Measurement / Experiment container

Right now what we can do, and what is done in the notebook, to take a measurement is:

``` Python
qc.Loop(
...
).run( ... )
```

This is getting too restricted, we could start overloading the `run()` even more with all kinds of stuff but I have the feeling it will get too big.
There are a lot of things I might want to do. And i want to propose some kind of measurement container. That can take care of settings, metadata, storage, and notes and others.

This also encourages reusing loops, instead of `loop(...).run(...)`

I'd like to run a measurement like this.

``` Python
my_loop = qc.loop(c0[1:2:3])
your_loop = qc.loop(c0[1:2:3])

# change stuff when I dont want defaults
exp = qc.Measurement(
location = 'somewhere',
file_format = 'hdf5',
snapshot = True,
snapshot_update = False,
default_measurement = [p.ampl, p.phase]
)

# add metadata to this measurement
meta = {
'sample': 'xyz'
'logbook': {'paper': 'page 12',
'online': '//link'}
'notes': "This Measurement is really important!"
}
exp.add_info( meta )

# Define default / intended plots for this measurement
exp.add_plot(x=c0, y = m.amplitude, ...)
exp.add_plot(x=c0, y = m.amplitude.diff(), ..)

# set the loop that is actually run
exp.set_loop( my_loop )

# run the measurement:
# - Prepare whatever
# - Genereate and save the snapshot + metadata
# - Run the loop
exp.run()

exp.plot()
```

@alexcjohnson and others, what do you think about this?

Start by reviewing the existing qc.Loop(...).run(...) usage shown in the issue and the surrounding discussion about a Measurement container. The scope requires an agreed design for settings, metadata, storage, notes, plots, loops, snapshots, and run behavior before implementation; done means that design is accepted and its required behavior is covered.

Written by the indexing model from the issue text.