C++ Primer学习笔记第六章tquery.cpp程序的剖析
2004-11-12 11:40
459 查看
/*
tquery.c
我用的编译器是gnu c++ 和 vc2003，为了此程序能够执行需做以下的修改：
1 tquery.c -> tquery.cpp
2 <iostream.h> -> <iostream>, <fstream.h> -> <fstream> , <stddef.h> -><cstddef>
3 增加 #include <iterator> 和 using namespace std;
4 删除 allocator 和它前面的“，”号，注意在要在>>之间留一个空格，因为编译器不是神仙，它会把>>当成操作符
5 删除250行的diff_type，对于现在的编译器，它已经过时了
6 对于执行gnu c++ 执行：g++ -o tquery.ext tquery.cpp 〔ENTER〕
7 对于vc2003 执行：cl tquery.cpp 〔ENTER〕
*/
#include <algorithm> //提供泛型算法的接口，比如copy
#include <string>
#include <vector>
#include <utility> //pair的接口<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" />
#include <map>
#include <set>
#include <iostream>
#include <fstream>
#include <cstddef> //定义了NULL，size_t等类型
#include <ctype.h> //为了大小写字母的处理
#include <iterator> //因为运用了ostream_iterator模板对象所以这是必须的，而原程序没有包含
using namespace std; //为了方便，所以显式指定名字空间是标准名字空间
typedef pair<short,short> location; //这样做完全是为了方便理解，没有太大的意义
typedef vector<location> loc;
typedef vector<string> text;
typedef pair<text*,loc*> text_loc; //完整的类型是pair< vector<string>*, vector< pair<short,short> >* >
//如果是我写的话，肯定会把location,loc,text,text_loc排在一列，李破门先生为何要这么做？我想大概是个人的喜好问
//题，这样做有它的好处，就是不会把定义相互混淆
class TextQuery {
public:
TextQuery() { memset( this, 0, sizeof( TextQuery )); } /*我们知道memset是一个C语言的函数，这句把this指向的那块内存大小为sizeof（TextQuery）个单元初值设置为0，我对这个构造函数的理解是：因为此类的成员函数操纵的是一些string，所以把内存做这样的解释是为了提高string处理的效率，因为memset虽然返回的是void*指针，但是却被编译器解释成char*指针*/
static void filter_elements( string felems ) { filt_elems = felems; }//静态成员函数说：“我是为了filt_elems而存在”
void query_text();
void display_map_text();
void display_text_locations();
void doit() {
retrieve_text();
separate_words();
filter_text();
suffix_text();
strip_caps();
build_word_map();
}
private:
void retrieve_text();
void separate_words();
void filter_text();
void strip_caps();
void suffix_text();
void suffix_s( string& );
void build_word_map();
private:
vector<string> *lines_of_text;
text_loc *text_locations;
map<string,loc*,less<string> > *word_map;
static string filt_elems; //我是属于类的，所以我属于所有对象，但我与它们一点关系也没有，即//使它们不存在，我也存在, 而且我是唯一的，不会每个对象都有一份，所以我节省空间，注意不要在类里初始化我，//为什么？去问Bjarne Stroustrup，那是俺爹
};
string TextQuery::filt_elems( "/",.;:!?)(///" );
int main()
{
TextQuery tq;
tq.doit();
tq.query_text();
tq.display_map_text();
return 0;
}
//获得文本
/*我先大体的叙述一下retrieve_text函数是做什么和怎么做的，此函数读取输入的文本文件中的每一行，一行一行地记录其内容，为什么这么说呢，你想想看: 它把输入的文本文件一行一行的读出来，看做string，用push_back放在了vector中，简言之vector中的0号元素就是text文件的第一行，以此类推。
它是如何做的呢？首先它构建了一个ifstream对象，这个对象的本身代表输入文本文件自己，接着getline登场，注意getline有两个一个是为流提供的getline，一个是在<string>中为string提供的getline，此处用的是后者。getline顾名思义，就是得到文件的一行，其实那倒也不一定，取决于getline的第三个参数，默认的参数就是“/n",“/n"意味着getline就是名副其实的get line，接下来vector的push_back粉墨登场，它把从getline获取的每一行做为string放进vector容器中，此时vector的大小会自动的加1。最终的结果就是，输入的文本被放在了一个string vector容器中，而前面的类的私有成员：vector<string> *lines_of_text就是指向这个容器的指针，抢答：(*lines_of_text)[ 0 ]代表什么？欧，是的，它代表输入文本第一行，如果你要求严格一点，它代表输入文本第一行的一个拷贝。其它的部分很简单，就此略过。*/
