Add proper handling for version collisions in differing time ranges

Replicate how Python handles files that have the same dataset identifier but the same version. In this case it considers instances where it finds a version which is higher than the already recorded latest as a warning but continues with the same behavior as before. This shouldn't happen in most datasets due to the ordering we apply but I believe it may be possible in custom datasets in which case ingestion may not work properly. It will also properly handle instances where files with the same identifier and version are encountered. This is hopefully only the case where files which have the same attributes but differ in their time ranges and should be included in latest as separate files.