Add client class

[jm-rivera]

This PR introduces theDataCommonsClient class to interact with the Data Commons API, along with various utility functions and tests.

DataCommonsClient

The DataCommonsClient class would be the main entry point for most users. As described in the design doc:

The datacommons entrypoint would coordinate interactions with and between classes.
It would simply provide convenient access to the different endpoints, centralising where the api_key and dc_instance are provided, and binding a specific ‘client’ instance to a specific knowledge graph and API key.
It will also add convenient methods to perform additional, higher-level interactions with the data/api. For example, it can implement a method to get data as a nicely structured Pandas DataFrame (if the optional dependency is installed).

This implementation achieves those objectives but it shouldn't be seen as the end state of the class. The ultimate aim is to provide a range of convenience tools and functionality. For now, I wanted to keep this PR small and focused around defining how it will all come together, including clearly showcasing how the optional Pandas dependency will work.

For example, connecting to base DC:

from datacommons_client.client import DataCommonsClient

client = DataCommonsClient(dc_instance="datacommons.org", api_key="api-key-string")

And using that client, getting all property labels for a given node (docs example)

response = client.node.fetch_property_labels(node_dcids="geoId/06", out=False)

Which returns a NodeResponse object with the API response

NodeResponse(
data={'geoId/06': Properties(properties=['affectedPlace',
                                                      'containedInPlace',
                                                      'location',
                                                      'member',
                                                      'overlapsWith'])},
nextToken=None
)

Or which can produce a flatted list of properties for each entity:

properties = response.get_properties()

{'geoId/06': ['affectedPlace',
              'containedInPlace',
              'location',
              'member',
              'overlapsWith']}

As another example, getting statistical observations as a Pandas DataFrame from a custom DC instance

client = DataCommonsClient(dc_instance="datacommons.one.org")

response = client.observations_dataframe(
    variable_dcids="sdg/SI_POV_DAY1",
    date="latest",
    entity_dcids=["country/GTM", "country/NGA"],
)

That will return a pandas Data Frame shaped as follows

	date	entity	variable	value	facetId	importName	measurementMethod	observationPeriod	provenanceUrl	unit
0	2018	country/NGA	sdg/SI_POV_DAY1	30.9	3549866825	UN_SDG	SDG_G_G		https://unstats.un.org/sdgs/dataportal	SDG_PERCENT
1	2014	country/GTM	sdg/SI_POV_DAY1	9.5	3549866825	UN_SDG	SDG_G_G		https://unstats.un.org/sdgs/dataportal	SDG_PERCENT

Or we could also get all data for African countries for a specific indicator.

response = client.observations_dataframe(
    variable_dcids="sdg/SI_POV_DAY1",
    date="all",
    entity_type="Country",
    parent_entity="africa",
)

response.head(5)

	date	entity	variable	value	facetId	importName	measurementMethod	observationPeriod	provenanceUrl	unit
0	1996	country/CMR	sdg/SI_POV_DAY1	50.4	3549866825	UN_SDG	SDG_G_G		https://unstats.un.org/sdgs/dataportal	SDG_PERCENT
1	2001	country/CMR	sdg/SI_POV_DAY1	25.7	3549866825	UN_SDG	SDG_G_G		https://unstats.un.org/sdgs/dataportal	SDG_PERCENT
2	2007	country/CMR	sdg/SI_POV_DAY1	31.4	3549866825	UN_SDG	SDG_G_G		https://unstats.un.org/sdgs/dataportal	SDG_PERCENT
3	2014	country/CMR	sdg/SI_POV_DAY1	25.7	3549866825	UN_SDG	SDG_G_G		https://unstats.un.org/sdgs/dataportal	SDG_PERCENT
4	1992	country/CAF	sdg/SI_POV_DAY1	82.2	3549866825	UN_SDG	SDG_G_G		https://unstats.un.org/sdgs/dataportal	SDG_PERCENT

A lot of convenient features are possible but this PR does not incorporate them (the subject of future PRs)...

These examples are just scratching the surface to showcase the idea. Let me know if more would be helpful!

[hqpho]

Love the example calls in the comments and PR description!!