Add and test Jupyter notebooks for fabfed example workflows

[xi-yang]

We need a design that uses the Jupyter notebooks as an interface to present the fabfed to users.

• [x] 1. Install the fabfed package and add the needed configs through the JupyterHub terminal CLI. A single install script is preferred, but we can start with some manual steps.
• [x] 2. Create a set of notebooks that demo the examples provided in the FabFed Users Manual. They should cover all the providers and roughly the same demos we provided at KNIT-7.
• [x] 3. Test through all the notebooks.
• [x] 4. Create notebooks that use FabFed python function/methods to repeat the above task.

[zlion]

• Set up the initial structure of Jupiter notebook scripts.
• Installed the latest version of fabfed-py (1.7.0b) within notebook

[zlion]

• Created Jupyter notebooks for all providers: fabric, aws, gcp, sense-aws, sense-gcp, cloudlab and Chameleon.
• Tested workflows for all workflow cases described in the manual
• Reported the issues in GitHub

Now working on notebooks using FabFed python functions to create workflow.

[zlion]

• Finished the notebooks using FabFed library
• Under code reviewing.

[abessiari]

@zlion

The handling of fabric credentials should probably be done in a separate notebook. Other notebooks can point to it. The fabric credential notebook should have a quick overview section for each provider and its credentials .... Minimally it should what credentials attributes are needed.

[abessiari]

@zlion For the two quick scenarios we have discussed we should probably have a getting started section which would allow people to run the network stitching in less than two minutes ... No need to have more than one notebook. Users can set the config directory and do either chameleon or cloudlab

[abessiari]

@zlion The description for plan should something like this. It does a whole lot more than list the resources: I recommend: Creates an execution plan which lets you preview the changes before applying

[abessiari]

show is pretty important it displays the state .... it should have its own cell

fabfed_manager.show(session=session_name)

[abessiari]

show (state) and show sessions should be after apply and after destroy.

after destroy if all is well they will both display an emty list

[abessiari]

Listing resources is a few lines. We should have it in its own cell after validate.

The code here lists the resources in the order they will handled.

for resource_config in config.resource_configs:
            dependencies = []

            for dependency in resource_config.dependencies:
                dependencies.append(dependency.resource)

            print(resource_config.label, "depends on ", dependencies)

[abessiari]

For chameleon no need for two files. Add a variable site that defaults to STAR. The only difference is the site. One has STAR and the other has TACC. So basically one config under chameleon suffices. You can then go with the default STAR or if you want TACC you use variables. (edited)