I am trying to parse a line from a text file which of the form:
[int_:] [int_/int_] [(int_, string)] string [string:int_]...
Where [] are optional parameter but will contain tags such as (":", "(", ")", "/").
Also the last format is repeat format "key:value" combination. e.g.:
10: 0x1/2 (8, INC) rd API:2 SI:100
I am able to parse the whole line when all the parameter are available. But if any of the starting optional parameter is missing then the parser fails.
How can I ignore the optional parameters in the Boost Spirit library? (i.e. skip the assignment of optional variables to default values.)
These are the qi grammar rules:
quoted_string = lexeme[+(char_ -(lit(' ') | lit(')')))];
hex_num = ((lit("0x") | lit("0X")) >> hex) | uint_;
start = (hex_num >> lit(":"))
>> (hex_num >> lit("/") >> hex_num )
>> lit("(") >> hex_num >> lit(",") >> quoted_string >> lit(")")
>> quoted_string
>> quoted_string;
qi::rule<Iterator, std::string(), ascii::space_type> quoted_string;
qi::rule<Iterator, uint32_t(), ascii::space_type> hex_num;
qi::rule<Iterator, employee(), ascii::space_type> start;
Model your AST node to reflect the Parser tree:
struct ratio_t { uint32_t a,b; };
struct opcode_t { uint32_t id; std::string name; };
struct Node {
uint32_t label; // prefix:
boost::optional<ratio_t> ratio; // a/b
boost::optional<opcode_t> opcode; // (id, name)
std::string extra;
std::multimap<std::string, uint32_t> params;
};
(Just making stuff up as I go, because I can only guess what the data means. I'm assuming employee, hex_num and quoted_string are somehow remnants from sample code you started with).
Now when you adapt these structures:
BOOST_FUSION_ADAPT_STRUCT(AST::ratio_t, a, b)
BOOST_FUSION_ADAPT_STRUCT(AST::opcode_t, id, name)
BOOST_FUSION_ADAPT_STRUCT(AST::Node, label, ratio, opcode, extra, params)
You can simply parse into it with an analogous parse tree:
// lexemes
unquoted_string = +(graph - ')');
num = (no_case[ "0x" ] >> hex) | uint_;
param = +(graph - ':') >> ':' >> num;
// skipping productions
opcode = '(' >> num >> ',' >> unquoted_string >> ')';
ratio = num >> '/' >> num;
prefix = (num >> ':') | attr(0); // defaults to 0
start = prefix
>> -ratio
>> -opcode
>> unquoted_string
>> *param;
Now when you parse these test cases:
for (std::string const input : {
"10: 0x1/2 (8, INC) rd API:2 SI:100",
"10: 0x1/2 (8, INC) rd API:2",
"10: 0x1/2 (8, INC) rd",
"10: 0x1/2 rd API:2 SI:100",
"10: rd API:2 SI:100",
"0x1/2 rd API:2 SI:100",
"rd API:2 SI:100",
})
{
It f = input.begin(), l = input.end();
AST::Node data;
bool ok = qi::phrase_parse(f, l, p, qi::ascii::space, data);
if (ok) {
std::cout << "Parse success: " << data << "\n";
}
else {
std::cout << "Parse failure ('" << input << "')\n";
}
if (f!=l) {
std::cout << "Remaining unparsed input: '" << std::string(f,l) << "'\n";
}
}
you get:
Parse success: 10: 1/2 (8, 'INC') rd API:2 SI:100
Parse success: 10: 1/2 (8, 'INC') rd API:2
Parse success: 10: 1/2 (8, 'INC') rd
Parse success: 10: 1/2 -- rd API:2 SI:100
Parse success: 10: -- -- rd API:2 SI:100
Parse success: 0: 1/2 -- rd API:2 SI:100
Parse success: 0: -- -- rd API:2 SI:100
#include <boost/spirit/include/qi.hpp>
#include <boost/fusion/adapted.hpp>
#include <boost/optional/optional_io.hpp>
#include <map>
namespace qi = boost::spirit::qi;
namespace AST {
struct ratio_t { uint32_t a,b; };
struct opcode_t { uint32_t id; std::string name; };
struct Node {
uint32_t label; // prefix:
boost::optional<ratio_t> ratio; // a/b
boost::optional<opcode_t> opcode; // (id, name)
std::string extra;
std::multimap<std::string, uint32_t> params;
};
std::ostream& operator<<(std::ostream& os, ratio_t const& v) {
return os << v.a << "/" << v.b;
}
std::ostream& operator<<(std::ostream& os, opcode_t const& v) {
return os << "(" << v.id << ", '" << v.name << "')";
}
std::ostream& operator<<(std::ostream& os, Node const& v) {
os << v.label << ": " << v.ratio << " " << v.opcode << " " << v.extra;
for (auto& p : v.params) os << " " << p.first << ":" << p.second;
return os;
}
}
BOOST_FUSION_ADAPT_STRUCT(AST::ratio_t, a, b)
BOOST_FUSION_ADAPT_STRUCT(AST::opcode_t, id, name)
BOOST_FUSION_ADAPT_STRUCT(AST::Node, label, ratio, opcode, extra, params)
template <typename It, typename Skipper = qi::ascii::space_type>
struct P : qi::grammar<It, AST::Node(), Skipper> {
P() : P::base_type(start)
{
using namespace qi;
// lexemes
unquoted_string = +(graph - ')');
num = (no_case[ "0x" ] >> hex) | uint_;
param = +(graph - ':') >> ':' >> num;
// skipping productions
opcode = '(' >> num >> ',' >> unquoted_string >> ')';
ratio = num >> '/' >> num;
prefix = (num >> ':') | attr(0); // defaults to 0
start = prefix
>> -ratio
>> -opcode
>> unquoted_string
>> *param;
BOOST_SPIRIT_DEBUG_NODES((start)(unquoted_string)(num)(prefix)(ratio)(opcode)(param))
}
private:
qi::rule<It, AST::ratio_t(), Skipper> ratio;
qi::rule<It, AST::opcode_t(), Skipper> opcode;
qi::rule<It, AST::Node(), Skipper> start;
qi::rule<It, uint32_t(), Skipper> prefix;
//lexemes
qi::rule<It, std::string()> unquoted_string;
qi::rule<It, uint32_t()> num;
qi::rule<It, std::pair<std::string, uint32_t>> param;
};
int main() {
using It = std::string::const_iterator;
P<It> const p;
for (std::string const input : {
"10: 0x1/2 (8, INC) rd API:2 SI:100",
"10: 0x1/2 (8, INC) rd API:2",
"10: 0x1/2 (8, INC) rd",
"10: 0x1/2 rd API:2 SI:100",
"10: rd API:2 SI:100",
"0x1/2 rd API:2 SI:100",
"rd API:2 SI:100",
})
{
It f = input.begin(), l = input.end();
AST::Node data;
bool ok = qi::phrase_parse(f, l, p, qi::ascii::space, data);
if (ok) {
std::cout << "Parse success: " << data << "\n";
}
else {
std::cout << "Parse failure ('" << input << "')\n";
}
if (f!=l) {
std::cout << "Remaining unparsed input: '" << std::string(f,l) << "'\n";
}
}
}