fix: use correct antpairs when finding bad pairs

[steven-murray]

This fixes a couple of things related to using nonavg data:

• it fixes the autocorrelation plot, and makes it a little faster (it was trying to plot many many panels previously)
• it fixes an error where the full list of antpairs was being used instead of just the ones for this chunk
• it tries to improve the memory consumption in the very last plot by closing figures after showing them.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[jsdillon]

Looks ok to me. Since it's just plotting code, I'm not really sure if it needs peer-review, but let me know if you want me to look more closely at any particular part.