`missing_stderr` and `missing_sample_size` in `jhu-csse` reporting incorrect values

[dshemetov]

Working on JIT A/B tests cmu-delphi/delphi-epidata#947, I found this.

Actual Behavior:

As an example, query the following
https://delphi.cmu.edu/epidata/covidcast/?signal=jhu-csse:confirmed_incidence_num&time=day:20200220&geo_type=county&geo_values=02100

This is an instance of missing_stderr and missing_sample_size reporting code 0 (NOT_MISSING), which should really be code 1 (NOT APPLICABLE).

As another example, query the following
https://delphi.cmu.edu/epidata/covidcast/?signal=jhu-csse:confirmed_incidence_num&time=day:20200123&geo_type=county&geo_values=02100

This is an instance of missing_stderr and missing_sample_size reporting code 5 (OTHER), which should really be code 1 (NOT APPLICABLE).

I don't know how prevalent this is, but for this geo, both cases were very common.

Expected behavior

We should update all derived signals' missing_* fields to be consistent. This should just be a post-processing step after the derived values are computed.

Two options:

1. fix in live code
2. wait for this to be auto-fixed by JIT