Home - News ( United States | United Kingdom | Italy | Germany ) - Football scores

Showing content from https://github.com/danielaparker/jsoncons/ below:

danielaparker/jsoncons: A C++, header-only library for constructing JSON and JSON-like data formats, with JSON Pointer, JSON Patch, JSON Schema, JSONPath, JMESPath, CSV, MessagePack, CBOR, BSON, UBJSON

For the examples below you need to include some header files and initialize a string of JSON data:

jsoncons supports transforming JSON texts into C++ data structures. The functions decode_json and encode_json convert strings or streams of JSON data to C++ data structures and back. Decode and encode work for all C++ classes that implement jsoncons reflection traits. defined. jsoncons already supports many types in the standard library, and your own types will be supported too if you specialize json_type_traits in the jsoncons namespace.

namespace ns {
    enum class hiking_experience {beginner,intermediate,advanced};

    class hiking_reputon
    {
        std::string rater_;
        hiking_experience assertion_;
        std::string rated_;
        double rating_;
        std::optional<std::chrono::seconds> generated_; // assumes C++17, if not use jsoncons::optional
        std::optional<std::chrono::seconds> expires_;
    public:
        hiking_reputon(const std::string& rater,
            hiking_experience assertion,
            const std::string& rated,
            double rating,
            const std::optional<std::chrono::seconds>& generated = 
                std::optional<std::chrono::seconds>(),
            const std::optional<std::chrono::seconds>& expires = 
                std::optional<std::chrono::seconds>())
            : rater_(rater), assertion_(assertion), rated_(rated), rating_(rating),
              generated_(generated), expires_(expires)
        {
        }

        const std::string& rater() const {return rater_;}
        hiking_experience assertion() const {return assertion_;}
        const std::string& rated() const {return rated_;}
        double rating() const {return rating_;}
        std::optional<std::chrono::seconds> generated() const {return generated_;}
        std::optional<std::chrono::seconds> expires() const {return expires_;}

        friend bool operator==(const hiking_reputon& lhs, const hiking_reputon& rhs)
        {
            return lhs.rater_ == rhs.rater_ && lhs.assertion_ == rhs.assertion_ && 
                   lhs.rated_ == rhs.rated_ && lhs.rating_ == rhs.rating_ &&
                   lhs.confidence_ == rhs.confidence_ && lhs.expires_ == rhs.expires_;
        }

        friend bool operator!=(const hiking_reputon& lhs, const hiking_reputon& rhs)
        {
            return !(lhs == rhs);
        };
    };

    class hiking_reputation
    {
        std::string application_;
        std::vector<hiking_reputon> reputons_;
    public:
        hiking_reputation(const std::string& application, 
            const std::vector<hiking_reputon>& reputons)
            : application_(application), 
              reputons_(reputons)
        {}

        const std::string& application() const { return application_;}
        const std::vector<hiking_reputon>& reputons() const { return reputons_;}
    };

} // namespace ns

// Declare the traits. Specify which data members need to be serialized.

JSONCONS_ENUM_TRAITS(ns::hiking_experience, beginner, intermediate, advanced)
// First four members listed are mandatory, generated and expires are optional
JSONCONS_N_CTOR_GETTER_TRAITS(ns::hiking_reputon, 4, rater, assertion, rated, rating, 
                              generated, expires)

// All members are mandatory
JSONCONS_ALL_CTOR_GETTER_TRAITS(ns::hiking_reputation, application, reputons)

int main()
{
    // Decode the string of data into a c++ structure
    ns::hiking_reputation v = decode_json<ns::hiking_reputation>(data);

    // Iterate over reputons array value
    std::cout << "(1)\n";
    for (const auto& item : v.reputons())
    {
        std::cout << item.rated() << ", " << item.rating();
        if (item.generated())
        {
            std::cout << ", " << (*item.generated()).count();
        }
        std::cout << "\n";
    }

    // Encode the c++ structure into a string
    std::string s;
    encode_json(v, s, indenting::indent);
    std::cout << "(2)\n";
    std::cout << s << "\n";
}

This example makes use of the convenience macros JSONCONS_ENUM_TRAITS, JSONCONS_N_CTOR_GETTER_TRAITS, and JSONCONS_ALL_CTOR_GETTER_TRAITS to specialize the json_type_traits for the enum type ns::hiking_experience, the class ns::hiking_reputon (with some non-mandatory members), and the class ns::hiking_reputation (with all mandatory members.) The macro JSONCONS_ENUM_TRAITS generates the code from the enum identifiers, and the macros JSONCONS_N_CTOR_GETTER_TRAITS and JSONCONS_ALL_CTOR_GETTER_TRAITS generate the code from the get functions and a constructor. These macro declarations must be placed outside any namespace blocks.

A typical pull parsing application will repeatedly process the current() event and call next() to advance to the next event, until done() returns true.

You can apply a filter to a cursor using the pipe syntax (e.g., cursor | filter1 | filter2 | ...)

