Wordクエリ2

#pragma once
#include <map>
#include <string>
#include <Windows.h>
using namespace std;

typedef map<string, string>	MAP_WORD;

class CPraseWord
{
public:
	CPraseWord(void);
	~CPraseWord(void);

public:
	BOOL ParseWord(const char* lpszFile);
	BOOL GetWordInfo(const char* lpszWord, string& strInfo);

private:
	MAP_WORD* GetMap(const char ch);
	BOOL AddToMap(const char* lpszStr, const int nLine, const int nWord);

private:
	MAP_WORD	m_mapWord[27];
};


#include "stdafx.h"
#include "PraseWord.h"
#include "macro.h"

#define MAX_WORD_LEN 100

CPraseWord::CPraseWord(void)
{
}

CPraseWord::~CPraseWord(void)
{
}

BOOL CPraseWord::ParseWord(const char* lpszFile)
{
	ASSERT_RETURN(lpszFile, FALSE);

	FILE* fp = fopen(lpszFile, "r");

	int nWord = 1;
	int nLine = 1;
	char ch = '\0';
	BOOL bLastSp = FALSE;
	int nCharPos = 0;

	char szWord[MAX_WORD_LEN] = {0};

	while ( EOF != (ch = fgetc(fp)) )
	{
		switch (ch)
		{
		case '
':
			nLine++;
			nWord = 1;
			bLastSp = FALSE;
			nCharPos = 0;

			AddToMap(szWord, nLine, nWord);

			break;

		case ' ':
		case '\r':
		case ',':
		case ':':
		case '#':
		case '!':
		case '+':
		case '-':
		case '*':
		case '&':
		case '(':
		case ')':
		case '@':
		case '$':
			bLastSp = TRUE;
			nCharPos = 0;

			AddToMap(szWord, nLine, nWord);
			
			break;

		default:
            if (bLastSp)
			{
				nWord++;
				nCharPos = 0;
				memset(szWord, 0, sizeof(szWord));
			}

			szWord[nCharPos++] = ch;

			bLastSp = FALSE;
			
			break;
		}		
	}

	if (NULL != fp)
	{
		fclose(fp);
		fp = NULL;
	}	

	return TRUE;
}

BOOL CPraseWord::AddToMap(const char* lpszStr, const int nLine, const int nWord)
{
	ASSERT_RETURN(lpszStr, FALSE);

	MAP_WORD* pMap = GetMap(lpszStr[0]);	

	if (NULL == pMap)
	{
		return FALSE;
	}

	char szPlace[100] = {0};
	::_snprintf(szPlace, sizeof(szPlace), "{%d, %d}", nLine, nWord);

	MAP_WORD::iterator it = pMap->find(lpszStr);

	if (pMap->end() == it)
	{
		(*pMap)[lpszStr] = szPlace;
	}
	else
	{
		it->second += ",";
		it->second += szPlace;
	}

	return TRUE;
}

BOOL CPraseWord::GetWordInfo(const char* lpszWord, string& strInfo)
{
	ASSERT_RETURN(lpszWord, FALSE);

	MAP_WORD* pMap = GetMap(lpszWord[0]);

	if (NULL == pMap)
	{
		return FALSE;
	}

	MAP_WORD::iterator it = pMap->find(lpszWord);

	if (pMap->end() == it)
	{
		return FALSE;
	}

	strInfo = it->second;
	return TRUE;
}

MAP_WORD* CPraseWord::GetMap(const char ch)
{
	int nSub = ch - 'a';

	if ( (nSub >= 0) && (nSub < 26) )
	{
		return &m_mapWord[nSub];
	}

	nSub = ch - 'A';

	if ( (nSub >= 0) && (nSub < 26) )
	{
		return &m_mapWord[nSub];
	}

	if (0 == ch)
	{
		return NULL;
	}

	return &m_mapWord[26];
}



// w_m.cpp : Defines the entry point for the console application.
//
#include "stdafx.h"
#include "PraseWord.h"
#include <iostream>
#include <string>
#include <Windows.h>
#include <time.h>
using namespace std;

size_t time_stat(bool bBegin = true);

int main(int argc, char* argv[])
{
	CPraseWord pd;
	char* lpszFile = "./kjv.txt";

	if (argc >= 2)
	{
		lpszFile = argv[1];
	}

	//DWORD dwStart = GetTickCount();

	time_stat(true);
	if (!pd.ParseWord(lpszFile))
	{
		cout << "prase file failed..." << endl;
		return -1;
	}
	
	cout << "parase file cost time:" << time_stat(false) << endl;
	

	//DWORD dwEnd = GetTickCount();
	//cout << "parase file cost time:" << dwEnd - dwStart << endl;

	while (TRUE)
	{
		char szWord[100] = {0};
		cout << "please input find parse word:";
		cin >> szWord;

		string str;
		//dwStart = GetTickCount();
		time_stat(true);

		pd.GetWordInfo(szWord, str);
		//cout << "search cost time:" << time_stat(false) << endl;
		size_t nCostTime = time_stat(false);
		
		//dwEnd = GetTickCount();
		//cout << "search cost time:" << dwEnd - dwStart << endl;

		if (str.empty())
		{
			cout << "can't find " << szWord << endl;
		}
		else
		{
			cout << str << endl;
		}
		
		cout << endl << endl;
		cout << "search cost time:" << nCostTime << endl;
		cout << "search total cost time:" << time_stat(false) << endl;
		//dwEnd = GetTickCount();
		//cout << "search total cost time:" << dwEnd - dwStart << endl;
	}

	return 0;
}

int gettimeofday(struct timeval *tv, void *tzp)
{
	time_t clock;
	struct tm tm;
	SYSTEMTIME wtm;

	GetLocalTime(&wtm);
	tm.tm_year = wtm.wYear - 1900;
	tm.tm_mon  = wtm.wMonth - 1;
	tm.tm_mday = wtm.wDay;
	tm.tm_hour = wtm.wHour;
	tm.tm_min  = wtm.wMinute;
	tm.tm_sec  = wtm.wSecond;
	tm.tm_isdst= -1;
	clock = mktime(&tm);
	tv->tv_sec = (long)clock;
	tv->tv_usec = wtm.wMilliseconds * 1000;

	return 0;
}


size_t time_stat(bool bBegin)
{
	static struct timeval st_time_begin, st_time_end;
	static char szInfo[128];

	if (bBegin)
	{
		gettimeofday(&st_time_begin, NULL);
		return 0;
	}
	else
	{
		gettimeofday(&st_time_end, NULL);
		return (st_time_end.tv_sec - st_time_begin.tv_sec) * 1000
			+ (st_time_end.tv_usec - st_time_begin.tv_usec) / 1000;
	}
}