Update tests

Author: ConnectedSystems

src/Nodes/Ensembles/WeightedEnsembleNode.jl

@@ -1,10 +1,12 @@
-Base.@kwdef mutable struct WeightedEnsembleNode{N<:NetworkNode, P, A<:Real} <: EnsembleNode
+import ..Analysis: TemporalCrossSection
+
+Base.@kwdef mutable struct WeightedEnsembleNode{N<:NetworkNode,P,A<:Real} <: EnsembleNode
     name::String
     area::A
 
     instances::Array{N} = NetworkNode[]
     weights::Array{P} = [
-        Param(1.0; bounds=[0.0,1.0])
+        Param(1.0; bounds=[0.0, 1.0])
     ]
 
     # Default to normalized weighted sum
@@ -25,8 +27,8 @@ function create_node(
 end
 
 function WeightedEnsembleNode(nodes::Vector{<:NetworkNode}; weights::Vector{Float64},
-                      bounds::Union{Nothing, Vector{Tuple}}=nothing,
-                      comb_method::Union{Nothing, Function}=nothing)::WeightedEnsembleNode
+    bounds::Union{Nothing,Vector{Tuple}}=nothing,
+    comb_method::Union{Nothing,Function}=nothing)::WeightedEnsembleNode
     if isnothing(bounds)
         num_nodes = length(nodes)
         @assert(length(weights) == num_nodes, "Number of nodes do not match provided number of weights")
@@ -39,7 +41,7 @@ function WeightedEnsembleNode(nodes::Vector{<:NetworkNode}; weights::Vector{Floa
     p_weights = [Param(w; bounds=b) for (w, b) in zip(weights, bounds)]
 
     n1 = nodes[1]
-    tmp = WeightedEnsembleNode{NetworkNode, Param, Float64}(;
+    tmp = WeightedEnsembleNode{NetworkNode,Param,Float64}(;
         name=n1.name,
         area=n1.area,
         instances=nodes,
@@ -157,12 +159,12 @@ function calibrate!(
     ensemble::WeightedEnsembleNode,
     climate::Climate,
     calib_data::DataFrame,
-    metric::Union{AbstractDict{String, C}, C};
+    metric::Union{AbstractDict{String,C},C};
     kwargs...
 ) where {C<:Function}
     return invoke(
         calibrate!,
-        Tuple{NetworkNode, Climate, DataFrame, typeof(metric)},
+        Tuple{NetworkNode,Climate,DataFrame,typeof(metric)},
         ensemble,
         climate,
         calib_data,
@@ -208,7 +210,7 @@ function apply_bias_correction(
     period=monthday
 ) where {T<:Real}
     dates = timesteps(climate)
-    tcs = TemporalCrossSection(dates, obs, ensemble.outflow)
+    tcs = TemporalCrossSection(dates, obs, ensemble.outflow; period)
 
     return ensemble.outflow .+ (-tcs.ts)
 end

test/test_calibration.jl

@@ -3,7 +3,7 @@ using Dates, DataFrames, CSV, YAML
 using Streamfall
 
 
-DATA_PATH = joinpath(@__DIR__, "data/hymod")
+DATA_PATH = joinpath(dirname(dirname(pathof(Streamfall))), "test/data/hymod")
 
 @testset "Single node calibration" begin
     # Load and generate stream network
@@ -13,8 +13,8 @@ DATA_PATH = joinpath(@__DIR__, "data/hymod")
     # Load climate data
     date_format = "YYYY-mm-dd"
     obs_data = CSV.File(joinpath(DATA_PATH, "leaf_river_data.csv"),
-                        comment="#",
-                        dateformat=date_format) |> DataFrame
+        comment="#",
+        dateformat=date_format) |> DataFrame
 
     # Column names must be the gauge name
     rename!(obs_data, ["leaf_river_outflow" => "leaf_river"])

test/test_data_op.jl

@@ -6,7 +6,7 @@ using Streamfall
 
 @testset "Data alignment" begin
 
-    here = @__DIR__
+    here = joinpath(dirname(dirname(pathof(Streamfall))), "test")
     climate_data = joinpath(here, "data/campaspe/climate/climate_historic.csv")
     dam_data_loc = joinpath(here, "data/campaspe/gauges")
 

test/test_networks.jl

@@ -7,7 +7,7 @@ using DataFrames
 using Streamfall
 
 
-DATA_PATH = joinpath(@__DIR__, "data/hymod")
+DATA_PATH = joinpath(dirname(dirname(pathof(Streamfall))), "test/data/hymod")
 
 
 @testset "Network loading" begin
@@ -71,7 +71,7 @@ end
 
 
 @testset "Reloading a network spec (HyMod)" begin
-    spec::Dict{String, Union{Dict{String, Any}, Any}} = Streamfall.extract_network_spec(sn)
+    spec::Dict{String,Union{Dict{String,Any},Any}} = Streamfall.extract_network_spec(sn)
     @test spec isa Dict
     @test haskey(spec, "leaf_river") || "Known node does not exist"
 
@@ -97,7 +97,7 @@ end
     )
 
     # Column name must match node name
-    rename!(obs_data, ["leaf_river_outflow"=>"leaf_river"])
+    rename!(obs_data, ["leaf_river_outflow" => "leaf_river"])
     climate_data = obs_data[:, ["Date", "leaf_river_P", "leaf_river_ET"]]
     climate = Climate(climate_data, "_P", "_ET")
 
@@ -109,9 +109,46 @@ end
 
 @testset "Recursing IHACRESNode upstream" begin
     begin
-        include("../examples/run_nodes.jl")
-        # Ensure example does not error out
+        data_path = joinpath(dirname(dirname(pathof(Streamfall))), "test/data/campaspe/")
+
+        # Load and generate stream network
+        sn = load_network("Example Network", joinpath(data_path, "campaspe_network.yml"))
+
+        climate = Climate("../test/data/campaspe/climate/climate.csv", "_rain", "_evap")
+
+        # Historic flows and dam level data
+        calib_data = CSV.read(
+            joinpath(data_path, "gauges", "outflow_and_level.csv"),
+            DataFrame;
+            comment="#"
+        )
 
+        # Historic extractions from the dam
+        extraction_data = CSV.read(
+            joinpath(data_path, "gauges", "dam_extraction.csv"),
+            DataFrame;
+            comment="#"
+        )
+
+        @info "Running example stream..."
+
+        dam_id, dam_node = sn["406000"]
+        Streamfall.run_node!(sn, dam_id, climate; extraction=extraction_data)
+
+        # Extract data for comparison with 1-year burn-in period
+        dam_obs = calib_data[:, "406000"][366:end]
+        dam_sim = dam_node.level[366:end]
+
+        nnse_score = Streamfall.NNSE(dam_obs, dam_sim)
+        nse_score = Streamfall.NSE(dam_obs, dam_sim)
+        rmse_score = Streamfall.RMSE(dam_obs, dam_sim)
+
+        @info "Obj Func Scores:" rmse_score nnse_score nse_score
+
+        nse = round(nse_score, digits=4)
+        rmse = round(rmse_score, digits=4)
+
+        # Ensure example does not error out
         reset!(sn)
         run_basin!(sn, climate; extraction=extraction_data)
         @test Streamfall.RMSE(dam_obs, sn[dam_id].level[366:end]) < 2.5