Julia 1.5+ and JuMP 0.21+ compat

Project.toml

@@ -0,0 +1,14 @@
+name = "PSA"
+uuid = "4c80e164-764c-4b93-bf2f-9dd64506ca8b"
+authors = ["Tom Brown <tom.brown@kit.edu>"]
+version = "0.1.0"
+
+[deps]
+CSV = "336ed68f-0bac-5ca0-87d4-7b16caf5d00b"
+Clp = "e2554f3b-3117-50c0-817c-e040a3ddf72d"
+Conda = "8f4d0f93-b110-5947-807f-2305c1781a2d"
+DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
+JuMP = "4076af6c-e467-56ae-b986-b466b2749572"
+LightGraphs = "093fc24a-ae57-5d10-9952-331d41423f4d"
+LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
+PyCall = "438e738f-606a-5dbb-bf0a-cddfbfd45ab0"

examples/ac-dc-lopf.jl

@@ -1,18 +1,17 @@
 
 # Cf. PyPSA example https://github.com/PyPSA/PyPSA/tree/master/examples/ac-dc-meshed
 
-push!(LOAD_PATH, "/home/tom/fias/lib")
-
 import PSA
 
 using Clp
+using JuMP
 
 network = PSA.import_network("/home/tom/fias/lib/pypsa/examples/ac-dc-meshed/ac-dc-data/")
 
-solver = ClpSolver()
+solver = Clp.Optimizer
 
 m = PSA.lopf(network, solver)
 
-print(m.objVal)
+println(objective_value(m))
 
-print(network.generators_t["p"])
+println(network.generators_t["p"])

src/PSA.jl

@@ -146,14 +146,15 @@ function import_network(folder)
     !ispath("$folder") ? error("Path not existent") : nothing
     components = static_components(network)
     for component=components
-        if ispath("$folder/$component.csv")
+        componentpath = joinpath("$folder", "$component.csv")
+        if ispath(componentpath)
             # fallback for missing values
             try
-                setfield!(network,component,CSV.read("$folder/$component.csv"; truestring="True",
-                                                    falsestring="False"))
+                setfield!(network,component,CSV.read(componentpath, DataFrame; truestrings=["True"],
+                                                    falsestrings=["False"]))
             catch y
                 if (typeof(y)==Missings.MissingException) | (typeof(y) == BoundsError)
-                    setfield!(network,component,readtable("$folder/$component.csv"; truestrings=["True"],
+                    setfield!(network,component,readtable(componentpath; truestrings=["True"],
                                                     falsestrings=["False"]))
 
                 end
@@ -164,7 +165,7 @@ function import_network(folder)
         for name in names(df)
             if name == :name
                 nothing
-            elseif typeof(df[name]) == Array{Union{Int64, Missings.Missing},1}
+            elseif typeof(df[!, name]) == Array{Union{Int64, Missings.Missing},1}
                 println("converting column $name of $component from Int to Float")
                 df[name] = float(df[name])
             end
@@ -178,7 +179,7 @@ function import_network(folder)
                 # fallback for missing values for a non-null column type, might be deprecated soon
                 try
                     getfield(network,component_t)[attr]= (
-                    CSV.read("$folder/$component-$attr.csv"; truestring="True", falsestring="False") )
+                    CSV.read("$folder/$component-$attr.csv"; truestrings=["True"], falsestrings=["False"]) )
                 catch y
                     if (typeof(y)==Missings.MissingException) | (typeof(y) == BoundsError)
                         getfield(network,component_t)[attr]= (
@@ -189,7 +190,7 @@ function import_network(folder)
         end
     end
     initializer = Network()
-    for field=setdiff(fieldnames(network), fieldnames(initializer))
+    for field=setdiff(fieldnames(typeof(network)), fieldnames(typeof(initializer)))
         setfield!(network, field, getfield(initializer, field))
     end
     return network

src/auxilliaries.jl

@@ -1,11 +1,11 @@
 using LightGraphs
 using PyCall
 const networkx = PyNULL()
-copy!(networkx, pyimport("networkx" ))
+copy!(networkx, pyimport("networkx"))
 
 
 function time_dependent_components(network)
-    fields = String.(fieldnames(network))
+    fields = String.(fieldnames(typeof(network)))
     components = []
     for field=fields
         if field[end-1:end] == "_t"
@@ -16,7 +16,7 @@ function time_dependent_components(network)
 end
 
 function static_components(network)
-    fields = String.(fieldnames(network))
+    fields = String.(fieldnames(typeof(network)))
     components = []
     for field=fields
         if field[end-1:end] != "_t"
@@ -41,26 +41,26 @@ end
 function align_component_order!(network)
     components_t = time_dependent_components(network)
     for comp=components_t
-        order = Symbol.(getfield(network, Symbol(String(comp)[1:end-2]))[:name])
+        order = Symbol.(getfield(network, Symbol(String(comp)[1:end-2]))[!, :name])
         for attr in keys(getfield(network, comp))
-            if length(getfield(network,comp)[attr])==length(order)
+            if nrow(getfield(network,comp)[attr])==length(order)
                 getfield(network,comp)[attr]= getfield(network,comp)[attr][:, order]
             end
         end
     end
 end
 
