Use circo to display composition diagrams

docs/src/ebm_melt/ebm_melt.md

@@ -107,7 +107,7 @@ ice_dynamics = apex(oapply(ice_dynamics_composition_diagram,
   [Open(halfar_eq2, [:Γ,:n]),
    Open(glens_law, [:Γ,:n])]))
 
-to_graphviz(ice_dynamics_composition_diagram)
+draw_composition(ice_dynamics_composition_diagram)
 ```
 
 The composition pattern tells you how to couple the variables and introduces namespaces that we will use later when supplying initial conditions.
@@ -166,7 +166,7 @@ budyko_sellers = apex(oapply(budyko_sellers_composition_diagram,
    Open(heat_transfer, [:Tₛ, :HT, :cosϕᵖ]),
    Open(insolation, [:Q, :cosϕᵖ])]))
 
-to_graphviz(budyko_sellers_composition_diagram)
+draw_composition(budyko_sellers_composition_diagram)
 ```
 
 Our full model can then be composed by adding terms for melting of water. We will assume that the meltwater is transported by diffusion because the transport of meltwater is so much faster than the melting process itself. If you wanted to increase the physical realism of this model, using a different model of melting and water transport would be a good place to start.
@@ -194,7 +194,7 @@ budyko_sellers_halfar_water_composition_diagram = @relation () begin
   halfar(A, h, melt)
 end
 
-to_graphviz(budyko_sellers_halfar_water_composition_diagram)
+draw_composition(budyko_sellers_halfar_water_composition_diagram)
 ```
 
 ``` @example DEC
@@ -256,7 +256,7 @@ u₀ = ComponentArray(
   h = h₀,
   melting_water = water)
 
-# the underscore sparate words are the namespaces that were introduced with oapply.
+# The underscore-separated words are the namespaces that were introduced by oapply.
 constants_and_parameters = (
   budyko_sellers_absorbed_radiation_α = α,
   budyko_sellers_outgoing_radiation_A = A,
@@ -271,7 +271,6 @@ constants_and_parameters = (
   melting_Dₕ₂ₒ = Dₕ₂ₒ)
 
 # Define how symbols map to Julia functions
-
 function generate(sd, my_symbol; hodge=GeometricHodge())
   op = @match my_symbol begin
     :♯ => begin