Dependency Monkey inspecting Python dependencies of TensorFlow

Having Python dependencies correctly installed and operational is crucial to ship any Python application. In this video, we will take a look at evaluating the quality of Python dependencies of TensorFlow. A service called Dependency Monkey can compute combinations of software packages respecting dependency graph structure constructed out of direct as well as transitive dependencies of a TensorFlow application stack. Subsequently, the service builds a container image with the specified set of packages and verifies the correctness of the dependencies installed. The example shows issues that can arise when TensorFlow is installed with a specific set of packages but Dependency Monkey can do a similar job for any Python application stack.