void
TextQuery::
retrieve_text()
{
string file_name;
cout << "please enter file name: ";
cin >> file_name;
ifstream infile( file_name.c_str(), ios::in ); //infile是一个ifstream对象,当然ifstream是一个由typedef定义的模板类，它锁定了char类型做为参数，构造函数需要一个const char*，ios::in的意义是首先ios是一个锁定char类型的流模板，而in是告诉流要为输入而打开一个文件
if ( !infile ) {
cerr << "oops! unable to open file "
<< file_name << " -- bailing out!/n";
exit( -1 );
}
else cout << "/n";
lines_of_text = new vector<string>;
string textline;
while ( getline( infile, textline, '/n' ))
lines_of_text->push_back( textline ); //把一行放入lines_of_text所指向的string vector中，vector自加1
}
//分离出单个的词并捕获其位置（？行，？列）
/*
问题1：如何分离出单个词？
答：根据词间的空格
问题2：如何迭代？
答：for循环迭代行，while循环寻找每一行的空格
*/
void
TextQuery::
separate_words()
{
vector<string> *words = new vector<string>;
vector<location> *locations = new vector<location>;
for ( short line_pos = 0; line_pos < lines_of_text->size(); line_pos++ ) /*对于书中给出的输入文本文件text_file->size()应该是6，因为string vector就是6行所以对于for循环就是for（short line_pos ＝ 0；line_pos < 6; line_pos++）,共循环6次来迭代每一行*/
{
short word_pos = 0;
string textline = (*lines_of_text)[ line_pos ];
string::size_type eol = textline.length();
string::size_type pos = 0, prev_pos = 0;
while (( pos = textline.find_first_of( ' ', pos )) != string::npos ) /*例如：对于第一行 ：
Alice Emma has long flowing red hair. Her Daddy says eol: 52 pos: 5 line: 0 word: 0 substring: Alice 意思是说，数到52,第一行结束，数到第5时出现个空格（记住Lippman先生说过要从0开始数），是第0行，此时标记出第一个词，位置是0，内容是Alice，再把Alice放进words中，words是一个string vector，想想为什么要放进vector中，因为要给单独的词一个代号比如0号，1号，2号.......*/
{
words->push_back( textline.substr( prev_pos, pos - prev_pos ));
locations->push_back( make_pair( line_pos, word_pos )); /*locations是一个pair vector，执行完此句locations就变成｛word[0] =（0, 0）......｝了*/
word_pos++; pos++; prev_pos = pos;
}
words->push_back( textline.substr( prev_pos, pos - prev_pos ));
locations->push_back( make_pair( line_pos, word_pos ));
}
text_locations = new text_loc( words, locations );
/*
text_loc完整的类型是：
pair< vector<string>*, vector< pair<short,short> >* >,
就像这样:（word[0]，(0,0)[0]）,(word[1], (0,1)[1]) 。。。。。。
*/
//滤去标点符号
void
TextQuery::
filter_text()
{
if ( filt_elems.empty() )
return;
vector<string> *words = text_locations->first;
vector<string>::iterator iter = words->begin();
vector<string>::iterator iter_end = words->end();
while ( iter != iter_end )
{
string::size_type pos = 0;
while (( pos = (*iter).find_first_of( filt_elems, pos )) != string::npos ) //find_first_of如果没有找到匹配的项会返回string::npos所以可以用string::npos做为标志
(*iter).erase(pos,1);
iter++;
}
}
//处理后缀
void
TextQuery::
suffix_text()
{
vector<string> *words = text_locations->first;
vector<string>::iterator iter = words->begin();
vector<string>::iterator iter_end = words->end();
while ( iter != iter_end )
{
// if 3 or less characters, let it be
if ( (*iter).size() <= 3 ) { iter++; continue; }
if ( (*iter)[ (*iter).size()-1 ] == 's' )
suffix_s( *iter );
// additional suffix handling goes here ...
