With the LiteParsedDesugarer refactor, we will have some pipeline parses being used only for query shape. This is because we will then desugar the lite parsed pipeline and "re-parse" the pipeline, this time for optimization/execution.

On this first parse for query shape only, it will be necessary to skip incrementing certain counters and calls, such as planShardedSearch. As such, it's necessary to track the "type" of parse in expCtx.