Improve method of reading parameters with epidist_db()

[joshwlambert]

This is a huge file. I understand the functionality of most of the code, but due to the length I may have missed potential issues - other reviewers, please also give this a look.

What bothers me is that nearly all of the code is being called for every epidist_db() call. I can't help but feel that there must be a better way to read from the database than to read the JSON each time and format entries into <epidist>.

One possibility is to filter the nested list of JSON data before formatting it.

Another option is to store the parameter library as package data, with the list of <epidist>s as a list-column in a data.frame that has a limited number of columns, such as the author, year, disease, and sample size. The JSON file could be an intermediate format for external contributions. This would make the data more R specific, so might not be the direction to go.

Originally posted by @pratikunterwegs in #197 (comment)

[joshwlambert]

This issue is mostly a duplicate of #198. Closing this issue as it is addressed in #207.