A python framework that provides the ability to compose computation data flows as DAG. This modular DAG representation currently enables (but not restricted to) reuse for computation logics, visualization of the computation graph and unit testing for specific computation nodes in a potentially complex computation flow. The package also enables users to add more DAG operation on the top of the ones mentioned above (e.g., a light-weight process for dynamic type-check of the data going in and out of computation nodes). The package is open-sourced, and may be found at https://github.com/yodasco/computation_dag


MP3