C++ implementation of the Jupyter Kernel protocol
xeus
is a library meant to facilitate the implementation of kernels for Jupyter. It takes the
burden of implementing the Jupyter Kernel protocol so developers can focus on implementing the
interpreter part of the kernel.
An example of kernel built with xeus is xeus-cling
,
a kernel for the C++ programming language based on the cling
C++ interpreter.
xeus
has been packaged on all platforms for the conda package manager.
conda install xeus -c QuantStack
xeus
enables custom kernel authors to implement Jupyter kernels more easily. It takes the burden of implementing the Jupyter Kernel protocol so developers can focus on implementing the interpreter part of the Kernel.
The easiest way to get started with a new kernel is to inherit from the base interpreter class xeus::xinterpreter
and implement the private virtual methods:
execute_request_impl
complete_request_impl
inspect_request_impl
history_request_impl
is_complete_request_impl
as seen in the echo kernel provided as an example.
#include "xeus/xinterpreter.hpp"
using xeus::xinterpreter;
using xeus::xjson;
using xeus::xhistory_arguments;
namespace echo_kernel
{
class echo_interpreter : public xinterpreter
{
public:
echo_interpreter() = default;
virtual ~echo_interpreter() = default;
private:
void configure() override;
xjson execute_request_impl(int execution_counter,
const std::string& code,
bool silent,
bool store_history,
const xjson::node_type* user_expressions,
bool allow_stdin) override;
xjson complete_request_impl(const std::string& code,
int cursor_pos) override;
xjson inspect_request_impl(const std::string& code,
int cursor_pos,
int detail_level) override;
xjson history_request_impl(const xhistory_arguments& args) override;
xjson is_complete_request_impl(const std::string& code) override;
xjson kernel_info_request_impl() override;
void input_reply_impl(const std::string& value) override;
};
}
Kernel authors can then rebind to the native APIs of the interpreter that is being interfaced, providing richer information than with the classical approach of a wrapper kernel capturing textual output.
xeus
depends on the following libraries: libzmq
,
cppzmq
, cryptopp
,
nlohmann_json
, and xtl
.
xeus | libzmq | cppzmq | cryptopp | xtl | nlohmann json |
---|---|---|---|---|---|
master | 4.2.3 | 4.2.3 | 5.6.5 | ^0.4.0 | ^3.1.1 |
0.11.0 | 4.2.3 | 4.2.3 | 5.6.5 | ^0.4.0 | ^3.1.1 |
0.10.x | 4.2.3 | 4.2.3 | 5.6.5 | ^0.4.0 | |
0.9.x | 4.2.3 | 4.2.2 | 5.6.5 | ^0.3.4 | |
0.8.x | 4.2.1 | 4.2.1 | 5.6.5 | ^0.3.4 |
On Linux platforms, xeus
also requires libuuid
, which is available in all linux distributions (uuid-dev
on Debian).
We have packaged all these dependencies for the conda package manager. The simplest way to install them with conda is to run:
conda install cmake zeromq cppzmq cryptopp nlohmann_json -c QuantStack
On Linux platform, you will also need:
conda install util-linux -c conda-forge
Once you have installed the dependencies, you can build and install xeus
:
cmake -D BUILD_EXAMPLES=ON -D CMAKE_BUILD_TYPE=Release
make
make install
If you need the xeus
library only, you can omit the BUILD_EXAMPLES
settings.
The dependencies can also be installed from source. Simply clone the directories and run the following cmake and make instructions.
cmake -D WITH_PERF_TOOL=OFF -D ZMQ_BUILD_TESTS=OFF -D ENABLE_CPACK=OFF
-D CMAKE_BUILD_TYPE=Release
make
make install
cppzmq
is a header only library:
cmake -D CMAKE_BUILD_TYPE=Release
make install
nlohmann_json
is a header only library
cmake
make install
cryptopp
must be built as a static library. Building cryptopp as a shared library is not supported on Windows.
cmake -D BUILD_SHARED=OFF -D BUILD_TESTING=OFF -D CMAKE_BUILD_TYPE=Release
make
make install
xtl
is a header only library:
cmake -D CMAKE_BUILD_TYPE
make install
We use a shared copyright model that enables all contributors to maintain the copyright on their contributions.
This software is licensed under the BSD-3-Clause license. See the LICENSE file for details.