Fix RadMon time series plot problem

[EdwardSafford-NOAA]

There are two problems with some of the RadMon time series plots:

1. The time plot for the sdv for mean correction is often all 0 values. The sdv values are created by the extraction executable so the error must be occurring there.
2. Some time series plots do not use the user specified number of cycles but rather always plot 30 day intervals.

I'll tag @HaixiaLiu-NOAA so you will be notified automatically of my findings and progress on this.

[EdwardSafford-NOAA]

I think I've gotten a handle on the problem. Two things seem to be going on. First, the text file that is generated for the javascript is formatting the sdv values as F9.3. Many of the sdv values are quite small and limiting the decimal points to 3 ends up rounding the value to 0.

But, when I change the format to better support the data, the javascript runs into a load error parsing the date field of the data file. Why that becomes a problem with a change in format for the data fields is not yet clear. There is something quite subtle going on.

[EdwardSafford-NOAA]

I've just issued a PR to fix the sdv plots. There are still some issues related to the data values. The values are small enough that the mouse-over legend often results in a value of 0 being displayed when the plotted curve clearly indicates the value is non-zero. I might be able to tweak the javascript and get that to display better.

Longer term it's worth considering if the time series plots should revert to static plots. Python is much more adept at formatting values than is javascript. The trade off is that the javascript readily supports on-the-fly comparisons.