Update NetCDFOutputWriter docstring and clobber default

[ali-ramadhan]

Docstring did not match the actual constructor. They do now.

@suyashbire1 Do you think clobber should be true or false by default?

Also updated the docstring to make it clear that outputs needs to be a Dict for now.

Resolves #553

[codecov]

Codecov Report

Merging #556 into master will increase coverage by 3.74%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #556      +/-   ##
==========================================
+ Coverage   67.72%   71.47%   +3.74%     
==========================================
  Files          69       69              
  Lines        1952     2033      +81     
==========================================
+ Hits         1322     1453     +131     
+ Misses        630      580      -50

Impacted Files	Coverage Δ
src/OutputWriters/netcdf_output_writer.jl	86.53% <100%> (ø)	⬆️
src/utils.jl	75.56% <0%> (+3.99%)	⬆️
src/Diagnostics/cfl.jl	71.42% <0%> (+4.76%)	⬆️
src/Operators/interpolation_operators.jl	64.28% <0%> (+4.76%)	⬆️
src/Operators/derivative_operators.jl	90.62% <0%> (+32%)	⬆️
...lementations/anisotropic_biharmonic_diffusivity.jl	100% <0%> (+50%)	⬆️
...ure_implementations/leith_enstrophy_diffusivity.jl	98.38% <0%> (+66.12%)	⬆️
src/Operators/laplacian_operators.jl	100% <0%> (+85.71%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a511fdd...e187583. Read the comment docs.

[glwagner]

clobber=false is probably safer.

[glwagner]

why can't we handle named tuples... ?

[ali-ramadhan]

Would be easy to add support for named tuple. I just removed them from the docstring because of #553

We can add support for named tuples in this PR as well.

[glwagner]

Named tuples are trivial to support; but maybe users don't care. Personally, I much prefer them in the JLD2OutputWriter.

[ali-ramadhan]

clobber=false is probably safer.

That's what I was thinking too but apparently with NCDatasets.jl the other modes are append ("a") and read-only ("r") so if you're creating files for the first time (the most common use case) then clobber must be used. Otherwise you get NetCDF errors when it tries to open a non-existent file.

Named tuples are trivial to support; but maybe users don't care. Personally, I much prefer them in the JLD2OutputWriter.

I actually wasn't able to get named tuples to work with JLD2 but I opened an issue about that: #562

I don't see the big advantage of a NamedTuple over a Dict in this case so I'm going to keep it simple and stick to named tuples here for now. The original point of this PR was that the documentation was wrong.

NetCDFOutputWriter needs to be refactored a little anyways, and I'd like to add some features to it as well so I'll add support for named tuples then.