dbt init --adapter default to the available adapter

[jtcohen6]

Describe the bug

In v0.18+, dbt init takes an additional argument --adapter so as to generate profiles.yml according to the adapter type specified.

Steps To Reproduce

Fresh install of a non-core plugin:

pip install dbt-spark

Initialize a project:

$ dbt init my_project
Running with dbt=0.18.0
Creating dbt configuration folder at /Users/jerco/.dbt
With sample profiles.yml for redshift
Encountered an error:
Runtime Error
  Could not find adapter type redshift!

Expected behavior

The --adapter argument should check the adapters available in the context

System information

Which database are you using dbt with?

• postgres
• redshift
• bigquery
• snowflake
• other (any community-contrib plugin)

The output of dbt --version:

dbt --version
installed version: 0.18.0
   latest version: 0.18.0

Up to date!

Plugins:
  - spark: 0.18.0

[jtcohen6]

Another issue: If the adapter plugin doesn't define a sample profiles.yml, dbt doesn't allow the user to run dbt init. Instead, dbt should fall back to a generic one, or do anything smarter than disallowing use of the init command entirely (cc @amychen1776).

[jtcohen6]

Copying from #3434:

If the --adapter argument is not passed, I think our options are:

• init checks the adapters available, rather than falling back to redshift as the default
• init falls back to a more reasonable, generic default (not redshift) that ships inside of core