Client Library Generation

AutoRest Configuration

When generating the AutoRest libraries, there are a few transformation made to the OpenAPI JSON in the form of a "literate configuration" .yaml file.
More details can be found here: http://azure.github.io/autorest/user/literate-file-formats/configuration.html

Here are the transformation made (with some information on why):

Skip URL encoding on the relativePath query parameter - this transformation is made to allow for OData query parameters to be passed via the relativePath parameter without issue (e.g. Customer&$top=10)

- from: swagger-document
  where: $..parameters[?(@.name == "relativePath")]
  transform: $['x-ms-skip-url-encoding'] = true

Add a missing item-type for a identifiers array query parameter that was not retaining the value 

- from: swagger-document
  where: $..parameters[?(@.name == "identifiers")].schema
  transform: |
    $.items = { 'type': 'string' }

Remove a redundant API operation (this operation is identical in functionality to Get-entities-with-OData-query-parameters, and we are able to control the URL encoding of that operation. The difference is only in format of the URI

- from: swagger-document
  where: $..paths[?(@.get.operationId == "Get-entities-with-OData-path")]
  transform: $ = null

AutoRest does not recognize none as a valid value for collectionFormat - per the OpenAPI specification, csv is the default: https://swagger.io/docs/specification/2-0/describing-parameters/#array. During configuration this is enforced.

- from: code-model-v1
  where: $..[?(@.collectionFormat == "none")].collectionFormat
  transform: $ = 'csv'

In the config, we specify generation of Java, Python, C# (csharp), and NodeJS libraries.