Project factory additions, project module reuse implementation

[ludoo]

This PR extends the project factory to support additional use cases, and expands its internal context interpolation:

• GCS buckets are added as a project-level primitive (automation buckets are ofc still available)
• interpolation between project and automation level resources, and with the new bucket resource has been extended to cover all possible use cases

This is the supported interpolations, the table has been added to the module docs:

The changes in the project factory require splitting its project handling in two separate module calls to avoid cycles, which surfaced the old issue with data sources in tests. This PR also addresses that, by changing the project module interface so that when a project is reused, the data source can be disabled if the project name and number are known. This is an example from one the updated project module examples:

module "create-project" {
  source          = "./fabric/modules/project"
  billing_account = var.billing_account_id
  name            = "project"
  parent          = var.folder_id
  prefix          = var.prefix
}

module "project" {
  source          = "./fabric/modules/project"
  name            = module.create-project.project_id
  prefix          = var.prefix
  # default behavior, uses a data source internally
  # project_reuse = {}
  # avoid use of a data source when project attributes are available
  project_reuse = {
    use_data_source = false
    project_attributes = {
      name   = module.create-project.name
      number = module.create-project.number
    }
  }

A small change to the tfdoc tool is also part of this PR, to allow optionally skipping specific headings in the generated table of contents.

[wiktorn]

❤️