Wrong values matched to measurement variables in multipointcoverage query

[Mikkolehtimaki]

When using the python interface, the output matches wrong names of measurement variables to measurement results. This can be seen for example when snow depth has negative values or relative humidity might is not between 0-100.

Run for example the command

./fmiopendata --stored_query=fmi::observations::weather::multipointcoverage --bbox=19,59,20,60 --start=2018-08-08T14:00:00Z --end=2018-08-08T14:00:00Z data

and the result shows snow depth: 12.3mm in August 2018.

I tracked the cause to fmiopendata.parse_multipointcoverage() in fmiopendata.py. It assumes a certain order of parameters when it parses the xml response. It seems to need the output of fmiopendata.get_parameters() to return a dict with value names corresponding to the indices of measurement time series. However, the order of value names returned from get_parameters() seems to vary, causing mismatch of variable-value pairs.