CDAWeb DE1_62MS_MAGA-GMS fails because of availability

[jbfaden]

I found where DE1_62MS_MAGA-GMS fails because it stops searching granules too early. In https://cottagesystems.com/~jbf/hapi/p/cdaweb//data/cdfmetafile/info/DE1_62MS_MAGA-GMS.json, there's an entry:

      {
        "Name": "https://cdaweb.gsfc.nasa.gov/sp_phys/data/de/de1/magnetic_fields_maga/62ms_magagms_cdaweb/1983/de1_62ms_maga-gms_19831220_v01.cdf",
        "StartTime": "1983-12-20T10:10:19Z",
        "EndTime": "1991-12-20T23:35:11Z"
      },

where all other granules are one day long. This causes the code to scan through all granules to find the one example day in 1991, causing timeouts.

[jbfaden]

I think what I will do is to detect this and drop this day.

[jbfaden]

I put in a test for overlaps, and there are quite a few that overlap. I think I might need to detect non-monotonic datasets and sort them.