We're seeing downstream failures due to yesterday's initial go at linting (#597). Specifically, the removal of numpy & scipy from libpysal.common.py is causing failures in esda.getisord where they are being access through libpysal, not directly. Also, libpysal.cg.KDTree is being accessed through libpysal.common instead directly.

Affected downstream:

• giddy
• spaghetti
• splot
• spopt
• spreg

This will be a fix in esda and I'll open a PR over there shortly. I'd like to leave this open until we fix it over there.