Reimplement get_relation to handle custom schemas

[jtcohen6]

• Fixes Custom schemas: table already exists #38
• Closes Custom schema: table already exists #42

[drewbanin]

Looks great! Ship it :D

[drewbanin]

(but also, fix the pep8 issues that found their way into master if you can)

[jtcohen6]

Yup, looks like the pep8 errors (in connections.py) shipped with some other code. I'll just push a fix to this branch.

[jtcohen6]

@drewbanin Could you look at the failing unit test test_http_connection?

[drewbanin]

@jtcohen6 I think this was broken when this line changed:

https://github.com/fishtown-analytics/dbt-spark/blob/0171a25adb37c76449cafc9e07bfbf4e3f98970a/dbt/adapters/spark/connections.py#L19-L21

We just want to call .strip() on that string to remove the leading and trailing whitespace, eg:

SPARK_CONNECTION_URL = '''
https://{host}:{port}/sql/protocolv1/o/{organization}/{cluster}
'''.strip()

When I pop open the debugger, I see that this whitespace impedes the parsing of the url: