NBToPackage

Software implementation, questions and answers

1. Having a separate class for every single plot seems like a bit of overkill, what are we gaining by doing things this way?

It gives us a way to configure the plots, and keep track of which configuration was used for each plot. Basically it give us the provenance tracking & reproducibility.

2. The plot-specific function _make_plot(self, cat) function only take one argument, namely the input catalog, but the function the user sees __call__(self, cat, **kwargs) takes any number of argument through the **kwargs mechanism. Why is this useful?

This give us a way to latch and store the configuration used to make the plot.

3. What do we get from having the PlotConfig class?

It gives us a way to associated a name for a plot with a Plotter subclass and some configuration parameters.

4. What do we get from having the PlotCollection class?

It gives use a top-level thing to talk to, which keeps track of all the plots we have.

5. What did we come up with for dealing with common parameters?

We stuck them in a single file, to make it easy to find them.

6. What is your overall take on the final package?

Well, I’m pretty happy with it.