使用Boost :: Spirit解析异构数据

Question

我想弄清楚如何解决以下问题.

我有以下格式的结构:

struct Data
{
     time_t timestamp;
     string id;
     boost::optional<int> data1;
     boost::optional<string> data2;
     // etc...
};

这应该用以下格式的单行字符串解析:

human_readable_timestamp;id;key1=value1 key2=value2.....

当然,键的排序不必与结构中元素的顺序相匹配.

Boost :: Spirit是否适合此类数据？我该如何处理？我已经完成了这些示例,但我无法从示例中获得符合我要求的代码.

Answer 1

您可以使用置换解析器.我在这里做了一个非常相似的例子:

• 使用C++和BOOST读取JSON文件

如果你有重复键,那么使用a更有意义 Kleene*或许

1. 用语义动作来分配属性/或/
2. 使用属性自定义点来分配结果
3. PS.另请参阅Spirit Repository中的关键字解析器(Boost Qi使用函数编写规则)

如果你不想使用语义动作(提升精神:"语义动作是邪恶的"？)你可以稍微调整结构,以便在使用排列时匹配自动合成的属性类型data元素:

struct Data
{
    boost::posix_time::ptime timestamp;
    std::string id;
    struct Fields {
        boost::optional<int> data1;
        boost::optional<std::string> data2;
    } fields;
};

现在解析器可以是:

    timestamp = stream;

    text  = lexeme [ '"' >> *~char_('"') >> '"' ];
    data1 = "key1" >> lit('=') >> int_;
    data2 = "key2" >> lit('=') >> text;
    id    = lexeme [ *~char_(';') ];

    start = timestamp >> ';' >> id >> ';' >> (data1 ^ data2);

UPDATE

评论,使其"有弹性".我最终改变了排列解析器,并采用第一种编号方法(具有语义动作方法的Kleene星).

    id     = lexeme [ *~char_(';') ];

    auto data1 = bind(&Data::Fields::data1, _val);
    auto data2 = bind(&Data::Fields::data2, _val);

    other  = lexeme [ +(graph-'=') ] >> '=' >> (real_|int_|text);

    fields = *(
                ("key1" >> lit('=') >> int_) [ data1 = _1 ]
              | ("key2" >> lit('=') >> text) [ data2 = _1 ]
              | other
              );

    start  = timestamp >> ';' >> id >> -(';' >> fields);

这改变了以下方面:

• 为了能够跳过"其他"字段,我需要为"其他"字段提出合理的语法:

  other  = lexeme [ +(graph-'=') ] >> '=' >> (real_|int_|text);
  

  (允许一个由任何非空格组成的键除外=,后面跟着=,后跟数字(eager)或文本).

• 我扩展了文本的概念,以支持流行的引用/转义方案:

  text   = lexeme [ 
              '"' >> *('\\' >> char_ | ~char_('"')) >> '"'
            | "'" >> *('\\' >> char_ | ~char_("'")) >> "'"
            | *graph 
         ];
  

• 它允许重复相同的键(在这种情况下,它保留最后看到的有效值).

• 如果您想禁止无效值,请替换>> int_或>> text使用> int_或> text(期望解析器).

我已经通过一些具有挑战性的案例扩展了测试用例:

    2015-Jan-26 00:00:00;id
    2015-Jan-26 14:59:24;id;key2="value"
    2015-Jan-26 14:59:24;id;key2="value" key1=42
    2015-Jan-26 14:59:24;id;key2="value" key1=42 something=awful __=4.74e-10 blarg;{blo;bloop='whatever \'ignor\'ed' key2="new} \"value\""
    2015-Jan-26 14:59:24.123;id;key1=42 key2="value" 

它现在打印

----------------------------------------
Parsing '2015-Jan-26 00:00:00;id'
Parsing success
2015-Jan-26 00:00:00    id
data1: --
data2: --
----------------------------------------
Parsing '2015-Jan-26 14:59:24;id;key2="value"'
Parsing success
2015-Jan-26 14:59:24    id
data1: --
data2:  value
----------------------------------------
Parsing '2015-Jan-26 14:59:24;id;key2="value" key1=42'
Parsing success
2015-Jan-26 14:59:24    id
data1:  42
data2:  value
----------------------------------------
Parsing '2015-Jan-26 14:59:24;id;key2="value" key1=42 something=awful __=4.74e-10 blarg;{blo;bloop='whatever \'ignor\'ed' key2="new} \"value\""'
Parsing success
2015-Jan-26 14:59:24    id
data1:  42
data2:  new} "value"
----------------------------------------
Parsing '2015-Jan-26 14:59:24.123;id;key1=42 key2="value" '
Parsing success
2015-Jan-26 14:59:24.123000 id
data1:  42
data2:  value

