目的:熟练掌握自上而下的语法分析方法,并能用程序实现。
要求:
1. 使用的文法如下:
E->TE'
E'->+TE'|#
T->FT'
T'->*FT'|#
F->(E)|id
2. 对于任意给定的输入串(词法记号流)进行语法分析,递归下降方法和非递归预测分析方法可以任选其一来实现。
3. 要有一定的错误处理功能。即对错误能提示,并且能在一定程度上忽略尽量少的记号来进行接下来的分析。可以参考书上介绍的同步记号集合来处理。
可能的出错情况:idid*id, id**id, (id+id, +id*+id ……
4. 输入串以#结尾,输出推导过程中使用到的产生式。例如:
输入:id+id*id#
输出:E->TE'
T->FT'
F->id
E'->+TE'
T->FT'
.....
如果输入串有错误,则在输出中要体现是跳过输入串的某些记号了,还是弹栈,弹出某个非终结符或者是终结符了,同时给出相应的出错提示信息。比如:
idid*id对应的出错信息是:“输入串跳过记号id,用户多输入了一个id”;
id**id对应的出错信息是:“弹栈,弹出非终结符F,用户少输入了一个id”
(id+id对应的出错信息是:“弹栈,弹出终结符 ) ,用户少输入了一个右括号(或者说,括号不匹配)”
套路:还是先把代码给贴出来,回头再找时间写篇文章总结一下,分析一下数据结构和算法。代码写的乱七八糟,不过基本功能实现了。
#pragma once
#include <iostream>
#include <algorithm>
#include <fstream>
#include <map>
#include <set>
#include <string>
#include <stack>
using namespace std;
/*----------------------------------------全局变量存放终结符和非终结符---------------------------------------------------*/
set<string> non_terminal; //存放非终结符
set<string> productions; //存放产生式
std::map<string, string> match_map; //存放非终结符和其对应的产生式的文法的键值对
std::map<string, set<string>> first; //string:非终结符;set<string>:非终结符所对应的first集合
std::map<string, set<string>> follow; //string:非终结符;set<string>:非终结符所对应的follow集合
bool is_not_changing = false;
/**
typedef struct _nonTerm_input //非终结符_输入符号
{
string nonTerm; //非终结符
string input; //输入符号
friend bool operator < (struct _nonTerm_input const &a, struct _nonTerm_input const &b) {
if (a.nonTerm == b.nonTerm && a.input == b.input) {
return false;
}
return true;
}
}nonTerm_input;
set<nonTerm_input> structs; //非终结符.size()*终结符.size()个struct
*/
std::map<map<string, string>, string> analyze_table; //分析表
///std::map<nonTerm_input, string> analyze_table; //分析表
stack<string> analyze_stack; //分析栈
/*------------------------------------------------------------------------------------------------------------------*/
void divide_words(string grammar, map<string, string>& match_map) {
for (int i = 0; i < (int)grammar.length(); ++i) {
if (grammar[i] == '-' && grammar[i + 1] == '>') {
/* code */
string left = grammar.substr(0, i); //一句文法的左边即非终结符
string right = grammar.substr(i + 2, grammar.length() - 1); //一句文法的右边即非终结符对应的产生式
non_terminal.insert(left); //插入非终结符集合里
productions.insert(right); //插入产生式集合里
match_map.insert(make_pair(left, right)); //将一句文法里的非终结符和其对应的产生式作为键值对插入到匹配map里
break;
}
}
}
/*将被'|'隔开的产生式拆分成对应多个的单词*/
void divide_right(string grammar_right, set<string>& small_right) {
/*或许可以用grammar.find_first_of一个一个找|,然后用substr分开子串,最后再insert到small_right中去*/
size_t found = grammar_right.find('|');
if (found != string::npos) {
int i = 0;
string temp = "\0";
while ((size_t)i < grammar_right.length()) {
if (grammar_right[i] != '|') {
temp += grammar_right[i];
i = i + 1;
}
else {
i = i + 1;
small_right.insert(temp);
temp = "\0";
}
if (i == grammar_right.length()) {
small_right.insert(temp);
temp = "\0";
}
}
}
else {
small_right.insert(grammar_right);
}
}
/*对每个非终结符non_term寻找它的非终结符集合first*/
void find_first(string non_term, set<string>& first) {
set<string> or_words; //存放产生式中被'|'隔开的单词
auto search = match_map.find(non_term);
if (search != match_map.end()) {
divide_right(search->second, or_words);
//匹配非终结符是否在or_words的开头
for (set<string>::iterator i = or_words.begin(); i != or_words.end(); i++) {
for (set<string>::iterator j = non_terminal.begin(); j != non_terminal.end(); j++) {
if ((*i).find(*j) == 0) { //在or_words[i]的开头找到了一个非终结符
//递归寻找非终结符j的first集合
find_first((*j), first);
}
else { //在or_words[i]的开头如果没有找到非终结符,即终结符
if ((*i)[0] >= 'a' && (*i)[0] <= 'z') {
first.insert(*i);
}
switch ((*i)[0]) {
case '(':
first.insert(string("("));
break;
case ')':
first.insert(string(")"));
break;
case '+':