iter++;
}
}
//处理后缀的辅助函数
void
TextQuery::
suffix_s( string &word )
{
string::size_type spos = 0;
string::size_type pos3 = word.size()-3;
// "ous", "ss", "is", "ius"
string suffixes( "oussisius" );
if ( ! word.compare( pos3, 3, suffixes, spos, 3 ) ||
! word.compare( pos3, 3, suffixes, spos+6, 3 ) ||
! word.compare( pos3+1, 2, suffixes, spos+2, 2 ) ||
! word.compare( pos3+1, 2, suffixes, spos+4, 2 ))
return;
string ies( "ies" );
if ( ! word.compare( pos3, 3, ies ))
{
word.replace( pos3, 3, 1, 'y' );
return;
}
string ses( "ses" );
if ( ! word.compare( pos3, 3, ses ))
{
word.erase( pos3+1, 2 );
return;
}
// erase ending 's'
word.erase( pos3+2 );
// watch out for "'s"
if ( word[ pos3+1 ] == '/'' )
word.erase( pos3+1 );
}
//处理大写字母
void
TextQuery::
strip_caps()
{
vector<string> *words = text_locations->first;
vector<string>::iterator iter = words->begin();
vector<string>::iterator iter_end = words->end();
string caps( "ABCDEFGHIJKLMNOPQRSTUVWXYZ" );
while ( iter != iter_end ) {
string::size_type pos = 0;
while (( pos = (*iter).find_first_of( caps, pos )) != string::npos )
(*iter)[ pos ] = tolower( (*iter)[pos] );
++iter;
}
}
void
TextQuery::
build_word_map()
{
word_map = new map< string, loc*, less<string> >;
typedef map<string,loc*,less<string> >::value_type value_type;
typedef set<string,less<string> >::difference_type diff_type;
set<string,less<string> > exclusion_set;
ifstream infile( "exclusion_set" );
if ( !infile )
{
static string default_excluded_words[25] = {
"the","and","but","that","then","are","been",
"can","can't","cannot","could","did","for",
"had","have","him","his","her","its","into",
"were","which","when","with","would"
};
cerr << "warning! unable to open word exclusion file! -- "
<< "using default set/n";
copy( default_excluded_words, default_excluded_words+25, inserter( exclusion_set, exclusion_set.begin() ));
}
else {
istream_iterator< string, diff_type > input_set( infile ), eos;
copy( input_set, eos, inserter( exclusion_set, exclusion_set.begin() ));
}
// iterate through the the words, entering the key/pair
vector<string,allocator> *text_words = text_locations->first;
vector<location,allocator> *text_locs = text_locations->second;
register int elem_cnt = text_words->size();
for ( int ix = 0; ix < elem_cnt; ++ix )
{
string textword = ( *text_words )[ ix ];
// exclusion strategies
// less than 3 character or in exclusion set
if ( textword.size() < 3 ||
exclusion_set.count( textword ))
continue;
if ( ! word_map->count((*text_words)[ix] ))
{ // not present, add it:
loc *ploc = new vector<location,allocator>;
ploc->push_back( (*text_locs)[ix] );
word_map->insert( value_type( (*text_words)[ix], ploc ));
}
else (*word_map)[(*text_words)[ix]]->push_back( (*text_locs)[ix] );
}
}
//处理用户的查询，此函数像是shell
void
TextQuery::
query_text()
{
string query_text;
do {
cout << "enter a word against which to search the text./n"
<< "to quit, enter a single character ==> ";
cin >> query_text;
if ( query_text.size() < 2 ) break;
string caps( "ABCDEFGHIJKLMNOPQRSTUVWXYZ" );
string::size_type pos = 0;
while (( pos = query_text.find_first_of( caps, pos )) != string::npos )
query_text[ pos ] = tolower( query_text[pos] );
// if we index into map, query_text is entered, if absent
// not at all what we should wish for ...
if ( !word_map->count( query_text )) {
cout << "/nSorry. There are no entries for "
<< query_text << "./n/n";
continue;
}
loc *ploc = (*word_map)[ query_text ];
set<short,less<short>,allocator> occurrence_lines;
loc::iterator liter = ploc->begin(), liter_end = ploc->end();
while ( liter != liter_end ) {
occurrence_lines.insert(occurrence_lines.end(), (*liter).first);
++liter;
}
register int size = occurrence_lines.size();
cout << "/n" << query_text
<< " occurs " << size
<< (size == 1 ? " time:" : " times:")
<< "/n/n";
set<short,less<short>,allocator>::iterator it=occurrence_lines.begin();
for ( ; it != occurrence_lines.end(); ++it ) {
int line = *it;
cout << "/t( line "
// don't confound user with text lines starting at 0 ...
