[LASP-Luke]

This PR introduces metadata fetching for xarray dataset global and variable metadata schemas which are applied to xarray datasets before returning them from the read_data function.

This metadata functionality has not yet been implemented for join requests.

Some minor imports cleanup and docstring additions were also made in this PR.

[LASP-Luke]

Here's an example of the metadata attributes added during a request for a TSIS (xarray) dataset:

Global attributes:

id: tsis_tsi_6hr
temporal_coverage: 2018-01-11/

Variable attributes:

time:
id: time
type: double
units: JD
title: Nominal Time

tsi_1au:
id: tsi_1au
type: double
units: W/m^2
missing_value: 0
title: Total Solar Irradiance (TSI) at 1-AU

instrument_accuracy_1au:
id: instrument_accuracy_1au
type: double
units: W/m^2
missing_value: 0
title: Instrument Accuracy in 1-AU TSI, 1 sigma

instrument_precision_1au:
id: instrument_precision_1au
type: double
units: W/m^2
missing_value: 0
title: Instrument Precision in TSI at 1-AU, 1 sigma

solar_standard_deviation_1au:
id: solar_standard_deviation_1au
type: double
units: W/m^2
missing_value: 0
title: Solar Standard Deviation in 1-AU TSI, 1 sigma

measurement_uncertainty_1au:
id: measurement_uncertainty_1au
type: double
units: W/m^2
missing_value: 0
title: Total Uncertainty in TSI at 1-AU, 1 sigma

tsi_true_earth:
id: tsi_true_earth
type: double
units: W/m^2
missing_value: 0
title: Total Solar Irradiance at Earth distance

instrument_accuracy_true_earth:
id: instrument_accuracy_true_earth
type: double
units: W/m^2
missing_value: 0
title: Instrument Accuracy at Earth distance, 1 sigma

instrument_precision_true_earth:
id: instrument_precision_true_earth
type: double
units: W/m^2
missing_value: 0
title: Instrument Precision at Earth distance, 1 sigma

solar_standard_deviation_true_earth:
id: solar_standard_deviation_true_earth
type: double
units: W/m^2
missing_value: 0
title: Solar Standard Deviation in TSI at Earth, 1 sigma

measurement_uncertainty_true_earth:
id: measurement_uncertainty_true_earth
type: double
units: W/m^2
missing_value: 0
title: Total Uncertainty in TSI at Earth distance, 1 sigma

avg_measurement_date:
id: avg_measurement_date
type: double
units: JD
missing_value: 0
title: Average Data Time

std_dev_measurement_date:
id: std_dev_measurement_date
type: double
units: days
missing_value: 0
title: Stdev of Average Data Time, 1 sigma

provisional_flag:
id: provisional_flag
type: int
title: Provisional Flag, 1=provisional data, 0=final data

index:

[dlindhol]

I fought with PyCharm a while and had to tweak some imports. (Any advice would be appreciated.) Otherwise, it looks good to me!

Could you also update the README with the xarray option? Mention that it will include metadata.

[dlindhol]

You might add "optional". Also, API keys are supported in addition to basic auth.

[dlindhol]

Did we discuss including optional title and description in the dataset metadata? Those would be handy.

Sorry for any scope creep.