Various bug fixes

[kaitejohnson]

This PR addresses a few semi-related issues including:

• improving the error messaging if a user provides hospital admissions data of insufficient length. This closes #99
• modified the get_draws_df() function to explicitly specify the column names to be joined to the posterior draws. This means that if there are extra columns being passed in as part of the input data for either wastewater or count datasets, they will not be returned as part of the stnadardized output of get_draws_df(). I think it is reasonable that extra metadata can be joined to this table in downstream, user-specific post-processing.
• modifying the package data to include an additional column location in each of the example data sets (ww_data and hosp_data. This was intended to recreate and test the bug described in #102.
• modifying the package ww_data so that sites are not orderd 1:n_sites but instead are in reverse order and don't start at 1. This was to try to recreate and test the bug described in #100. However, I wasn't able to reproduce this error.

This PR branches off of #108, and shouldn't be reviewed until that is merged.

[kaitejohnson]

Some key fixes and suggestions. Also, is there a unit test for the internal reordering?

modifying the package ww_data so that sites are not orderd 1:n_sites but instead are in reverse order and don't start at 1. This was to try to recreate and test the bug described in #100. However, I wasn't able to reproduce this error.

A unit test to check that the reordering worked as expected? I can add to the preprocessing tests. Note in #158 we will be setting the order of the populations to descending order by site population and I added a unit test to confirm that works as expected