<< line + 1 << " ) "
<< (*lines_of_text)[line] << endl;
}
cout << endl;
}
while ( ! query_text.empty() );
cout << "Ok, bye!/n";
}
void
TextQuery::
display_map_text()
{
typedef map<string,loc*,less<string>,allocator> map_text;
map_text::iterator iter = word_map->begin(), iter_end = word_map->end();
while ( iter != iter_end ) {
cout << "word: " << (*iter).first << " (";
int loc_cnt = 0;
loc *text_locs = (*iter).second;
loc::iterator liter = text_locs->begin(),
liter_end = text_locs->end();
while ( liter != liter_end )
{
if ( loc_cnt )
cout << ",";
else ++loc_cnt;
cout << "(" << (*liter).first
<< "," << (*liter).second << ")";
++liter;
}
cout << ")/n";
++iter;
}
cout << endl;
}
void
TextQuery::
display_text_locations()
{
vector<string,allocator> *text_words = text_locations->first;
vector<location,allocator> *text_locs = text_locations->second;
register int elem_cnt = text_words->size();
if ( elem_cnt != text_locs->size() )
{
cerr << "oops! internal error: word and position vectors "
<< "are of unequal size/n"
<< "words: " << elem_cnt << " "
<< "locs: " << text_locs->size()
<< " -- bailing out!/n";
exit( -2 );
}
for ( int ix = 0; ix < elem_cnt; ix++ )
{
cout << "word: " << (*text_words)[ ix ] << "/t"
<< "location: ("
<< (*text_locs)[ix].first << ","
<< (*text_locs)[ix].second << ")"
<< "/n";
}
cout << endl;
}
/*
sample input text:
------------------
Alice Emma has long flowing red hair. Her Daddy says
when the wind blows through her hair, it looks almost alive,
like a fiery bird in flight. A beautiful fiery bird, he tells her,
magical but untamed. "Daddy, shush, there is no such thing,"
she tells him, at the same time wanting him to tell her more.
Shyly, she asks, "I mean, Daddy, is there?"
---------------------
sample query session:
---------------------
please enter file name: alice_emma
warning! unable to open word exclusion file! -- using default set
enter a word against which to search the text.
to quit, enter a single character ==> alice
alice occurs 1 time:
( line 1 ) Alice Emma has long flowing red hair. Her Daddy says
enter a word against which to search the text.
to quit, enter a single character ==> daddy
daddy occurs 3 times:
( line 1 ) Alice Emma has long flowing red hair. Her Daddy says
( line 4 ) magical but untamed. "Daddy, shush, there is no such thing,"
( line 6 ) Shyly, she asks, "I mean, Daddy, is there?"
enter a word against which to search the text.
to quit, enter a single character ==> phoenix
Sorry. There are no entries for phoenix.
enter a word against which to search the text.
to quit, enter a single character ==> .
Ok, bye!
----------------------------------------------------------
sample text map after: (a) stripping out punctuation,
(b) eliminating semantically neutral words such as `the`,
(c) suffixing, so that fixes and fix become fix, and
(d) removal of capitalization
-----------------------------------------------------------
word: alice ((0,0))
word: alive ((1,10))
word: almost ((1,9))
word: ask ((5,2))
word: beautiful ((2,7))
word: bird ((2,3),(2,9))
word: blow ((1,3))
word: daddy ((0,8),(3,3),(5,5))
word: emma ((0,1))
word: fiery ((2,2),(2,8))
word: flight ((2,5))
word: flowing ((0,4))
word: hair ((0,6),(1,6))
word: has ((0,2))
word: like ((2,0))
word: long ((0,3))
word: look ((1,8))
word: magical ((3,0))
word: mean ((5,4))
word: more ((4,12))
word: red ((0,5))
word: same ((4,5))
word: say ((0,9))
word: she ((4,0),(5,1))
word: shush ((3,4))
word: shyly ((5,0))
word: such ((3,8))
word: tell ((2,11),(4,1),(4,10))
word: there ((3,5),(5,7))
word: thing ((3,9))
word: through ((1,4))
word: time ((4,6))
word: untamed ((3,2))
word: wanting ((4,7))
word: wind ((1,2))
*/
tquery.c
我用的编译器是gnu c++ 和 vc2003，为了此程序能够执行需做以下的修改：
1 tquery.c -> tquery.cpp
2 <iostream.h> -> <iostream>, <fstream.h> -> <fstream> , <stddef.h> -><cstddef>
3 增加 #include <iterator> 和 using namespace std;
4 删除 allocator 和它前面的“，”号，注意在要在>>之间留一个空格，因为编译器不是神仙，它会把>>当成操作符
5 删除250行的diff_type，对于现在的编译器，它已经过时了
6 对于执行gnu c++ 执行：g++ -o tquery.ext tquery.cpp 〔ENTER〕
7 对于vc2003 执行：cl tquery.cpp 〔ENTER〕
*/
#include <algorithm> //提供泛型算法的接口，比如copy
#include <string>
#include <vector>
#include <utility> //pair的接口<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" />
#include <map>
#include <set>
#include <iostream>
#include <fstream>
#include <cstddef> //定义了NULL，size_t等类型
#include <ctype.h> //为了大小写字母的处理
#include <iterator> //因为运用了ostream_iterator模板对象所以这是必须的，而原程序没有包含
using namespace std; //为了方便，所以显式指定名字空间是标准名字空间
typedef pair<short,short> location; //这样做完全是为了方便理解，没有太大的意义
typedef vector<location> loc;
typedef vector<string> text;
typedef pair<text*,loc*> text_loc; //完整的类型是pair< vector<string>*, vector< pair<short,short> >* >
//如果是我写的话，肯定会把location,loc,text,text_loc排在一列，李破门先生为何要这么做？我想大概是个人的喜好问
//题，这样做有它的好处，就是不会把定义相互混淆
class TextQuery {
public:
TextQuery() { memset( this, 0, sizeof( TextQuery )); } /*我们知道memset是一个C语言的函数，这句把this指向的那块内存大小为sizeof（TextQuery）个单元初值设置为0，我对这个构造函数的理解是：因为此类的成员函数操纵的是一些string，所以把内存做这样的解释是为了提高string处理的效率，因为memset虽然返回的是void*指针，但是却被编译器解释成char*指针*/
static void filter_elements( string felems ) { filt_elems = felems; }//静态成员函数说：“我是为了filt_elems而存在”
void query_text();
void display_map_text();
void display_text_locations();
void doit() {
retrieve_text();
separate_words();
filter_text();
suffix_text();
strip_caps();
build_word_map();
}
private:
void retrieve_text();
void separate_words();
void filter_text();
void strip_caps();
void suffix_text();
void suffix_s( string& );
void build_word_map();
private:
vector<string> *lines_of_text;
text_loc *text_locations;
map<string,loc*,less<string> > *word_map;
static string filt_elems; //我是属于类的，所以我属于所有对象，但我与它们一点关系也没有，即//使它们不存在，我也存在, 而且我是唯一的，不会每个对象都有一份，所以我节省空间，注意不要在类里初始化我，//为什么？去问Bjarne Stroustrup，那是俺爹
};
string TextQuery::filt_elems( "/",.;:!?)(///" );
int main()
{
TextQuery tq;
tq.doit();
tq.query_text();
tq.display_map_text();
return 0;
}
//获得文本
/*我先大体的叙述一下retrieve_text函数是做什么和怎么做的，此函数读取输入的文本文件中的每一行，一行一行地记录其内容，为什么这么说呢，你想想看: 它把输入的文本文件一行一行的读出来，看做string，用push_back放在了vector中，简言之vector中的0号元素就是text文件的第一行，以此类推。
它是如何做的呢？首先它构建了一个ifstream对象，这个对象的本身代表输入文本文件自己，接着getline登场，注意getline有两个一个是为流提供的getline，一个是在<string>中为string提供的getline，此处用的是后者。getline顾名思义，就是得到文件的一行，其实那倒也不一定，取决于getline的第三个参数，默认的参数就是“/n",“/n"意味着getline就是名副其实的get line，接下来vector的push_back粉墨登场，它把从getline获取的每一行做为string放进vector容器中，此时vector的大小会自动的加1。最终的结果就是，输入的文本被放在了一个string vector容器中，而前面的类的私有成员：vector<string> *lines_of_text就是指向这个容器的指针，抢答：(*lines_of_text)[ 0 ]代表什么？欧，是的，它代表输入文本第一行，如果你要求严格一点，它代表输入文本第一行的一个拷贝。其它的部分很简单，就此略过。*/
