Visualize workflow diagram in Jupyter notebook or Terminal

Users may want to check the current workflow diagram inside the Jupyter notebook.

For example,

...

@resource(cpu=1)
@env(pip_packages=["torch", "numpy", "monai", "pillow"])
class App(Application):
    def compose(self):
        load_pil_op = LoadPILOperator()
        classifier_op = MedNISTClassifierOperator()

        self.add_flow(load_pil_op, classifier_op)

my_app = App()

my_app.show()  # or just 'my_app'

We would like to see the DAG in the Jupytergraph by executing my_app.show().

scikit-learn recently introduced awesome visualization of pipeline

• https://twitter.com/scikit_learn/status/1448413235720359939?s=21
• https://scikit-learn.org/dev/auto_examples/miscellaneous/plot_pipeline_display.html#sphx-glr-auto-examples-miscellaneous-plot-pipeline-display-py

References:

• https://networkx.org/documentation/stable/reference/drawing.html
• https://networkx.org/documentation/stable/auto_examples/drawing/plot_chess_masters.html
• https://networkx.org/documentation/stable/auto_examples/drawing/plot_directed.html
• https://stackoverflow.com/questions/62512760/how-to-label-edges-of-a-multigraph-in-networkx-and-matplotlib
• https://api-lab.dimensions.ai/cookbooks/8-organizations/3-Organizations-Collaboration-Network.html
• https://github.com/WestHealth/pyvis/issues/53