/home/emre/Desktop/nlp/hw01/main.cpp File Reference

Gebze Institute of Tech. CSE 484 Introduction to NLP HW01 Gram tests on Turkish word morphology and generating words from parsed words using foma and TRmorph. To use program you have to download and install foma. More...

#include <cstdio>
#include <iostream>
#include <cstdlib>
#include <vector>
#include <string>
#include <sstream>
#include <queue>
#include <iomanip>
#include <fstream>
#include "ngram.hpp"

Functions
void	initWords (char *fileName, vector< string > &vect)
bool	writeGeneratorFileScript (char *fileName, string str)
bool	editWordFile (char *fileName)
void	printLoading (int loading)
void	runScriptForEveryWord (vector< string > &vect, string argv2)
void	ngramParserFromFile (NGram ngram[])
void	wordGenerator (NGram ngram[], string fst, int oneG, int twoG, int threeG)
void	printResults (NGram ngram[], string fileName)
void	printNSize (NGram ngram[], int n)
int	main (int argc, char *argv[])

---

Detailed Description

Gebze Institute of Tech. CSE 484 Introduction to NLP HW01 Gram tests on Turkish word morphology and generating words from parsed words using foma and TRmorph. To use program you have to download and install foma.

Main Side of Program.

Author:

Emre Sercan Aslan <emre.aslan@e-aslan.net>

Date:

March 24, 2013

Version:

v1.0

See also:

References :

[1] Mans Hulden, Finite-State Compiler and C Library http://code.google.com/p/foma/

[2] Çağrı Çöltekin, TRmorph: A Turkish Morphological Analyzer https://github.com/coltekin/TRmorph

---

Function Documentation

bool editWordFile

(

char *

fileName

)

Opens and edits created word file. It is used for editing generated words' list.

Parameters:

fileName

Name of source file

Returns:

Returns true if process done correctly, otherwise false

void initWords	(	char *	fileName,
		vector< string > &	vect
	)

Reads words from text to initialize word list

Parameters:

filename	Name of source file
vect	Word list vector

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

Main routine of program.

*use case

*read file get words

*parse words

*analyse parsed parts

*generate words

void ngramParserFromFile

(

NGram

ngram[]

)

Analyses morphologies of words according to 1-Gram, 2-Gram and 3-Gram.

Parameters:

ngram

NGram instances to get 1-Gram, 2-Gram, 3-Gram results

See also:

runScriptForEveryWord()

void printLoading

(

int

loading

)

Prints loading state to terminal.

Parameters:

loading

Percent of jobs done

void printNSize	(	NGram	ngram[],
		int	n
	)

Prints just n number of results for per gram test.

Parameters:

ngram	NGram instances 1-Gram, 2-Gram, 3-Gram
n	Number of POS Tag will be printed to terminal for per gram

void printResults	(	NGram	ngram[],
		string	fileName
	)

Writes whole analyse result to file.

Parameters:

ngram	NGram instances 1-Gram, 2-Gram, 3-Gram
fileName	Name of file where results will be kept

void runScriptForEveryWord	(	vector< string > &	vect,
		string	argv2
	)

Runs flookup script for per word in word list. Writes parsing result to a file which is used in ngramParserFromFile function.

Parameters:

vect	Word list vector
argv2	File name like "trmorph.fst"

See also:

ngramParserFromFile() function

void wordGenerator	(	NGram	ngram[],
		string	fst,
		int	oneG,
		int	twoG,
		int	threeG
	)

Generates word generator codes and writes them to file to generate words. Function generates words using that codes. Runs foma script to generate words. Edits generated word file to delete inappropriate datas.

Parameters:

ngram	NGram instances 1-Gram, 2-Gram, 3-Gram
fst	File name like "trmorph.fst"
oneG	While generating word, use top n number POS Tags in 1Gram
twoG	While generating word, use top n number POS Tags in 2Gram
threeG	While generating word, use top n number POS Tags in 3Gram

bool writeGeneratorFileScript	(	char *	fileName,
		string	str
	)

Writes string to file. Writes root + tags to generate new words

Parameters:

fileName	Name of output file
str	Content that will be written to output file

Returns:

Returns true if process done correctly, otherwise false