Allow overriding/specifying Resource Identity values in Import checks

### terraform-plugin-testing version

```
github.com/hashicorp/terraform-plugin-testing v1.16.0
```

### Use cases

In some cases, importing should allow un-normalized values in a Resource Identity attribute while the RI itself would have a normalized value.

For example, in `aws_route53_record`, the `name` attribute for the resource can be subdomain-only or a fully-qualified domain name with or without trailing dot. When using an `import` block, the `name` in the Resource Identity should also support these forms.

Currently, testing `resource.ImportBlockWithResourceIdentity` directly uses the Resource Identity from the resource. It is not possible to test the other forms.

### Proposal

A mechanism similar to the existing `ImportStateIdFunc` could be used to specify a value for the Resource Identity. Some option to modify the existing Resource Identity would be useful, especially if some of the identifying attributes are `Computed`-only.

For example, for `aws_route53_record`, there are four attributes `zone_id`, `name`, `type`, and `set_identifier`. We would want to use the existing `zone_id`, `type`, and `set_identifier`, but be able to de-normalize the `name`. `zone_id` is a `Computed`-only attribute on the Record's parent resource.

Start by reading resource.ImportBlockWithResourceIdentity and the existing ImportStateIdFunc mechanism. Trace how Resource Identity values are obtained, including Computed-only attributes, and determine the API shape needed to override selected values. Done means import checks can exercise alternate, un-normalized identity values while preserving the other identity attributes.

Written by the indexing model from the issue text.