Resources for SQL Analytics queries and dashboards

[pietern]

New resources are:

• databricks_sql_query
• databricks_sql_visualization
• databricks_sql_dashboard
• databricks_sql_widget

[codecov]

Codecov Report

Merging #553 (acfeb2c) into master (fc595ac) will increase coverage by 0.08%.
The diff coverage is 83.51%.

@@            Coverage Diff             @@
##           master     #553      +/-   ##
==========================================
+ Coverage   82.55%   82.64%   +0.08%     
==========================================
  Files          79       85       +6     
  Lines        7032     7765     +733     
==========================================
+ Hits         5805     6417     +612     
- Misses        812      878      +66     
- Partials      415      470      +55     

Impacted Files	Coverage Δ
sqlanalytics/api/widget.go	63.63% <63.63%> (ø)
sqlanalytics/resource_visualization.go	71.27% <71.27%> (ø)
sqlanalytics/api/query.go	73.25% <73.25%> (ø)
sqlanalytics/resource_widget.go	83.45% <83.45%> (ø)
sqlanalytics/resource_dashboard.go	84.61% <84.61%> (ø)
sqlanalytics/resource_query.go	94.75% <94.75%> (ø)
provider/provider.go	95.65% <100.00%> (+0.05%)	⬆️
... and 3 more

[nfx]

First batch of changes requested :)

[nfx]

This PR drops coverage by 19% - https://app.codecov.io/gh/databrickslabs/terraform-provider-databricks/compare/553/tree/sqlanalytics, please bring code coverage to the level of resource_sql_endpoint.go.

The other 3 things:

• comment in api package
• please add an acceptance test for each resource, like in https://github.com/databrickslabs/terraform-provider-databricks/blob/master/sqlanalytics/acceptance/sql_endpoint_test.go - i want to be sure that none will break it :)
• please add documentation for each resource, like in https://github.com/databrickslabs/terraform-provider-databricks/blob/master/docs/resources/sql_endpoint.md - how else users will know about it? :)

[nfx]

Please split Wrapper into VisualizationAPI, QueryAPI, WidgetAPI structs and move entities and API wrappers to respective resource files, e.g. resource_widget.go will have NewWidgetsAPI factory, WidgetsAPI struct, WidgetEntity terraform entity for reflection, Widget api entity, and ResourceWidget. The main reason for this is that we have 82%+ code coverage requirement for all code that is written and gotestsum calculates coverage on the package level. I've figured it, when i was confident that reflect_resource is pretty well tested by other packages, but the coverage report didn't show any good results for it. api package coverage is missing from the report, meaning that there are long-term risks of changes going untested to production. This may not be ideal, but it is consistent with the rest of the codebase. Even the special package, dedicated to Accounts API only, follows the same conventions as packages that create resources within the workspace. We may reconsider package layout conventions if we want to have databricks-go-sdk and refactor all 7 packages in a separate pull request, adjusting resource.RetryContext along the way.

[nfx]

acceptance tests should have TestPreviewAcc prefix, so that we can test those separately.

[pietern]

Output of make test-preview:

preview
---
 * [ ] TestPreviewAccDashboard (593.86s)
 * [ ] TestPreviewAccPipelineResource_CreatePipeline (2.15s)
 * [ ] TestPreviewAccQuery (865.81s)
 * [ ] TestPreviewAccSQLEndpoint (519.09s)
 * [x] sqlanalytics.TestPreviewAccSQLEndpoints (443.40s)

The pipeline failure is a known problem (erroneous template).

The SQLA failures are all caused by a non-empty plan for the SQL endpoint:

    acceptance.go:160: Step 1/1 error: After applying this test step, the plan was not empty.
        stdout:


        An execution plan has been generated and is shown below.
        Resource actions are indicated with the following symbols:
          ~ update in-place

        Terraform will perform the following actions:

          # databricks_sql_endpoint.this will be updated in-place
          ~ resource "databricks_sql_endpoint" "this" {
              - auto_stop_mins   = 120 -> null
                id               = "024de5b61abeb9c3"
              - min_num_clusters = 1 -> null
                name             = "tf-thcbecidgbdii"
              - num_clusters     = 1 -> null
                # (6 unchanged attributes hidden)


              - tags {
                }
                # (1 unchanged block hidden)
            }

        Plan: 0 to add, 1 to change, 0 to destroy.