使用boost::spirit解析Newick语法

我认为您的程序中有很多cargo cult编码。例如，变体是完全无用的。所以我重写了一点，添加了一些评论来帮助你理解（我希望，如果不清楚，请不要犹豫，在评论中询问）。我把空间规格放在一边，因为我认为它对你来说毫无用处。

#include <boost/spirit/include/qi.hpp>
#include <boost/fusion/include/adapt_struct.hpp>
#include <boost/spirit/include/phoenix_core.hpp>
#include <boost/spirit/include/phoenix_operator.hpp>
#include <boost/spirit/include/phoenix_fusion.hpp>
#include <boost/spirit/include/phoenix_stl.hpp>
#include <vector>
#include <string>
#include <iostream>
namespace parser
{
    // Forward declaration for the vector
    struct ptree;
    // typedef to ease the writing
    typedef std::vector<ptree> children_vector;
    // The tree structure itseflf
    struct ptree
    {
        children_vector children;
        std::string name;
        double length;
    };
    // Streaming operator for printing the result
    std::ostream& operator<<(std::ostream& stream, const ptree& tree)
    {
        bool first = true;
        stream << "(" << tree.name << ": " << tree.length << " { ";
        for (auto child: tree.children)
        {
            stream << (first ? "" : "," ) << child;
            first = false;
        }
        stream << " }";
        return stream;
    }
}
// adapt the structure to fusion phoenix
BOOST_FUSION_ADAPT_STRUCT(
    parser::ptree,
    (parser::children_vector, children)
    (std::string, name)
    (double, length)
)
namespace parser
{
    // namespace aliasing to shorten the names
    namespace qi = boost::spirit::qi;    
    namespace phoenix = boost::phoenix;
    // This grammar parse string to a ptree
    struct newick_grammar : qi::grammar<std::string::const_iterator, ptree()>
    {
    public:
        newick_grammar() 
            : newick_grammar::base_type(tree) // We try to parse the tree rule
        {                
            using phoenix::at_c; // Access nth field of structure
            using phoenix::push_back; // Push into vector
            // For label use %= to assign the result of the parse to the string
            label %= qi::lexeme[+(qi::char_ - ':' - ')' - ',')]; 
            // For branch length use %= to assign the result of the parse to the
            // double
            branch_length %= ':' >> qi::double_;
            // When parsing the subtree just assign the elements that have been
            // built in the subrules
            subtree = 
                // Assign vector of children to the first element of the struct
                -descendant_list [at_c<0>(qi::_val) = qi::_1 ] 
                // Assign the label to the second element
                >> -label [ at_c<1>(qi::_val) = qi::_1 ]
                // Assign the branch length to the third element 
                >> -branch_length [ at_c<2>(qi::_val) = qi::_1 ];
            // Descendant list is a vector of ptree, we just push back the
            // created ptrees into the vector
            descendant_list = 
                '(' >> subtree [ push_back(qi::_val, qi::_1) ]
                >> *(',' >> subtree [ push_back(qi::_val, qi::_1) ])
                >> ')';
            // The tree receive the whole subtree using %=
            tree %= subtree  >> ';' ;
        }
    private:
        // Here are the various grammar rules typed by the element they do
        // generate
        qi::rule<std::string::const_iterator, ptree()> tree, subtree;
        qi::rule<std::string::const_iterator, children_vector()> descendant_list;
        qi::rule<std::string::const_iterator, double()> branch_length;
        qi::rule<std::string::const_iterator, std::string()> label;
    };
}
int main(int argc, char const *argv[])
{
    namespace qi = boost::spirit::qi;
    std::string str;
    while (getline(std::cin, str))
    {
        // Instantiate grammar and tree
        parser::newick_grammar grammar;
        parser::ptree tree;
        // Parse
        bool result = qi::phrase_parse(str.cbegin(), str.cend(), grammar, qi::space,  tree);
        // Print the result
        std::cout << "Parsing result: " << std::boolalpha << result << std::endl;
        std::cout << tree << std::endl;
    }
    return 0;
}

以下是您的样本输出：

$ ./a.exe
(((One:0.1,Two:0.2)Sub1:0.3,(Three:0.4,Four:0.5)Sub2:0.6)Sub3:0.7,Five:0.8)Root:0.9;
Parsing result: true
(Root: 0.9 { (Sub3: 0.7 { (Sub1: 0.3 { (One: 0.1 {  },(Two: 0.2 {  } },(Sub2: 0.6 { (Three: 0.4 {  },(Four: 0.5 {  } } },(Five: 0.8 {  } }