void
TextQuery::
retrieve_text()
{
string file_name;
cout << "please enter file name: ";
cin >> file_name;
ifstream infile( file_name.c_str(), ios::in ); //infile是一个ifstream对象,当然ifstream是一个由typedef定义的模板类，它锁定了char类型做为参数，构造函数需要一个const char*，ios::in的意义是首先ios是一个锁定char类型的流模板，而in是告诉流要为输入而打开一个文件
if ( !infile ) {
cerr << "oops! unable to open file "
<< file_name << " -- bailing out!/n";
exit( -1 );
}
else cout << "/n";
lines_of_text = new vector<string>;
string textline;
while ( getline( infile, textline, '/n' ))
lines_of_text->push_back( textline ); //把一行放入lines_of_text所指向的string vector中，vector自加1
}
//分离出单个的词并捕获其位置（？行，？列）
/*
问题1：如何分离出单个词？
答：根据词间的空格
问题2：如何迭代？
答：for循环迭代行，while循环寻找每一行的空格
*/
void
TextQuery::
separate_words()
{
vector<string> *words = new vector<string>;
vector<location> *locations = new vector<location>;
for ( short line_pos = 0; line_pos < lines_of_text->size(); line_pos++ ) /*对于书中给出的输入文本文件text_file->size()应该是6，因为string vector就是6行所以对于for循环就是for（short line_pos ＝ 0；line_pos < 6; line_pos++）,共循环6次来迭代每一行*/
{
short word_pos = 0;
string textline = (*lines_of_text)[ line_pos ];
string::size_type eol = textline.length();
string::size_type pos = 0, prev_pos = 0;
while (( pos = textline.find_first_of( ' ', pos )) != string::npos ) /*例如：对于第一行 ：
Alice Emma has long flowing red hair. Her Daddy says eol: 52 pos: 5 line: 0 word: 0 substring: Alice 意思是说，数到52,第一行结束，数到第5时出现个空格（记住Lippman先生说过要从0开始数），是第0行，此时标记出第一个词，位置是0，内容是Alice，再把Alice放进words中，words是一个string vector，想想为什么要放进vector中，因为要给单独的词一个代号比如0号，1号，2号.......*/
{
words->push_back( textline.substr( prev_pos, pos - prev_pos ));
locations->push_back( make_pair( line_pos, word_pos )); /*locations是一个pair vector，执行完此句locations就变成｛word[0] =（0, 0）......｝了*/
word_pos++; pos++; prev_pos = pos;
}
words->push_back( textline.substr( prev_pos, pos - prev_pos ));
locations->push_back( make_pair( line_pos, word_pos ));
}
text_locations = new text_loc( words, locations );
/*
text_loc完整的类型是：
pair< vector<string>*, vector< pair<short,short> >* >,
就像这样:（word[0]，(0,0)[0]）,(word[1], (0,1)[1]) 。。。。。。
*/
//滤去标点符号
void
TextQuery::
filter_text()
{
if ( filt_elems.empty() )
return;
vector<string> *words = text_locations->first;
vector<string>::iterator iter = words->begin();
vector<string>::iterator iter_end = words->end();
while ( iter != iter_end )
{
string::size_type pos = 0;
while (( pos = (*iter).find_first_of( filt_elems, pos )) != string::npos ) //find_first_of如果没有找到匹配的项会返回string::npos所以可以用string::npos做为标志
(*iter).erase(pos,1);
iter++;
}
}
//处理后缀
void
TextQuery::
suffix_text()
{
vector<string> *words = text_locations->first;
vector<string>::iterator iter = words->begin();
vector<string>::iterator iter_end = words->end();
while ( iter != iter_end )
{
// if 3 or less characters, let it be
if ( (*iter).size() <= 3 ) { iter++; continue; }
if ( (*iter)[ (*iter).size()-1 ] == 's' )
suffix_s( *iter );
// additional suffix handling goes here ...