For the examples below you need to include some header files and initialize a buffer of CBOR data:

#include <iomanip>
#include <iostream>
#include <jsoncons/json.hpp>
#include <jsoncons_ext/cbor/cbor.hpp>
#include <jsoncons_ext/jsonpath/jsonpath.hpp>

using namespace jsoncons; // for convenience

const std::vector<uint8_t> data = {
    0x9f, // Start indefinte length array
      0x83, // Array of length 3
        0x63, // String value of length 3
          0x66,0x6f,0x6f, // "foo" 
        0x44, // Byte string value of length 4
          0x50,0x75,0x73,0x73, // 'P''u''s''s'
        0xc5, // Tag 5 (bigfloat)
          0x82, // Array of length 2
            0x20, // -1
            0x03, // 3   
      0x83, // Another array of length 3
        0x63, // String value of length 3
          0x62,0x61,0x72, // "bar"
        0xd6, // Expected conversion to base64
        0x44, // Byte string value of length 4
          0x50,0x75,0x73,0x73, // 'P''u''s''s'
        0xc4, // Tag 4 (decimal fraction)
          0x82, // Array of length 2
            0x38, // Negative integer of length 1
              0x1c, // -29
            0xc2, // Tag 2 (positive bignum)
              0x4d, // Byte string value of length 13
                0x01,0x8e,0xe9,0x0f,0xf6,0xc3,0x73,0xe0,0xee,0x4e,0x3f,0x0a,0xd2,
    0xff // "break"
};

jsoncons allows you to work with the CBOR data similarly to JSON data:

• As a variant-like data structure, basic_json

• As a strongly typed C++ data structure that implements json_type_traits

• With cursor-level access to a stream of parse events

int main()
{
    // Parse the CBOR data into a json value
    json j = cbor::decode_cbor<json>(data);

    // Pretty print
    std::cout << "(1)\n" << pretty_print(j) << "\n\n";

    // Iterate over rows
    std::cout << "(2)\n";
    for (const auto& row : j.array_range())
    {
        std::cout << row[1].as<jsoncons::byte_string>() << " (" << row[1].tag() << ")\n";
    }
    std::cout << "\n";

    // Select the third column with JSONPath
    std::cout << "(3)\n";
    json result = jsonpath::json_query(j,"$[*][2]");
    std::cout << pretty_print(result) << "\n\n";

    // Serialize back to CBOR
    std::vector<uint8_t> buffer;
    cbor::encode_cbor(j, buffer);
    std::cout << "(4)\n" << byte_string_view(buffer) << "\n\n";
}

Output:

(1)
[
    ["foo", "UHVzcw", "0x3p-1"],
    ["bar", "UHVzcw==", "1.23456789012345678901234567890"]
]

(2)
50,75,73,73 (n/a)
50,75,73,73 (base64)

(3)
[
    "0x3p-1",
    "1.23456789012345678901234567890"
]

(4)
82,83,63,66,6f,6f,44,50,75,73,73,c5,82,20,03,83,63,62,61,72,d6,44,50,75,73,73,c4,82,38,1c,c2,4d,01,8e,e9,0f,f6,c3,73,e0,ee,4e,3f,0a,d2

int main()
{
    // Parse the string of data into a std::vector<std::tuple<std::string,jsoncons::byte_string,std::string>> value
    auto val = cbor::decode_cbor<std::vector<std::tuple<std::string,jsoncons::byte_string,std::string>>>(data);

    std::cout << "(1)\n";
    for (const auto& row : val)
    {
        std::cout << std::get<0>(row) << ", " << std::get<1>(row) << ", " << std::get<2>(row) << "\n";
    }
    std::cout << "\n";

    // Serialize back to CBOR
    std::vector<uint8_t> buffer;
    cbor::encode_cbor(val, buffer);
    std::cout << "(2)\n" << byte_string_view(buffer) << "\n\n";
}

Output:

(1)
foo, 50,75,73,73, 0x3p-1
bar, 50,75,73,73, 1.23456789012345678901234567890

(2)
82,9f,63,66,6f,6f,44,50,75,73,73,66,30,78,33,70,2d,31,ff,9f,63,62,61,72,44,50,75,73,73,78,1f,31,2e,32,33,34,35,36,37,38,39,30,31,32,33,34,35,36,37,38,39,30,31,32,33,34,35,36,37,38,39,30,ff

Note that when decoding the bigfloat and decimal fraction into a std::string, we lose the semantic information that the variant like data structure preserved with a tag, so serializing back to CBOR produces a text string.

