Variable's scale_factor and add_offset are not taken into account during reading

[laclouis5]

For instance a variable with storage type SHORT and with defined scale_factorand add_offsetis not interpreted as FLOAT during reading, which is not the expected behavior. As a comparison, the Python package NetCDF4 does the right conversion during reading.

[visr]

You're right, they are not taken into account. I'll close this as a duplicate of #39. There is also some discussion here: #131 (comment). I think the options are right now to either do it manually using this package, or use https://github.com/Alexander-Barth/NCDatasets.jl.