std::vector< Ngram >	order_ngrams
	Ngrams read with read_order_ngrams().
std::vector< int >	sorted_order
	Indices of sorted ngrams.
SymbolMap< std::string, int >	symbol_map
	The symbols in the model.
bool	end_reached
	Have we reached the \ keyword.
Private Attributes
FILE *	m_file
	The file to read from.
int	m_current_order
	The order ngram just read from the file.
int	m_ngrams_read
	The number of ngrams read for current order.
Classes
struct	Header
	Information in the ARPA header. More...
struct	Ngram
	The ngram read at the last call of read_ngram(). More...
struct	Options
	Configurable options. More...

---

Detailed Description

Reads ARPA language model file.

The following events are treated as error:

• An n-gram contains a symbol that is not introduced in 1-grams

The following events are accepted silently:

• An n-gram has non-zero backoff, but no continuation
• Duplicate n-grams

---

Constructor & Destructor Documentation

bit::ArpaReader::ArpaReader (  FILE *  file = NULL  )

Construct a reader associated with a possible file.

---

Member Function Documentation

void bit::ArpaReader::read_header (   )

Read the header of the ARPA file.

bool bit::ArpaReader::read_ngram (   )

Read next ngram from the ARPA file. Returns: false if no more ngrams on this order. Next call will start reading next order. See end_reached to check end of file.

bool bit::ArpaReader::read_order_ngrams (  bool  sort = false  )

Read all ngrams of the next order. Returns: false if could not read more ngrams

void bit::ArpaReader::reset (  FILE *  file  )

Reset the structure and prepare reading from file.

---

Member Data Documentation

bool bit::ArpaReader::end_reached

Have we reached the \ keyword.

struct bit::ArpaReader::Header bit::ArpaReader::header

Information in the ARPA header.

int bit::ArpaReader::m_current_order [private]

The order ngram just read from the file.

FILE* bit::ArpaReader::m_file [private]

The file to read from.

int bit::ArpaReader::m_ngrams_read [private]

The number of ngrams read for current order.

struct bit::ArpaReader::Ngram bit::ArpaReader::ngram

The ngram read at the last call of read_ngram().

struct bit::ArpaReader::Options bit::ArpaReader::opt

Configurable options.

std::vector<Ngram> bit::ArpaReader::order_ngrams

Ngrams read with read_order_ngrams().

std::vector<int> bit::ArpaReader::sorted_order

Indices of sorted ngrams.

SymbolMap<std::string, int> bit::ArpaReader::symbol_map

The symbols in the model.

---

The documentation for this class was generated from the following files:

• lm/ArpaReader.hh
• lm/ArpaReader.cc

---