This is an engine that converts data of one structure to another, based on a configuration file which describes how.
The configuration file can be written in either the native protobuf format or a condensed Whistle Data Transformation Language which is transpiled to protobuf configs for you.
The engine accepts data in JSON format and outputs it in JSON format. For information on the mapping configuration, look at the protobuf files in the proto directory.
This repository is organized into several packages that together enable you to author Whistle configs, extend existing mapping configurations, and test configs within a Jupyter notebook environment.
We highly recommend that you start by setting up your Jupyter Notebook environment using the published docker images and executing the example notebook. Once setup, work through the Whistle Data Transformation Language Codelab to get yourself familiar with Whistle. As you author more Whistle configs, use the Whistle Data Transformation Language Reference to deepen your understanding of the language.
This project consists of three components, the mapping engine, the mapping language, and Jupyter notebook UI extensions and magic commands. If you want to build the mapping engine and mapping language packages:
Make sure you have installed and added to PATH
- Golang (>= 1.17)
- Java JDK (>= 8)
- Protobuf Compiler
protoc(>= 3.11.4) - Clang (>= 11.0.1-2)
Then run build_all.sh.
This command will build and run the tests of the above packages. In addition, there are a set of JupyterLab UI extensions and magic commands that simplify the authoring workflow. The extensions are packaged into a set of pre-built and published docker images that contain and Jupyter notebook extensions/magic commands and does not require you to build the mapping engine and mapping library packages. For more details about each package, please refer to their individual READMEs for more information.
A language reference is available: Whistle Data Transformation Language Reference
Please refer to the Whistle Data Transformation Language Codelab for instructions on how to run the mapping engine and for getting familiar with the mapping language.
Whistle configs can be executed in Apache Beam. Please refer to the Whistle Dataflow Pipelines Repo for sample pipelines.
Want to help the Google Cloud Healthcare and Life Sciences team improve Whistle? Please email: whistle-feedback@google.com to connect with the Whistle team for a further discussion on your experience with Whistle.
Apache License, Version 2.0