ashgen / kdb-templates

A C++17 template library for converting to/from native C++ types to KDB

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

KDB C++ Template Library v0.0.1

NOTE: The current version of this library is an initial prototype that is being worked on. The APIs and feature set are not completely stable or well documented.

Introduction

Please Note! This project requires a C++17 compatible compiler to build as it's using constexpr and std::variant. You can check compiler compatiblility with these features at cppreference.com.

A template library for efficiently converting to/from native C++ types to KDB without any knowledge of the underlying KDB apis. This is a header only library, so you do not need to install anything to use it in your project. You can just copy the kdb.hpp header in the release directory into your own project.

Features

  • Header Only: The library has no dependencies other than copying a single header file into your project.
  • Zero Overhead: The library aims to generate code that's as fast as what you would write by hand. Features that you don't use cost nothing.
  • STL Enabled: Supports conversion of all the basic STL containers!
  • No Macros!: The macros that are in the k.h header file often conflict with template parameters used by popular C++ libraries (boost, eigen etc.). These are undefined by default in this library for better C++ compatibility.
  • Extensible: You can teach the library to convert your own types and they will work seamlessly with the existing types.

Examples

Basic Types

// just include the kdb.hpp header after k.h in your project to pull in all of the functionality
#include "k.h"
#include "kdb.hpp"

// the new type aliases live under the kdb::type namespace 
using kdb::type;

// some convenient type aliases are defined since the macros have been removed by the library
atom_bool  z = true;        // 1 byte boolean        (kdb: b)
atom_char  c = 'e';         // 1 byte character      (kdb: c) 
atom_short h = 372;         // 2 byte integer        (kdb: h)
atom_int   i = 65738;       // 4 byte integer        (kdb: i)
atom_long  j = 7245645623;  // 8 byte integer        (kdb: j)
atom_real  x = 174.26;      // 4 byte floating point (kdb: e)
atom_float y = 827784.123;  // 8 byte floating point (kdb: f)

// there are equivalent list types for all of the atom types - these are aliases for a std::vector
// and support all of the operations you are used to for manipulating lists.
list_bool  zl = { false, true, false };
list_char  cl = { 'a', 'b', 'c' };
list_short hl = { 1, 2, 3, 4, 5 };
list_int   il = { 1, 2, 3, 4, 5 };
list_long  jl = { 1, 2, 3, 4, 5 };
list_real  xl = { 26.4, 26.1, 85.3 };
list_float yl = { 100, 57.2, 73.274 };

// there are variant types that can hold any type of object and will be converted to the correct 
// type dynamically (note: there is some overhead vs using concrete types due to checking the 
// type before conversion). These are more useful when converting to C++ native types
atom_any a = "example";
atom_any b = 42;

// the list_mixed type holds atom_any objects and maps to a list of type 0 in kdb.
list_mixed ml = { 42, "example", false };
ml.push_back(73.5);
ml.push_back('a');

Conversions

atom_float f = 42.5;
list_float fl = { 42.5, 12.7, 85.2 };
list_mixed ml = { 42, "example", false };

// conversion to a K object is performed via from_native. This will generate
// code for the type at compile time as though you had written the conversion
// by hand.
K kf  = from_native(f);
K kfl = from_native(fl);
K kml = from_native(ml);

// supported types can be nested arbitrarily. all of the std:: types and
// collections are supported by default. it's easy to add your own custom
// types to this too.
template<typename T1, typename T2>
using dict = std::unordered_map<T1,T2>;

dict<atom_symbol, list_float> weather = { 
        {"NYC", {77, 64.5, 71.4, 74.7}},
        {"LDN", {63.2, 51.7, 65.8, 71.5}},
        {"HK",  {87.5, 83.61, 81.2, 76, 79.9}}
};

// conversion is exactly the same as with the simple types
K weather_table = from_native(weather);

// We can convert K datastructures back into native C++ types using to_native.
// In this case the type is inferred from the second argument that we are placing
// the result into.
dict<atom_symbol, list_float> result;
to_native(weather_table, result);

