mPyPl

Monadic Pipeline Library for Python

This library was created by a team of enthusiastic software developers / data scientists at Microsoft, who wanted to simplify tasks of data processing and creating complex data pipelines. The library is inspired by the following main ideas:

• Using functional approach to data processing (which implies immutability, lazy evaluation, etc.)
• Using pipe module in Python to achieve data pipelines similar to F#.
• Data pipeline uses dictionaries with different fields as base type, new operations would typically enrich data and add new fields by using apply function. Those dictionaries are similar to monads, and apply is similar to lift operation on monads. Thus the naming of the library.

Credits

Principal developers of mPyPl:

• Dmitri Soshnikov
• Yana Valieva
• Tim Scarfe