RML IO - Konverter Frontend

This is a RML IO frontend for the konverter relational algebra executor. The frontend parses valid RML IO formatted mapping files and generates execution plans that can be executed with the konverter backend.

Prerequisites

• C++ Compiler: Install a C++ compiler (e.g., using sudo apt install build-essential).
• Konverter Backend: Download and build the konverter backend and place the required shared libraries in the backend/ directory.

Usage

Direct Usage

1. Prepare the Backend:

   • Download and build the konverter backend.
   • Place the required files (e.g., libexecutor.so, librapartitioner.so, libthreadexecutor.so) in the backend/ directory.

2. Build the Standalone C++ Files:

   • Run the build_standalone.sh script to compile the generated C++ files.

3. Run the Pipeline:

   • Execute the complete end-to-end pipeline with:

     python3 rml_frontend.py -m path/to/mapping.ttl

Compilation into a Standalone Executable

1. Prepare the Backend:

   • Download and build the konverter backend.
   • Place the required files in the backend/ directory.

2. Build the C++ Files:

   • Run the build.sh script to compile the generated C++ files.

3. Bundle with Nuitka:

   • The build process will bundle everything into a standalone executable (e.g., rml_frontend.bin).

4. Run the Executable:

   • Execute the standalone executable using:

     ./rml_frontend.bin -m path/to/mapping.ttl

Notes

• Shared Libraries: Ensure that the shared libraries from the backend are correctly located.
• Build Scripts: The provided build_standalone.sh and build.sh scripts are required for compiling and packaging the frontend.

Benchmarks

We benchmarked the RML frontend and the konverter backend against FlexRML and Morph-KGC. The results are shown below. Missing values indicate a timeout.

Contributing

Feel free to open issues or submit pull requests if you encounter any problems or have suggestions for improvements.

License

The code uses the serd library for RDF parsing, which is licensed under the ISC License and is downloaded during the build process.

Our code is licensed under the GNU Affero General Public License version 3.