Master thesis project repository. Test and development of new decoders using soft information.
To clone this repository, run the following command in your terminal:
git clone --recursive https://github.ibm.com/Maurice-Hanisch/Soft-Info.git
For macOS:
brew install bazel
brew install armadillo
brew install mlpack
brew install llvm # for clang and parrallelism
Then add the Brew install path to your include path variable:
/opt/homebrew/include
# also add include for omp.h
/opt/homebrew/opt/llvm/bin/../include/c++/v1
/opt/homebrew/Cellar/llvm/17.0.6/lib/clang/17/include
/Library/Developer/CommandLineTools/SDKs/MacOSX14.sdk/usr/include
NOTE: /opt/homebrew is the default path on Apple Silicon Macs, if you are using an Intel Mac, the path is /usr/local
CAUTION: if building with '-DCMAKE_CXX_FLAGS=-fopenmp'
the kernel will crash when throwing errors!
After cloning the repository, navigate to the project directory and install the required Python packages using Pipenv: (WARNING THIS MAY TAKE A WHILE BECAUSE C++ PARTS ARE COMPILED)
pipenv install
If you wish to change C++ parts, even if the soft_information_decoding package is editable you need to rebuild it using the following command:
pip install .
Install the stim wheel using pip. It is located in 'libs/PyMatching/bazel-bin/external/stim'.
pip install /path/to/stim-0.0.dev0-py3-none-any.whl