Replace standard deviation in operator estimation with standard error

[msohaibalam]

Resolves Issue #864. Also introduces a few other minor cosmetic changes:

• Changes the name of the optional argument readout_symmetrize to symmetrize_readout in measure_observables(..), to make it read more similar to the optional argument calibrate_readout
• Updates the warning message in _exhaustive_symmetrization to use f-strings instead of the old way of using format

[notmgsk]

Did you use a lorem ipsum generator for the title

[marcusps]

LGTM

[marcusps]

@joshcombes any thoughts?

[msohaibalam]

Did you use a lorem ipsum generator for the title

@notmgsk Watch out, my next few PRs are going to be titled Expecto patronum, Wingardium Leviosa and hocus pocus.

[msohaibalam]

All (@notmgsk, @marcusps, @joshcombes), I'd like to merge this in and move to other PRs. If there's any other thing you'd like to add to this one, please do so by tomorrow or forever hold your peace!

[karalekas]

So this will break any code that uses these attributes, but I guess we're assuming that most of our users are internal for this feature set?

[msohaibalam]

@karalekas Correct. The only other place that I'm aware of where this code is used is in forest-benchmarking. I've created a separate PR for forest-benchmarking (rigetti/forest-benchmarking#88) that would resolve any breaking changes in forest-benchmarking caused by merging this PR.