Preparation of the parsed mongo::Pipeline seems to be one logical step in the overall run_aggregate path. We should break it into a helper function to clean up the high-level path, whether the input is the aggregation plan and the resolved collation, and the output is the pipeline (which includes the expression context). In summary, this step will make the expression context, parse the pipeline, register query stats, optimize and rewrite the pipeline as necessary, and coordinate small checks along the way.