// you can also call to_native and specify the type explicitly.
auto result = to_native<dict<atom_symbol, list_float>>(weather_table)

Adding custom types

// The KDB_REGISTER macro can be used to teach the library about custom structures
// that you would like to serialize. This macro will generate code to convert the
// data to a mixed list containing the items.
//
// The macro takes a list of fields to include in the conversion. Any fields you
// don't place in the macro will be skipped and will remain uninitialized after
// a call to to_native.
struct User {
    int id;
    std::string name;
};
KDB_REGISTER(User, id, name)

struct SessionStatus {
    float timeout;
    int sessionId;
    std::string sessionName;
    std::vector<User> users;
};
KDB_REGISTER(SessionStatus, timeout, sessionId, sessionName, users)

// after registration, conversion of these types is as simple as calling from_native
SessionStatus session = { 3050.0, 100, "example session name", {}};
K kobj = from_native(session);

// the new type is also handled automatically within containers
std::vector<SessionStatus> data_list = {
        { 3050.0, 100, "example session name", {{"Alice", 1}, {"Bob", 2}}},
        { 1000.0, 85,  "another session name", {{"Peter", 7}}},
        { 8472.0, 10,  "more session names", {{"Sarah", 3}}}
};
K kobj = from_native(data_list);

// these complex custom types can be converted back into C++ types with full
// typechecking of the datastructure.
std::vector<SessionStatus> result;
to_native(kobj, result);

Using the library

The simple way to use this library is to just copy/vendor the release/kdb.hpp file into your own project. You will have to make sure any flags required for the C++14/17 features that are used are enabled in your build system.

You can also add an external dependency using the FetchContent module in CMake. This will download the library at configuration time and will make the headers available through the kdb-cpp target. Using this cmake target will ensure that all of the required flags are propagated to any targets that link against it.

include(FetchContent)

# Download the library from this repository
FetchContent_Declare(kdb-cpp
    GIT_REPOSITORY https://github.com/markrooney/kdb-templates.git
    GIT_TAG v0.0.1)

# Populate the content
FetchContent_GetProperties(kdb-cpp)
if(NOT kdb-cpp_POPULATED)
    FetchContent_Populate(kdb-cpp)
    add_subdirectory(${kdb-cpp_SOURCE_DIR} ${kdb-cpp_BINARY_DIR} EXCLUDE_FROM_ALL)
endif()

# Link against kdb-cpp anywhere that you need the header file to be available
target_link_libraries(<your-library-name> PRIVATE kdb-cpp)

Building

  1. Checkout the project with submodules. This will also checkout vcpkg for you.
# clone the project from github along with any submodules
git clone --recurse-submodules https://github.com/markrooney/kdb-templates.git
# change directory into the project
cd kdb-templates
  1. Bootstrap and initialize vcpkg for your system.
# initialize vcpkg on your system by running the bootstrap script
./thirdparty/vcpkg/bootstrap-vcpkg.sh
  1. Run cmake to build the executables.
# create the build directory and initialize cmake there.
cmake -DCMAKE_BUILD_TYPE=Release -B build .
# perform the build inside the build directory
cmake --build build
  1. Run the benchmark and test executables.
# run the benchmarks
./build/benchmarking
# run the tests
./build/test-runner

Combining headers into one file

If you have made a change to the library and would like to rebuild the combined header file, you can do so using the amalgamate scripts in the thirdparty folder. This should create a new kdb.hpp file in the release/ folder.

python thirdparty/amalgamate/amalgamate.py -c config.json -s .

License

Copyright 2020 Mark Rooney

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

 http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Contact

If you have any suggestions or issues relating to the library, please open an issue on GitHub.

About

A C++17 template library for converting to/from native C++ types to KDB

License:Apache License 2.0


Languages

Language:C++ 97.5%Language:q 1.9%Language:CMake 0.6%