Add calibration offsets and radii as properties on BNO055 library

[sjirwin]

issue #21

builds on work by @katlings and replaces her PR
implements the approach suggested by @deshipu and addresses feedback from @caternuson

note - assignment is fighting against BNO055 auto-calibration so can have odd behavior

Adafruit CircuitPython 4.0.0-rc.1 on 2019-04-23; Adafruit Feather M4 Express with samd51j19
>>> import board, busio
>>> import adafruit_bno055 as bno055
>>> i2c = busio.I2C(board.SCL, board.SDA)
>>> bno = bno055.BNO055(i2c)
>>> bno.radius_magnetometer
0
>>> bno.radius_magnetometer = 40
>>> bno.radius_magnetometer = 40
>>> bno.radius_magnetometer
1
>>> bno.radius_magnetometer
0
>>> bno.offsets_accelerometer = 1, -1, 1
>>> bno.offsets_accelerometer
(1, -1, 1)

[tannewt]

Looking pretty good. Looks like the type checks need to be swapped to isinstance() for lint.

[caternuson]

note - assignment is fighting against BNO055 auto-calibration so can have odd behavior

I think that has to do with running in one of the "sensor fusion" modes, like NDOF.

The previous driver didn't do any checking for current mode. Not sure if that's worth adding here?

[evaherrada]

@caternuson this seems like something useful, as long as the auto-calibration doesn't mess with it. Is this something we should see about merging or close?

[caternuson]

Maybe have it as a follow on PR, with a new issue for tracking, just so we can get this PR closed off.

[kattni]

@caternuson Are you able to test this if you have not already?

[caternuson]

@kattni I didn't test this. Looks like that was notionally done in first post above. I can if you want me to though. In general, I'm OK with this once it passes travis. I think that's where things stalled.

EDIT - oh wait, that was done. @tannewt You OK with the changes?

[kattni]

@tannewt Can you please approve these changes if you're satisfied? Thanks!

[tannewt]

Looks good! Thanks!