iter++;
}
}
//处理后缀的辅助函数
void
TextQuery::
suffix_s( string &word )
{
string::size_type spos = 0;
string::size_type pos3 = word.size()-3;
// "ous", "ss", "is", "ius"
string suffixes( "oussisius" );
if ( ! word.compare( pos3, 3, suffixes, spos, 3 ) ||
! word.compare( pos3, 3, suffixes, spos+6, 3 ) ||
! word.compare( pos3+1, 2, suffixes, spos+2, 2 ) ||
! word.compare( pos3+1, 2, suffixes, spos+4, 2 ))
return;
string ies( "ies" );
if ( ! word.compare( pos3, 3, ies ))
{
word.replace( pos3, 3, 1, 'y' );
return;
}
string ses( "ses" );
if ( ! word.compare( pos3, 3, ses ))
{
word.erase( pos3+1, 2 );
return;
}
// erase ending 's'
word.erase( pos3+2 );
// watch out for "'s"
if ( word[ pos3+1 ] == '/'' )
word.erase( pos3+1 );
}
//处理大写字母
void
TextQuery::
strip_caps()
{
vector<string> *words = text_locations->first;
vector<string>::iterator iter = words->begin();
vector<string>::iterator iter_end = words->end();
string caps( "ABCDEFGHIJKLMNOPQRSTUVWXYZ" );
while ( iter != iter_end ) {
string::size_type pos = 0;
while (( pos = (*iter).find_first_of( caps, pos )) != string::npos )
(*iter)[ pos ] = tolower( (*iter)[pos] );
++iter;
}
}
void
TextQuery::
build_word_map()
{
word_map = new map< string, loc*, less<string> >;
typedef map<string,loc*,less<string> >::value_type value_type;
typedef set<string,less<string> >::difference_type diff_type;
set<string,less<string> > exclusion_set;
ifstream infile( "exclusion_set" );
if ( !infile )
{
static string default_excluded_words[25] = {
"the","and","but","that","then","are","been",
"can","can't","cannot","could","did","for",
"had","have","him","his","her","its","into",
"were","which","when","with","would"
};
cerr << "warning! unable to open word exclusion file! -- "
<< "using default set/n";
copy( default_excluded_words, default_excluded_words+25, inserter( exclusion_set, exclusion_set.begin() ));
}
else {
istream_iterator< string, diff_type > input_set( infile ), eos;
copy( input_set, eos, inserter( exclusion_set, exclusion_set.begin() ));
}
// iterate through the the words, entering the key/pair
vector<string,allocator> *text_words = text_locations->first;
vector<location,allocator> *text_locs = text_locations->second;
register int elem_cnt = text_words->size();
for ( int ix = 0; ix < elem_cnt; ++ix )
{
string textword = ( *text_words )[ ix ];
// exclusion strategies
// less than 3 character or in exclusion set
if ( textword.size() < 3 ||
exclusion_set.count( textword ))
continue;
if ( ! word_map->count((*text_words)[ix] ))
{ // not present, add it:
loc *ploc = new vector<location,allocator>;
ploc->push_back( (*text_locs)[ix] );
word_map->insert( value_type( (*text_words)[ix], ploc ));
}
else (*word_map)[(*text_words)[ix]]->push_back( (*text_locs)[ix] );
}
}
//处理用户的查询，此函数像是shell
void
TextQuery::
query_text()
{
string query_text;
do {
cout << "enter a word against which to search the text./n"
<< "to quit, enter a single character ==> ";
cin >> query_text;
if ( query_text.size() < 2 ) break;
string caps( "ABCDEFGHIJKLMNOPQRSTUVWXYZ" );
string::size_type pos = 0;
while (( pos = query_text.find_first_of( caps, pos )) != string::npos )
query_text[ pos ] = tolower( query_text[pos] );
// if we index into map, query_text is entered, if absent
// not at all what we should wish for ...
if ( !word_map->count( query_text )) {
cout << "/nSorry. There are no entries for "
<< query_text << "./n/n";
continue;
}
loc *ploc = (*word_map)[ query_text ];
set<short,less<short>,allocator> occurrence_lines;
loc::iterator liter = ploc->begin(), liter_end = ploc->end();
while ( liter != liter_end ) {
occurrence_lines.insert(occurrence_lines.end(), (*liter).first);
++liter;
}
register int size = occurrence_lines.size();
cout << "/n" << query_text
<< " occurs " << size
<< (size == 1 ? " time:" : " times:")
<< "/n/n";
set<short,less<short>,allocator>::iterator it=occurrence_lines.begin();
for ( ; it != occurrence_lines.end(); ++it ) {
int line = *it;
cout << "/t( line "
// don't confound user with text lines starting at 0 ...
