Rexl, also written RExL or REXL, stands for Research Expression Language. It shares heritage and some design goals with Microsoft Power Fx. It is designed to be embedded in applications.

This repository contains:

• The core Rexl functionality including lexer, parser, binder, and code generator.
• Flow graph functionality, with data nodes and formula nodes, where dependencies (edges) are defined by Rexl formulas.
• ONNX runtime integration with some sample models taken from the ONNX Model Zoo. The models are exposed as functions in the Rexl language.
• Solver integration, including a Boolean Satisfiability Solver (SAT solver), and several linear MIP solvers.
• Statement Rexl, also known as RexlScript.
• Harness functionality for embedding RexlScript in an application.
• Some sample applications, including:
  • A Jupyter Notebook kernel hosting RexlScript and usable with both Jupyter Lab and Jupyter Notebook.
  • RexlBench, a GUI application for editing and running RexlScript.
  • RexlRun, a command line application for running RexlScript.
  • DocBench, a GUI application for editing and evaluating flow graphs.
• Sample notebooks for Jupyter and sample scripts for RexlBench/RexlRun.
• Documentation.

• Overview
• Building
• Packages
• Releases
• Documentation
• Samples
• Contributing
• Trademarks

Rexl is a pure functional expression language. It is statically typed. Types are inferred, not directly declared by the expression. Expressions are compiled to a semantic representation known as bound nodes. Bound nodes can be compiled to CIL (the instruction set of the .Net runtime) and evaluated. We may also implement a bound node interpreter at some point. Another possible improvement is to delegate operations to "smart" data sources such as SQL Server.

There is a larger language called RexlScript. RexlScript is interpreted, not compiled. RexlScript supports control flow as well as declaration and execution of parallel tasks. RexlScript is supported by the Jupyter Notebook kernel, the sample applications RexlBench and RexlRun, and is used by some test suites.

Core Rexl consists of the following projects:

• Rexl.Base: type system, name handling and other utilities.
• Rexl.Bind: Contains the lexer, parser, and binder, as well as parse tree and semantic (bound) tree representations. This includes parsing of RexlScript.
• Rexl.Code: Contains MSIL/CIL code generation and type manager functionality. A type manager is responsible for mapping from Rexl type (known as DType) to .Net System.Type.

Other projects include:

• Rexl.Flow: Contains the flow graph functionality.
• Rexl.Onnx: Contains basic ONNX runtime integration including sample functions wrapping models from the ONNX Model Zoo.
• Rexl.Solve: Contains support for invoking a SAT solver and several linear MIP solvers.
• Rexl.Harness: Contains harness functionality for executing RexlScript. This includes support for RexlScript's task concept.

The Rexl.sln file can be built in the standard ways using Microsoft Visual Studio or the dotnet command line application. For example, from a command prompt:

• cd to the src directory.
• Run dotnet build Rexl.sln -c Debug to build the debug configuration.
• Run dotnet build Rexl.sln -c Release to build the release configuration.

To build in Linux and WSL (Windows Subsystem for Linux):

• cd to the src directory.
• Run dotnet build RexlCrossPlat.sln -c Debug to build the debug configuration.
• Run dotnet build RexlCrossPlat.sln -c Release to build the release configuration.

Similarly the tests can be run in the standard ways using Microsoft Visual Studio or the dotnet command line application:

• cd to the src directory.
• Run dotnet test Rexl.sln -c Debug to run the tests for the debug configuration.
• Run dotnet test Rexl.sln -c Release to run the tests for the release configuration.

To run tests under Linux or WSL, specify RexlCrossPlat.sln instead of Rexl.sln.

To run RexlScript in a Jupyter notebook:

• Build Debug or Release.
• cd to the src\Apps\Kernel directory.
• Register the Jupyter kernel by running RegisterKernel.cmd Debug or RegisterKernel.cmd Release.
• cd to the directory where you have notebooks or want to create notebooks.
• Run jupyter lab or jupyter notebook.
• Create or open a Rexl notebook.

To run the other sample applications, RexlBench, RexlRun, or DocBench simply run the corresponding .exe from within Visual Studio, from the command line, or from windows explorer.

Packages are published to nuget.org. The published packages include:

Name	Released Package
Microsoft.Rexl.Base
Microsoft.Rexl.Bind
Microsoft.Rexl.Code
Microsoft.Rexl.Flow
Microsoft.Rexl.Harness
Microsoft.RexlKernel.Base

There will be more in the future, for example, Microsoft.Rexl.Onnx and Microsoft.Rexl.Solve.

We publish releases of RexlKernel here in GitHub. A release consists of .zip files for supported operating system / architecture configurations. Each zip is self-contained. To install and use:

• Ensure that Python and Jupyter are installed. The latter can be accomplished with pip install jupyter.
• Download the appropiate RexlKernel_<OS>_<Arch>_Release.zip file.
• Extract the contents of the .zip to a folder.
• Open a command line shell and cd into the extracted folder.
• Run RexlKernel register.
• cd to where you would like to create notebook files.
• Run jupyter lab.
• In Jupyter Lab, you should see an icon for creating a Rexl notebook.

The samples directory contents may be dowloaded and used as a launch directory for jupyter lab.

Documentation for this project is in the docs directory.

There is a samples directory containing sample data, scripts, and notebooks. More will be added over time. To run sample scripts, use either RexlBench or RexlRun. To run notebooks, follow the instructions in Running Jupyter. We hope to add Jupyter kernel installer functionality soon, to simplify the process.

Another source of samples is the many test scripts containing examples of Rexl and RexlScript. For example, see src/Test/Rexl.Code.Test/Scripts/Block/General/Basic.txt for some sample RexlScript.

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.

When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact [email protected] with any additional questions or comments.

This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow Microsoft's Trademark & Brand Guidelines. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos are subject to those third-party's policies.