False positives when testing for ingestion completion

ap_verify tests whether particular types of data have been ingested, and only attempts ingestion if they have not. However, the checks currently address whether ingestion began, not whether it finished successfully. This may cause problems if ingestion fails partway through, or if the order in which datasets are ingested is changed during future refactoring.

Either the ingestion checks need to be rewritten to directly test ingestion completion, or IngestTask needs better handling of duplicate files.