Live On Coliru

//#define BOOST_SPIRIT_DEBUG
#include <boost/optional/optional_io.hpp>
#include <boost/date_time/posix_time/posix_time.hpp>
#include <boost/date_time/posix_time/posix_time_io.hpp>
#include <boost/fusion/adapted/struct.hpp>
#include <boost/spirit/include/qi.hpp>
#include <boost/spirit/include/phoenix.hpp>

namespace qi = boost::spirit::qi;
namespace phx = boost::phoenix;

struct Data
{
    boost::posix_time::ptime timestamp;
    std::string id;
    struct Fields {
        boost::optional<int> data1;
        boost::optional<std::string> data2;
    } fields;
};

BOOST_FUSION_ADAPT_STRUCT(Data::Fields,
        (boost::optional<int>, data1)
        (boost::optional<std::string>, data2)
    )

BOOST_FUSION_ADAPT_STRUCT(Data,
        (boost::posix_time::ptime, timestamp)
        (std::string, id)
        (Data::Fields, fields)
    )

template <typename It, typename Skipper = qi::space_type>
struct grammar : qi::grammar<It, Data(), Skipper> {
    grammar() : grammar::base_type(start) {
        using namespace qi;
        timestamp = stream;

        real_parser<double, strict_real_policies<double> > real_;

        text   = lexeme [ 
                    '"' >> *('\\' >> char_ | ~char_('"')) >> '"'
                  | "'" >> *('\\' >> char_ | ~char_("'")) >> "'"
                  | *graph 
               ];

        id     = lexeme [ *~char_(';') ];

        auto data1 = bind(&Data::Fields::data1, _val);
        auto data2 = bind(&Data::Fields::data2, _val);

        other  = lexeme [ +(graph-'=') ] >> '=' >> (real_|int_|text);

        fields = *(
                    ("key1" >> lit('=') >> int_) [ data1 = _1 ]
                  | ("key2" >> lit('=') >> text) [ data2 = _1 ]
                  | other
                  );

        start  = timestamp >> ';' >> id >> -(';' >> fields);

        BOOST_SPIRIT_DEBUG_NODES((timestamp)(id)(start)(text)(other)(fields))
    }
  private:
    qi::rule<It,                                 Skipper> other;
    qi::rule<It, std::string(),                  Skipper> text, id;
    qi::rule<It, boost::posix_time::ptime(),     Skipper> timestamp;
    qi::rule<It, Data::Fields(),                 Skipper> fields;
    qi::rule<It, Data(),                         Skipper> start;
};

int main() {
    using It = std::string::const_iterator;
    for (std::string const input : {
            "2015-Jan-26 00:00:00;id",
            "2015-Jan-26 14:59:24;id;key2=\"value\"",
            "2015-Jan-26 14:59:24;id;key2=\"value\" key1=42",
            "2015-Jan-26 14:59:24;id;key2=\"value\" key1=42 something=awful __=4.74e-10 blarg;{blo;bloop='whatever \\'ignor\\'ed' key2=\"new} \\\"value\\\"\"",
            "2015-Jan-26 14:59:24.123;id;key1=42 key2=\"value\" ",
            })
    {
        std::cout << "----------------------------------------\nParsing '" << input << "'\n";
        It f(input.begin()), l(input.end());
        Data parsed;
        bool ok = qi::phrase_parse(f,l,grammar<It>(),qi::space,parsed);

        if (ok) {
            std::cout << "Parsing success\n";
            std::cout << parsed.timestamp << "\t" << parsed.id << "\n";
            std::cout << "data1: " << parsed.fields.data1 << "\n";
            std::cout << "data2: " << parsed.fields.data2 << "\n";
        } else {
            std::cout << "Parsing failed\n";
        }

        if (f!=l)
            std::cout << "Remaining unparsed: '" << std::string(f,l) << "'\n";
    }
}