Shout_lm2bin Class Reference

With this class (main class of the application shout_lm2bin) an ARPA language model can be stored in a LanguageModel object and written to disk. More...

Inheritance diagram for Shout_lm2bin:

[legend]

List of all members.


Public Member Functions
	Shout_lm2bin (char lexName, char lmName, char *binLmName)
	~Shout_lm2bin ()
Protected Member Functions
int	getNumberOfValidEntries (char word1, char word2, FILE *file)
void	sortWords (int numberOfItems, int wordList, int sortList)
Protected Attributes
MemMappedFile *	memMap
int	numberOfWords
char **	vocabulary
StringLookup *	wordTree

Detailed Description

With this class (main class of the application shout_lm2bin) an ARPA language model can be stored in a LanguageModel object and written to disk.

Constructor & Destructor Documentation

Shout_lm2bin::Shout_lm2bin	(	char *	lexName,
		char *	lmName,
		char *	binLmName
	)

The main function of the shout_lm2bin application will call this constructor. All work is done by this constructor.

The lexical tree is needed to obtain the wordID for each word in the ARPA model. Looking up the IDs is done by just calling the getWordID() method from the LexicalTree object.

After loading the lexical tree, the ARPA model is loaded and the uni-, bi- and tri-gram probabilities and backoff values are loaded and translated into LanguageModel format.

Finally, the binary language model is stored to file in the format so that it can be loaded by LanguageModel.

References LMEntryType_1::backoff, LanguageModel::bi_Hash, LanguageModel::bi_lmData, LanguageModel::bi_tableLength, TrainHash::fillMapping(), TrainHash::finalizeHash(), LanguageModel::four_Hash, LanguageModel::four_lmData, LanguageModel::four_tableLength, WriteFileLittleBigEndian::freadEndianSafe(), WriteFileLittleBigEndian::fwriteEndianSafe(), TrainHash::getIndex(), StringLookup::getWordID(), TrainHash::initialiseMapping(), memMap, numberOfWords, LMEntryType_1::p, StringFunctions::RightTrim(), MemMappedFile::setWritePermission(), StringFunctions::splitList(), TrainHash::storeHash(), LanguageModel::tri_Hash, LanguageModel::tri_lmData, LanguageModel::tri_tableLength, LanguageModel::uni_lmData, LanguageModel::uni_tableLength, vocabulary, and wordTree.