Address a few bugs

[dshemetov]

Closes #69.
Closes #66.
Closes #65.
Closes #67.

On the last issue: request params get formatted differently depending on whether a vector or a list is handed as an argument. This is an attempt to make the outcome uniform. The issue is

r$> paste(sapply(c("290", "190"), toString), collapse = ",")
[1] "290,190"

r$> paste(sapply(list(c("290", "190")), toString), collapse = ",")
[1] "290, 190"

So I added some logic to check for vectors and use as.list instead. The problem is that the is.vector also yields TRUE for lists, so it's currently messing up the list-type handling. Would appreciate help/commits there.

[brookslogan]

The function you wanted was probably is.atomic. But I wonder if we can keep the logic a bit simpler, back out your change, and instead fix toString to be paste, collapse=",". And maybe also use vapply instead of sapply (FUN.VALUE would be character(1L)).

[brookslogan]

Looks good, though hope we don't depend on int zips/fips somewhere already. Committed the approach I described above. Do you remember what request revealed the fip_code typo? Wonder if we're missing some tests.

[dshemetov]

Thanks Logan! I don't recall what revealed the fip_code typo. To say that we're missing tests is an understatement: those two functions format_list and format_item are the only functions that have any tests in this repo at all.

Just going to merge and leave making unit tests to another PR.