first.insert(string("+"));
break;
case '*':
first.insert(string("*"));
break;
case '#':
first.insert(string("#"));
break;
default: //如果没有匹配到符号的话就把这个单词插入到first集合中
//first.insert(*i);
break;
}
continue; //找到之后跳出循环,避免进行多余的遍历浪费时间
}
}
}
}
}
//对非终结符的follow集进行初始化,开始符号的follow集初始化成{$},其余的初始化成空集
void initial_follow() {
for (set<string>::iterator i = non_terminal.begin(); i != non_terminal.end(); i++) {
if (i == non_terminal.begin()) {
set<string> startFollow;
startFollow.insert("$");
auto pair = make_pair(*i, startFollow);
follow.insert(pair);
}
set<string> temp_follow;
auto pair = make_pair(*i, temp_follow);
follow.insert(pair);
}
}
//判断一个非终结符的first集合中是不是含有空串#
bool first_contains_null(set<string> &first) {
auto find = first.find("#");
if (find != first.end()) {
return true;
}
return false;
}
//判断一个字符串str是否是非终结符,如果是返回true,否则返回false
bool is_non_terminal(string str) {
auto find = non_terminal.find(str);
if (find != non_terminal.end()) {
return true;
}
return false;
}
bool is_letter(char a) { //是否是小写字母
if (a >= 'a' && a <= 'z') {
return true;
}
return false;
}
bool is_cap_letter(char a) {
return (a >= 'A' && a <= 'Z') ? true : false;
}
//返回一个产生式的右部str的最后一个终结符或者非终结符
string find_last(string &str) {
if ("\0" == str) {
return "\0";
}
if ('\'' == str.at(str.size() - 1)) {
string s = str.substr(str.size() - 2, 2);
str = str.substr(0, str.size() - 2);
return s;
}
else if (is_letter( str.at(str.size() - 1) ) && is_letter( str.at( str.size() - 2) ) ) {
string s = str.substr(str.size() - 2, 2);
str = str.substr(0, str.size() - 2);
return s;
}
else {
string s = str.substr(str.size() - 1, 1);
str = str.substr(0, str.size() - 1);
return s;
}
}
int cal_follow_total_size() { //计算所有follow集合的总size
int total_size = 0;
for (map<string, set<string>>::iterator i = follow.begin(); i != follow.end(); i++) {
total_size += i->second.size();
}
return total_size;
}
void find_follow(std::map<string, set<string>>& Follow) {
while (!is_not_changing) {
int fomer_size = cal_follow_total_size();
for (std::map<string, string>::iterator i = match_map.begin(); i != match_map.end(); i++) {//对每一个产生式进行遍历
set<string> or_words;
string left = (*i).first; //左边的非终结符A
string right = (*i).second; //右边的产生式A->b1b2b3B...
divide_right(right, or_words);
for (set<string>::iterator j = or_words.begin(); j != or_words.end(); j++) {
set<string> temp = Follow.find(left)->second;
string str;
string word = *j;
for (; word != "\0"; ) {
str = find_last(word);
if (!is_non_terminal(str)) { //是终结符
temp.clear();
temp.insert(str);
}
else {
for (set<string>::iterator k = temp.begin(); k != temp.end(); k++) {
if ("#" != (*k)) {
(Follow.find(str)->second).insert(*k);
}
}
if (!first_contains_null(first.find(str)->second)) {
temp = first.find(str)->second;
}
else {
for (set<string>::iterator m = first.find(str)->second.begin(); m != first.find(str)->second.end(); m++) {
temp.insert(*m);
}
}
}
}
}
}
//判断是否发生变化
int latter_size = cal_follow_total_size();
is_not_changing = fomer_size == latter_size ? true : false;
}
}
/*
初始化预测分析表的表头,并用synch来指示从非终结符的FOLLOW集合中得到的同步记号,并将其填入表中
*/
void init_table(string input) { //input:输入串
set<string> input_symbol; //输入符号集合
for (int i = 0; i < (int)input.size(); i++) { //默认输入符号中id这种终结符最大长度为2
if (is_letter(input[i])) { //形如id的终结符
if (is_letter(input[i + 1])) {
input_symbol.insert(input.substr(i, 2));
++i;
}
else {
string m;
m = input[i];
input_symbol.insert(m);
}
}
else {
switch (input[i]) {
case '+':
input_symbol.insert("+");
break;
case '*':
input_symbol.insert("*");
break;
case '(':
input_symbol.insert("(");
break;
case ')':
input_symbol.insert(")");
break;
case '$': //结束符
