ENH: start of patsy branch for discussion

[sjsrey]

The justification for this PR is: to start a discussion and explore how to have a formula syntax across the library. This builds on the discussion in spreg with an eye towards what other packages in the library could benefit from this type of syntax.

[AppVeyorBot]

✅ Build libpysal 1.0.50 completed (commit 5b43eff7da by @sjsrey)

[codecov]

Codecov Report

Merging #260 into master will not change coverage by %.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master     #260   +/-   ##
=======================================
  Coverage   80.96%   80.96%           
=======================================
  Files         116      116           
  Lines       11582    11582           
=======================================
  Hits         9377     9377           
  Misses       2205     2205           

Impacted Files	Coverage Δ
libpysal/common.py	82.35% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 334716d...7b34a0d. Read the comment docs.

[knaaptime]

i'm returning to this issue as i'd like to start thinking about using mgwr simulation as an interpolation method over in tobler.

The critical piece is that patsy has now been succeeded by formulaic which seems both wonderfully fast and flexible. I haven't started testing it yet, but wanted to ping this issue again in case anyone was interested in hacking on this in the near future

cc @ljwolf @pedrovma @jGaboardi @TaylorOshan

[TaylorOshan]

Looks interesting @knaaptime . Could be fodder for a GSOC project @tdhoffman ?

[knaaptime]

would make a killer gsoc imo!

(also i accidentially tagged levi twice above instead of @darribas, so just @'ing him as well)

[tdhoffman]

I'm super interested in this for a GSoC project! @TaylorOshan and I have been tossing around something along these lines for a bit. Not sure how formulaic works, but I took a look at patsy at the time and it looked very dev-friendly. I think the real challenge of the project would be designing a library-wide semantics for the formulas, then the coding portion is just execution.

[knaaptime]

sweet! I agree on your take @tdhoffman, looking forward to seeing how this comes together :)

[TaylorOshan]

I do believe GSOC is on the docket for tomorrow's dev meeting @ 9am pacific, if you can make it @tdhoffman !

[tdhoffman]

I'll try to make it! I have a meeting from 8:30-9:30 pacific and at 10 pacific, so I'll hop on in the latter half