<< line + 1 << " ) "
<< (*lines_of_text)[line] << endl;
}
cout << endl;
}
while ( ! query_text.empty() );
cout << "Ok, bye!/n";
}
void
TextQuery::
display_map_text()
{
typedef map<string,loc*,less<string>,allocator> map_text;
map_text::iterator iter = word_map->begin(), iter_end = word_map->end();
while ( iter != iter_end ) {
cout << "word: " << (*iter).first << " (";
int loc_cnt = 0;
loc *text_locs = (*iter).second;
loc::iterator liter = text_locs->begin(),
liter_end = text_locs->end();
while ( liter != liter_end )
{
if ( loc_cnt )
cout << ",";
else ++loc_cnt;
cout << "(" << (*liter).first
<< "," << (*liter).second << ")";
++liter;
}
cout << ")/n";
++iter;
}
cout << endl;
}
void
TextQuery::
display_text_locations()
{
vector<string,allocator> *text_words = text_locations->first;
vector<location,allocator> *text_locs = text_locations->second;
register int elem_cnt = text_words->size();
if ( elem_cnt != text_locs->size() )
{
cerr << "oops! internal error: word and position vectors "
<< "are of unequal size/n"
<< "words: " << elem_cnt << " "
<< "locs: " << text_locs->size()
<< " -- bailing out!/n";
exit( -2 );
}
for ( int ix = 0; ix < elem_cnt; ix++ )
{
cout << "word: " << (*text_words)[ ix ] << "/t"
<< "location: ("
<< (*text_locs)[ix].first << ","
<< (*text_locs)[ix].second << ")"
<< "/n";
}
cout << endl;
}
/*
sample input text:
------------------
Alice Emma has long flowing red hair. Her Daddy says
when the wind blows through her hair, it looks almost alive,
like a fiery bird in flight. A beautiful fiery bird, he tells her,
magical but untamed. "Daddy, shush, there is no such thing,"
she tells him, at the same time wanting him to tell her more.
Shyly, she asks, "I mean, Daddy, is there?"
---------------------
sample query session:
---------------------
please enter file name: alice_emma
warning! unable to open word exclusion file! -- using default set
enter a word against which to search the text.
to quit, enter a single character ==> alice
alice occurs 1 time:
( line 1 ) Alice Emma has long flowing red hair. Her Daddy says
enter a word against which to search the text.
to quit, enter a single character ==> daddy
daddy occurs 3 times:
( line 1 ) Alice Emma has long flowing red hair. Her Daddy says
( line 4 ) magical but untamed. "Daddy, shush, there is no such thing,"
( line 6 ) Shyly, she asks, "I mean, Daddy, is there?"
enter a word against which to search the text.
to quit, enter a single character ==> phoenix
Sorry. There are no entries for phoenix.
enter a word against which to search the text.
to quit, enter a single character ==> .
Ok, bye!
----------------------------------------------------------
sample text map after: (a) stripping out punctuation,
(b) eliminating semantically neutral words such as `the`,
(c) suffixing, so that fixes and fix become fix, and
(d) removal of capitalization
-----------------------------------------------------------
word: alice ((0,0))
word: alive ((1,10))
word: almost ((1,9))
word: ask ((5,2))
word: beautiful ((2,7))
word: bird ((2,3),(2,9))
word: blow ((1,3))
word: daddy ((0,8),(3,3),(5,5))
word: emma ((0,1))
word: fiery ((2,2),(2,8))
word: flight ((2,5))
word: flowing ((0,4))
word: hair ((0,6),(1,6))
word: has ((0,2))
word: like ((2,0))
word: long ((0,3))
word: look ((1,8))
word: magical ((3,0))
word: mean ((5,4))
word: more ((4,12))
word: red ((0,5))
word: same ((4,5))
word: say ((0,9))
word: she ((4,0),(5,1))
word: shush ((3,4))
word: shyly ((5,0))
word: such ((3,8))
word: tell ((2,11),(4,1),(4,10))
word: there ((3,5),(5,7))
word: thing ((3,9))
word: through ((1,4))
word: time ((4,6))
word: untamed ((3,2))
word: wanting ((4,7))
word: wind ((1,2))
*/
内容来自用户分享和网络整理，不保证内容的准确性，如有侵权内容，可联系管理员处理 
相关文章推荐
- C++算法系列之日历生成的算法代码
- Vector 还是ArrayList
- Java测试规范(引用)
- Java面试题目
- delphi中的时间操作技术（2）
- 解析IP地址为主机域名
- 将某一主机域名解析为IP地址
- 输入输出流为什么要用缓冲区？
- 引用一个人的说法看中国软件需求基础
- 有关TrueDBGrid的问题向各位大虾请教
- 力争最简单把*、&说清楚
- 使用C#开发COM+组件
- Decal SDL-Delphi的范型类库-通用数据结构与算法类库（一）
- thinking in c++ 卷2
- thinking in c++卷2
- thinkng in c++卷2
- Thinking in C++ 卷2中文版
- 初学者的问题----csdn系列