In 2013, we were using Hadoop at Exalead, writing data processing pipelines with Cascading. But Cascading’s Java API was… not very friendly to use.
I wanted a more “fluent” API, so I wrote an OSS project on top of Cascading to provide a fluent API. I used Flapi, which is an API generator for Java, “generating ‘smart’ interfaces for improved fluency in your code”.
The cascading-flapi project was used at Exalead to help write our data processing pipelines… until switching to Spark.
Note that after talking about this API with the maintainers of Cascading, they started building their own wrapper a few months later: Fluid.
Project link: https://github.com/vbehar/cascading-flapi