 # auxilliary funcitons
-idx(dataframe) = Dict(zip(dataframe[:name], Iterators.countfrom(1)))
-rev_idx(dataframe) = Dict(zip(Iterators.countfrom(1), dataframe[:name]))
-idx_by(dataframe, col, values) = select_by(dataframe, col, values)[:idx]
+idx(dataframe) = Dict(zip(dataframe[!, :name], Iterators.countfrom(1)))
+rev_idx(dataframe) = Dict(zip(Iterators.countfrom(1), dataframe[!, :name]))
+idx_by(dataframe, col, values) = select_by(dataframe, col, values)[!, :idx]
 
 function select_by(dataframe, col, selector)
-    if length(findin(dataframe[col], selector))==0
+    if length(findall(in(selector), dataframe[!, col]))==0
         return dataframe[repeat(Bool[false],outer=nrow(dataframe)) , :]
     else
-        mdict = Dict(zip(dataframe[col], Iterators.countfrom(1)))
-        ids = Array{Int,1}(0)
+        mdict = Dict(zip(dataframe[!, col], Iterators.countfrom(1)))
+        ids = Array{Int,1}()
         for i in selector
             push!(ids, mdict[i])
         end
@@ -73,10 +73,10 @@ select_names(a, b) = select_by(a, :name, b)
 function append_idx_col!(dataframe)
     if typeof(dataframe)==Vector{DataFrames.DataFrame}
         for df in dataframe
-            df[:idx] = collect(1:nrow(df))
+            df[!, :idx] = collect(1:nrow(df))
         end
     else
-        dataframe[:idx] = collect(1:nrow(dataframe))
+        dataframe[!, :idx] = collect(1:nrow(dataframe))
     end
 end
 
@@ -89,16 +89,16 @@ function get_switchable_as_dense(network, component, attribute, snapshots=0)
     if in(attribute, keys(getfield(network, component_t)))
         dense = getfield(network, component_t)[attribute]
     end
-    cols = Symbol.(getfield(network, component)[:name])
+    cols = Symbol.(getfield(network, component)[!, :name])
     not_included = String.(setdiff(cols, names(dense)))
     if length(not_included)>0
         attribute = Symbol.(attribute)
         df = select_names(getfield(network, component), not_included)
-        df = names!(DataFrame(repmat(transpose(Array(df[attribute])), T)),
+        df = rename!(DataFrame(repeat(transpose(Array(df[!, attribute])), T)),
                 Symbol.(not_included))
         dense = [dense df]
     end
-    return dense[cols]
+    return dense[!, cols]
 end
 
 
@@ -128,7 +128,7 @@ end
 
 function calculate_dependent_values!(network)
     function set_default(dataframe, col, default)
-        !in(col, names(dataframe)) ? dataframe[col] = default : nothing
+        !(col in propertynames(dataframe)) ? insertcols!(dataframe, col .=> default) : nothing
     end
 
     #buses
@@ -138,7 +138,7 @@ function calculate_dependent_values!(network)
     end
 
     # generators
-    defaults = [(:p_nom_extendable, false), (:p_nom_max, Inf),(:commitable, false),
+    defaults = [(:p_nom_extendable, false), (:committable, false),
                 (:p_min_pu, 0), (:p_max_pu, 1), (:p_nom_min, 0),(:capital_cost, 0),
                 (:min_up_time, 0), (:min_down_time, 0), (:initial_status, true),
                 (:p_nom, 0.),(:marginal_cost, 0),(:p_nom_opt, 0.)]
@@ -147,16 +147,18 @@ function calculate_dependent_values!(network)
     end
 
     # lines
-    network.lines[:v_nom]=select_names(network.buses, network.lines[:bus0])[:v_nom]
+    network.lines[!, :v_nom]=select_names(network.buses, network.lines[!, :bus0])[!, :v_nom]
     defaults = [(:s_nom_extendable, false), (:s_nom_min, 0),(:s_nom_max, Inf), (:s_nom, 0.),
                 (:s_nom_min, 0), (:s_nom_max, Inf), (:capital_cost, 0), (:g, 0)]
     for (col, default) in defaults
         set_default(network.lines, col, default)
     end
-    network.lines[:x_pu] = network.lines[:x]./(network.lines[:v_nom].^2)
-    network.lines[:r_pu] = network.lines[:r]./(network.lines[:v_nom].^2)
-    network.lines[:b_pu] = network.lines[:b].*network.lines[:v_nom].^2
-    network.lines[:g_pu] = network.lines[:g].*network.lines[:v_nom].^2
+    pu(a, b) = a ./ b.^2
+    pu_inverse(a, b) = a .* b.^2
+    transform!(network.lines, [:x, :v_nom] => pu => :x_pu)
+    transform!(network.lines, [:r, :v_nom] => pu => :r_pu)
+    transform!(network.lines, [:b, :v_nom] => pu_inverse => :b_pu)
+    transform!(network.lines, [:g, :v_nom] => pu_inverse => :g_pu)
 
     # links
     defaults = [(:p_nom_extendable, false), (:p_nom_max, Inf), (:p_min_pu, 0),
@@ -187,7 +189,7 @@ function calculate_dependent_values!(network)
     # loads_t
     for df_name=keys(network.loads_t)
         if nrow(network.loads_t[df_name])>1
-            for bus=[bus for bus in network.loads[:name] if
+            for bus=[bus for bus in network.loads[!, :name] if
                 !in(Symbol(bus), names(network.loads_t[df_name]))]
                 set_default(network.loads_t[df_name], bus, 0)
             end
@@ -238,6 +240,7 @@ function ptdf_matrix(network)
 end
 
 function get_cycles(network)
+    copy!(networkx, pyimport("networkx"))
     busidx = idx(network.buses)
     g = networkx[:Graph]()
     g[:add_nodes_from](busidx)