Failing to run ExpHydro Model Tutorial

[olivierbonte]

When running the tutorial as defined in https://github.com/chooron/HydroModels.jl/blob/35ff1bbd24172db898c701da079d8bbbbd9d2bc7/docs/src/tutorials/run_a_exphydro_model.md, I run into a few issues

• The input data should be mapped to an input matrix, as in

  HydroModels.jl/test/run_lumped_model.jl

  Line 11 in 35ff1bb

  input_mat = Matrix(reduce(hcat, collect(input_ntp[[:temp, :lday, :prcp]]))')

• The object model is not defined. To get this running, one should first follow the building Exphydro tutorial (https://github.com/chooron/HydroModels.jl/blob/35ff1bbd24172db898c701da079d8bbbbd9d2bc7/docs/src/implements/build_exphydro_model_en.md) and then construct the model using model = HydroModel(; name=:exphydro, components=[snowpack_bucket,soilwater_bucket])
• Converting the output to a named tuple as defined in

  HydroModels.jl/docs/src/tutorials/run_a_exphydro_model.md

  Line 88 in 35ff1bb

  result = model(input, params, config=(solver=solver, timeidx=ts), convert_to_ntp=true)

  does not work for me (using v0.1.1.) , although the code does run when setting convert_to_ntp=true.

[chooron]

This is quite embarrassing. This issue is also related to the documentation synchronization problem after code updates. In v0.1.1's design, I set the input and output restrictions for components like bucket and model to be AbstractArray type (the convert_to_ntp setting was therefore removed). If you want to use NamedTuple as input like before, you can wrapper this component through NamedTupleIOAdapter.

[olivierbonte]

No worries, I think these sort of issues are very normal when in early stages of package development! The updates you made to the docs in 7727d7d make things a lot clearer already :) The only mistake I can still spot now is that input is not defined from the dataframe yet for use in

HydroModels.jl/docs/src/tutorials/run_exphydro_model.md

Line 100 in 7727d7d

result = exphydro_model(input, pas, config=(solver=solver, timeidx=ts))

This could be solved by adding

input_ntp = (lday=df[ts, "dayl(day)"], temp=df[ts, "tmean(C)"], prcp=df[ts, "prcp(mm/day)"])
input = Matrix(reduce(hcat, collect(input_ntp[[:temp, :lday, :prcp]]))')

somewhere before

[olivierbonte]

See PR #6

[olivierbonte]

Closed by PR #6