Fix duplication of mainstem path / headwater HUC / outlet HUC.
In preparing the mainstems data release, I've found that some messiness has been left behind in the mainstem ID, headwater HUC and outlet HUC sets. This doesn't appear as a problem until trying to summarize things on a mainstem path basis when we realize there are duplicates. It appears that the issue is being caused in the core math levelpaths methods.
I started from the workflow output and narrowed things down to the output of par_math_levelpaths. Next steps are to narrow down the component of the workflow with a small-ish sample dataset