algemaploom-rs


AlgeMapLoom: Weaving Mapping Languages with Algebraic Operators

About The Project

Mapping algebra provides operational semantics to the mapping process, opening the door to study of complexity and expressiveness of existing mapping languages. This project provides the CLI translator from RML and ShExML to mapping algebra.

Prerequisites

To compile the project on your own, you’ll need to have Rust toolchain installed.

For Linux-based users:

For the visualization of the generated mapping plans, you’ll need python version >= 3.10 and the following packages:

  1. graphviz

Running

  1. Download this repo
  2. Run cargo build at the root this repo 
    cd {repo dir}
cargo build --release
  3. Run the CLI translator app from the compiled translator binary 
    cd ./target/release/
./translator  file  <RML_DOCUMENT>

    For more information/options of CLI app:

    ./translator  -h
  4. Visualize the created mapping plan 
    dot -Tpng {generated dot file} > output.png
  5. Simple plain text format of the mapping plan for parsing 
    dot -Tplain {generated dot file} > output.txt

    (back to top)

Mapping plan JSON schema

The JSON schema describing the structure of the mapping plan generated by translating the input mapping document can be found at the schema documentation page

Test cases

RML

Currently, the translator generates valid mapping plans for the official RML test cases with mapping plans for CSV data sources (all test cases ending in CSV).

The generated mapping plans for the test cases are inside the /resources/csv-testcases.

ShExML

The translator can partially translate ShExML documents. The translator will make a best-effort translation if the ShExML document uses the following unsupported features. It will still generate a mapping plan which could be executed but the results won’t be complete.

The following features are not supported in translation yet:

1) Autoincrements 2) Query statements 3) Joins 4) Functions 5) Conditionals

Bindings

AlgeMapLoom provides bindings for Java, Python and Node.js. These can be enabled with the features jni, pyo3 and neon respectively.

If you build from source, you can run the build_java.sh, build_python.sh and build_nodejs.sh scripts respectively.

For cross-compiling Windows binaries from Linux you need to install the MinGW cross-compiler and have the correct build targets installed:

For usage, check out the src/java, src/python, and src/nodejs folders.

Acknowledgement

This software makes use of sophia_rs crate!

(back to top)

This site is open source. Improve this page.