rFBP - rFBP v1.0.4

C++ compilers supported :

Python version supported :

It is just a fix of the Pypi installation of the package. We add the version of the file via CMake configure or setup generation.

Scientific Software - Peer-reviewed - C++
Published by Nico-Curti over 5 years ago

rFBP - rFBP v1.0.3

This is the official release for the JOSS paper .

Paper Authors

• name: Nico Curti orcid: 0000-0001-5802-1195 affiliation: 1
• name: Daniele Dall'Olio orcid: 0000-0003-0196-6870 affiliation: 3
• name: Daniel Remondini orcid: 0000-0003-3185-7456 affiliation: 3
• name: Gastone Castellani orcid: 0000-0003-4892-925X affiliation: 2
• name: Enrico Giampieri orcid: 0000-0003-2269-2338 affiliation: 1

Added

• Add the pyproject.toml for the build requirements according to PEP-518.
• Add the Pypi badge for the latest releases.
• Add the Doxygen build via CMake.
• Upload the package to Zenodo .

Fixed

• Fix bibliography and paper document
• Fix minor issues in the documentation

Scientific Software - Peer-reviewed - C++
Published by Nico-Curti over 5 years ago

rFBP - rFBP v1.0.2

C++ compilers supported :

Python version supported :

In this release we improve the documentation of both C++ and Python APIs. You can find the full documentation here:

We follow the instructions of JOSS reviewers, solving issues related to the installation and description of the project.

The issue related to the building on MacOS environments is still opened (ref https://github.com/openjournals/joss-reviews/issues/2663#issuecomment-703316056).

Scientific Software - Peer-reviewed - C++
Published by Nico-Curti over 5 years ago

rFBP - rFBP v1.0.0

C++ compilers supported :

Python version supported :

We propose a C++ version of the Replicated Focusing Belief Propagation Julia package. Our implementation optimizes and extends the original library including multi-threading support and an easy-to-use interface to the main algorithm. To further improve the usage of our code, we propose also a Python wrap of the library with a full compatibility with the scikit-learn and scikit-optimize packages.

The rFBP project is written in C++ using a large amount of c++17 features. To enlarge the usability of our package we provide also a retro-compatibility of all the c++17 modules reaching an usability (tested) of our code from gcc 4.8.5+. The package installation can be performed via CMake or Makefile.

If you are using the CMake (recommended) installer the maximum version of C++ standard is automatic detected. The CMake installer provides also the export of the library: after the installation you can use this library into other CMake projects using a simple find_package function. The exported CMake library (rFBP::rfbp) is installed in the share/rFBP directory of the current project and the relative header files are available in the rFBP_INCLUDE_DIR variable.

The CMake installer provides also a rFBP.pc, useful if you want link to the rFBP using pkg-config.

You can also use the rFBP package in Python using the Cython wrap provided inside this project. The only requirements are the following:

• numpy >= 1.15
• cython >= 0.29
• scipy >= 1.2.1
• scikit-learn >= 0.20.3
• requests >= 2.22.0

The Cython version can be built and installed via CMake enabling the -DPYWRAP variable. The Python wrap guarantees also a good integration with the other common Machine Learning tools provided by scikit-learn Python package; in this way you can use the rFBP algorithm as an equivalent alternative also in other pipelines. Like other Machine Learning algorithm also the rFBP one depends on many parameters, i.e its hyper-parameters, which has to be tuned according to the given problem. The Python wrap of the library was written according to scikit-optimize Python package to allow an easy hyper-parameters optimization using the already implemented classical methods.

Scientific Software - Peer-reviewed - C++
Published by Nico-Curti over 5 years ago