input_symbol.insert("$");
break;
default:
break;
}
}
}
for (set<string>::iterator i = non_terminal.begin(); i != non_terminal.end(); i++) {
for (set<string>::iterator j = input_symbol.begin(); j != input_symbol.end(); j++) {
map<string, string> _temp_map;
auto pair = make_pair(*i, *j);
_temp_map.insert(pair);
if ((follow.find(*i))->second.find(*j) != (follow.find(*i))->second.end()) {//用ssynch来指示从非终结符的FOLLOW集合中得到的同步记号
analyze_table.insert(make_pair(_temp_map, "synch"));
}
else {
analyze_table.insert(make_pair(_temp_map, "")); //初步完成构建预测分析表的表头,analyze的第二个元素才是分析表中的元素
}
}
}
}
set<string> find_first_s(string left, string production) {
set<string> FIRST_S;
for (size_t i = 0; i < production.size(); i++) {
if (is_cap_letter(production.at(i))) {
string M;
if ('\'' == production.at(i + 1)) {
M = production.substr(i, 2);
}
else {
///M = "" + production.at(i);
M = production.at(i);
}
set<string> M_FIRST;
find_first(M, M_FIRST);
for (set<string>::iterator j = M_FIRST.begin(); j != M_FIRST.end(); j++) {
FIRST_S.insert(*j);
}
if (!first_contains_null(M_FIRST)) {
return FIRST_S;
}
}
else {
if (is_letter(production.at(i))) {
if (is_letter(production.at(i + 1))) {
FIRST_S.insert(production.substr(i, 2));
}
else {
FIRST_S.insert(production.substr(i, 1));
}
}
else {
string _temp_;
_temp_ = production.at(i);
FIRST_S.insert(_temp_);
}
return FIRST_S;
}
}
FIRST_S = follow.find(left)->second;
return FIRST_S;
}
void bulid_table() { //返回<非终结符,输入符号>位置所对应的产生式
for (set<string>::iterator nT = non_terminal.begin(); nT != non_terminal.end(); nT++) {
auto search = match_map.find(*nT);
if (search != match_map.end()) {
set<string> or_words;
divide_right(search->second, or_words);
for (set<string>::iterator _or_words_iterator = or_words.begin(); _or_words_iterator != or_words.end(); _or_words_iterator++) {
set<string> FIRST_S = find_first_s(*nT, *_or_words_iterator);
for (set<string>::iterator first_s = FIRST_S.begin(); first_s != FIRST_S.end(); first_s++) {
//对FIRST(Alpha)的每个终结符a,把A->a加入analyze_table[A,a]
if ("#" == (*first_s)) { //如果空串在FIRST(A->Alpha)中,则对FOLLOW(A)的每个终结符b(包括$),把A->Alpha加入analyze_table[A,b](包括analyze_table[M,$])
auto _find_follow = follow.find(*nT);
if (follow.end() != _find_follow) {
set<string> follow_A = _find_follow->second;
for (set<string>::iterator i = follow_A.begin(); i != follow_A.end(); i++) {
map<string, string> _temp;
auto _temp_pair = make_pair(*nT, *i);
_temp.insert(_temp_pair);
auto find_position = analyze_table.find(_temp);
if (analyze_table.end() != find_position) {
string element = *nT + "->#";
find_position->second = element;
}
}
}
}
else {
map<string, string> _temp_map;
_temp_map.insert(make_pair(*nT, *first_s));
auto find = analyze_table.find(_temp_map);
if (find != analyze_table.end()) {
string element = *nT + "->" + *_or_words_iterator;
find->second = element;
}
}
}
}
}
else {
continue;
}
}
}
/*输出analyze_table*/
void diaplay_table() {
cout << "非终结符\t输入符号\t同步记号" << endl;
for (map<map<string, string>, string>::iterator i = analyze_table.begin(); i != analyze_table.end(); i++) {
cout << i->first.begin()->first << "\t\t" << i->first.begin()->second;
cout << "\t\t" << (*i).second << "\n";
}
}
bool no_empty_in_table(string left, string right) {
map<string, string> _temp_map;
_temp_map.insert(make_pair(left, right));
auto find_element = analyze_table.find(_temp_map);
if (analyze_table.end() != find_element) {
if (find_element->second != "") {
return true;
}
else {
return false;
}
}
return false;
}
bool is_synch(string left, string right) {
map<string, string> _temp_map;
_temp_map.insert(make_pair(left, right));
auto find_element = analyze_table.find(_temp_map);
