Yet another C++ scientific computing library.
Objective: Provide the SciPy / NumPy API for std::array and std::vector.
Scicpp is a header only library. It's only dependence is the Eigen linear algebra library, also header only.
SciCpp uses the C++-17 standard. It is tested with the GCC and Clang compilers under Linux.
See documentation for more details.
To use SciCpp you can directly download a release zip file. Since SciCpp is a header library you can just unzip the content to a convenient place for you.
If you want to install it globally on your system you can unzip it in the /usr/include directory.
Else you can also save it directly into your project and add an include directive -I to the compiler to specify the path.
SciCpp requires the Eigen library for some functions. You can install it by running the script:
sudo bash scripts/install_eigen.sh
that will install it globally on your system.
If you want to use the plotting functions, you also need to install Sciplot:
sudo bash scripts/install_sciplot.sh
After cloning the repository, you can follow the circleci configuration file to setup your development environment. Only tested on Ubuntu 18.04 for now.
SciCpp tries to get as much test coverage as possible.
Each file library header file (ex. my_header.hpp) should have a
corresponding test file in the same folder (my_header.t.hpp).
To build and run the tests:
make test
List of tests to run can be found here. You can comment out the tests you don't want to run.
Scicpp uses the Nonius framework for benchmarking. The header is already included in the repo. However the Boost library is required, you can install it using:
sudo make setup_benchmark
To build and run the benchmarks:
make benchmark
List of benchmarks to run can be found here.
You can comment out the tests you don't want to run.
make EXAMPLE=estimate-zeros example
make format