A typical pull parsing application will repeatedly process the current() event and call next() to advance to the next event, until done() returns true.

int main()
{
    cbor::cbor_bytes_cursor cursor(data);
    for (; !cursor.done(); cursor.next())
    {
        const auto& event = cursor.current();
        switch (event.event_type())
        {
            case staj_event_type::begin_array:
                std::cout << event.event_type() << " " << "(" << event.tag() << ")\n";
                break;
            case staj_event_type::end_array:
                std::cout << event.event_type() << " " << "(" << event.tag() << ")\n";
                break;
            case staj_event_type::begin_object:
                std::cout << event.event_type() << " " << "(" << event.tag() << ")\n";
                break;
            case staj_event_type::end_object:
                std::cout << event.event_type() << " " << "(" << event.tag() << ")\n";
                break;
            case staj_event_type::key:
                // Or std::string_view, if supported
                std::cout << event.event_type() << ": " << event.get<jsoncons::string_view>() << " " << "(" << event.tag() << ")\n";
                break;
            case staj_event_type::string_value:
                // Or std::string_view, if supported
                std::cout << event.event_type() << ": " << event.get<jsoncons::string_view>() << " " << "(" << event.tag() << ")\n";
                break;
            case staj_event_type::byte_string_value:
                std::cout << event.event_type() << ": " << event.get<jsoncons::span<const uint8_t>>() << " " << "(" << event.tag() << ")\n";
                break;
            case staj_event_type::null_value:
                std::cout << event.event_type() << " " << "(" << event.tag() << ")\n";
                break;
            case staj_event_type::bool_value:
                std::cout << event.event_type() << ": " << std::boolalpha << event.get<bool>() << " " << "(" << event.tag() << ")\n";
                break;
            case staj_event_type::int64_value:
                std::cout << event.event_type() << ": " << event.get<int64_t>() << " " << "(" << event.tag() << ")\n";
                break;
            case staj_event_type::uint64_value:
                std::cout << event.event_type() << ": " << event.get<uint64_t>() << " " << "(" << event.tag() << ")\n";
                break;
            case staj_event_type::half_value:
            case staj_event_type::double_value:
                std::cout << event.event_type() << ": "  << event.get<double>() << " " << "(" << event.tag() << ")\n";
                break;
            default:
                std::cout << "Unhandled event type " << event.event_type() << " " << "(" << event.tag() << ")\n";
                break;
        }
    }
}

Output:

begin_array (n/a)
begin_array (n/a)
string_value: foo (n/a)
byte_string_value: 50,75,73,73 (n/a)
string_value: 0x3p-1 (bigfloat)
end_array (n/a)
begin_array (n/a)
string_value: bar (n/a)
byte_string_value: 50,75,73,73 (base64)
string_value: 1.23456789012345678901234567890 (bigdec)
end_array (n/a)
end_array (n/a)

You can apply a filter to a cursor using the pipe syntax,

int main()
{
    auto filter = [&](const staj_event& ev, const ser_context&) -> bool
    {
        return (ev.tag() == semantic_tag::bigdec) || (ev.tag() == semantic_tag::bigfloat);  
    };

    cbor::cbor_bytes_cursor cursor(data);
    auto filtered_c = cursor | filter;

    for (; !filtered_c.done(); filtered_c.next())
    {
        const auto& event = filtered_c.current();
        switch (event.event_type())
        {
            case staj_event_type::string_value:
                // Or std::string_view, if supported
                std::cout << event.event_type() << ": " << event.get<jsoncons::string_view>() << " " << "(" << event.tag() << ")\n";
                break;
            default:
                std::cout << "Unhandled event type " << event.event_type() << " " << "(" << event.tag() << ")\n";
                break;
        }
    }
}

Output:

string_value: 0x3p-1 (bigfloat)
string_value: 1.23456789012345678901234567890 (bigdec)

jsoncons requires a compiler with minimally C++11 support. It is tested in continuous integration on Github Actions and circleci. UndefinedBehaviorSanitizer (UBSan) diagnostics are enabled for selected gcc and clang builds. Since v0.151.0, it is integrated with Google OSS-fuzz, with coverage for all parsers and encoders.

CMake is a cross-platform build tool that generates makefiles and solutions for the compiler environment of your choice. On Windows you can download a Windows Installer package. On Linux it is usually available as a package, e.g., on Ubuntu,

sudo apt-get install cmake

Once cmake is installed, you can build and run the unit tests from the jsoncons directory,

On Windows:

> mkdir build
> cd build
> cmake .. -DJSONCONS_BUILD_TESTS=On
> cmake --build .
> ctest -C Debug --output-on-failure

On UNIX:

$ mkdir build
$ cd build
$ cmake .. -DJSONCONS_BUILD_TESTS=On
$ cmake --build .
$ ctest --output-on-failure

jsoncons uses the PVS-Studio static analyzer, provided free for open source projects.

A big thanks to the comp.lang.c++ community for help with implementation details.

The jsoncons platform dependent binary configuration draws on to the excellent MIT licensed tinycbor.

Thanks to Milo Yip, author of RapidJSON, for raising the quality of JSON libraries across the board, by publishing the benchmarks, and contacting this project (among others) to share the results.

The jsoncons implementation of the Grisu3 algorithm for printing floating-point numbers follows Florian Loitsch's MIT licensed grisu3_59_56 implementation, with minor modifications.

The macro JSONCONS_ALL_MEMBER_TRAITS follows the approach taken by Martin York's ThorsSerializer

The jsoncons implementations of BSON decimal128 to and from string, and ObjectId to and from string, are based on the Apache 2 licensed libbson.

Special thanks to our contributors

RetroSearch is an open source project built by @garambo | Open a GitHub Issue

Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo

HTML: 3.2 | Encoding: UTF-8 | Version: 0.7.4