HIRM visualizations

[emilyfertig]

Add plotting utilities to visualize HIRM results like Figures 5 and 6 of Saad et al.. This will require:

• Loading the output of util_io::to_txt into a Python HIRM (defined in src/hirm.py) or another Python data structure
• Adapting or forking the code in src/util_plot.py to reproduce figures like Figure 5, which shows the clustering of domains within each view, color-coded by output value. Only boolean output values are supported so far, but we can expand this to include color-coded categoricals and float values (not sure how to visualize strings.) The examples in the examples subdirectory generate these plots (and are still runnable because they use the Python HIRM, not the C++ one we've modified.)
• Coding Figure 6, which shows the posterior probability of shared cluster membership per domain. I don't think there is code to do this yet in the repo.

[emilyfertig]

Update: instead of loading the output of util_io::to_txt it might make sense to run/query the model with Pybind, which we'll want to do eventually anyway (or, if standing up Pybind is a lot of work, we could just stick with loading results from disk for now).