if (analyze_table.end() != find_element) {
if (find_element->second == "synch") {
return true;
}
else {
return false;
}
}
return false;
}
void error_empty(string redundant) {
cout << "出错!跳过!" << " 多输入了一个" << redundant << endl;
}
void error_synch(string synch) {
cout << "出错!弹栈!" << " 多输入了一个" << synch << endl;
}
void error_bracket() {
cout << "出错!括号不匹配!" << endl;
}
int main() {
/*读取文法文件*/
const char* filename = "wenfa.txt";
ifstream inFile(filename);
if (!inFile) {
cout << "\nFiled to open file " << filename;
return -1;
}
string st = "\0";
char buf[100];
while (!inFile.eof()) {
inFile.getline(buf, 20);
st = buf;
if (strlen(buf) == 0 || st == "end") {
break;
}
divide_words(st, match_map); //对每一行文法进行分析找出非终结符和对应的产生式
}
inFile.close();
/*遍历非终结符集合,为每个非终结符寻找first集合*/
for (set<string>::iterator i = non_terminal.begin(); i != non_terminal.end(); ++i) {
set<string> the_first; //当前非终结符的first集合
find_first(*i, the_first);
first.insert(make_pair(*i, the_first));
}
cout << "非终结符" << "\t" << "First集合" << endl;
for (map<string, set<string>>::iterator i = first.begin(); i != first.end(); i++) {
cout << "-------------------------" << endl;
cout << i->first << "\t|\t";
cout << "{ ";
//倒序输出first集合中的元素与文法中出现的顺序保持一致
for (set<string>::reverse_iterator j = (i->second).rbegin(); j != (i->second).rend(); j++) {
cout << *j << ", ";
}
cout << "\b\b }";
cout << endl;
}
cout << endl;
initial_follow();
find_follow(follow);
cout << "非终结符" << "\t" << "Follow集合" << endl;
for (map<string, set<string>>::iterator i = follow.begin(); i != follow.end(); i++) {
cout << "------------------------------" << endl;
cout << i->first << "\t|\t";
cout << "{ ";
//倒序输出first集合中的元素与文法中出现的顺序保持一致
for (set<string>::reverse_iterator j = (i->second).rbegin(); j != (i->second).rend(); j++) {
cout << *j << ", ";
}
cout << "\b\b }";
cout << endl;
}
cout << "PLEASE ENTER YOUR WORDS" << endl;
string w; //输入串
while (true) {
char temp;
cin >> temp;
if ('#' == temp) {
break;
}
w = w + temp;
}
w = w + '$'; //在最后再加上$符
cout << "WHAT YOUR INPUT STREAM IS: " << endl;
cout << w << endl;
init_table("id+id*id*(id*id)$"); //用文法中的一些符号进行初始化
bulid_table();
diaplay_table();
analyze_stack.push("$"); //压入$符
analyze_stack.push(*non_terminal.begin()); //压入文法的开始符号
int token = 0;
while (analyze_stack.size() > 0) {
string left;
string right;
left = analyze_stack.top();
if (is_letter(w.at(token))) {
if (is_letter(w.at(token + 1))) {
//形如id的
right = w.substr(token, 2);
token++;
}
else {
right = w.at(token);
}
}
else {
right = w.at(token);
}
if (left == "(" || left == ")") {
error_bracket();
break;
}
map<string, string> found_pair;
found_pair.insert(make_pair(left, right));
bool _is_empty = no_empty_in_table(left, right); //true为不空,false为空
if (analyze_table.find(found_pair) != analyze_table.end() && !_is_empty) { //如果为空的话
token++;
error_empty(right);
}
else if (is_synch(left, right)) {
analyze_stack.pop();
error_synch(right);
}
else {
bool flag = false;
//递归直到left和right相同
while (left != right) {
found_pair.insert(make_pair(left, right));
string element = analyze_table.find(found_pair)->second;
cout << element << endl;
string ele_right = "\0";
for (int i = 0; i < (int)element.length(); ++i) {
if (element[i] == '-' && element[i + 1] == '>') {
ele_right = element.substr(i + 2, element.length() - 1);
break;
}
}
if (ele_right == "#") {
analyze_stack.pop();
flag = true;
break;
}
else {
analyze_stack.pop();
while (ele_right != "\0") {
analyze_stack.push(find_last(ele_right));
}
left = analyze_stack.top();
}
found_pair.clear();
}
if (!flag) {
analyze_stack.pop();
token++;
}
found_pair.clear();
}
}
return 0;
}