stoptree.cc

一个增量文本聚类的算法。 参考文献： Wai-chiu Wong, Ada Wai-chee Fu, Incremental Document Clustering for Web Page Cl

#include <stdio.h>
#include <stdlib.h>
#include <ctype.h>
#include <string.h>
#include <iostream.h>
#include "StopTree.h"
#include "Word.h"

StopTree::StopTree()
{
    FILE *in;
    char str[MaxWordLen];

    stopRoot = NULL;
    if((in=fopen("../MyLib/stoplist","r")) == NULL){
        cerr << "Cannot open the stoplist file\n";
        exit(1);
    }

    while(!feof(in)){
	fscanf(in,"%s",&str);
	insert(&stopRoot,str);
    }

//    printTree(stopRoot);
    fclose(in);
}

StopTree::~StopTree()
{
}

void StopTree::insert(Word **root,char *str)
{
    if((* root) == NULL){
    	(* root) = new Word(str);
    }else{
	int s;
	s = strcmp((* root)->nameList->name,str);
	if(s > 0){
	    insert(&((* root)->left),str);
	}else if(s < 0){
	    insert(&((* root)->right),str);
	}
    }
}

void StopTree::printTree(Word *root)
{
    if(root->left != NULL){
	printTree(root->left);
    }

    printf("%s\n",root->nameList->name);

    if(root->right != NULL){
	printTree(root->right);
    }
}

int StopTree::inStopList(Word *root,char *str)
{
    int l=0;

    if(root == NULL){
        l = 0;
    }else if(strcmp(root->nameList->name,str) > 0){
        l = inStopList(root->left,str);
    }else if(strcmp(root->nameList->name,str) < 0){
        l = inStopList(root->right,str);
    }else{
	l = 1;
    }

    return l;
}