get_weight with AbstractEdge (#46)

Quick PR to allow `get_weight` to be called with any `AbstractEdge`
using the Interface defined in `Graphs`. This fixes my issue #45.
(assuming it is an issue and not a conscious design decision.)

Author: henrik-wolf

docs/src/tutorial.md

@@ -28,6 +28,13 @@ julia> get_weight(g, 1, 2)
 0.5
 ```
 
+Or by passing in any `e::AbstractEdge`:
+
+```jldoctest tuto
+julia> get_weight(g, Edge(1, 2))
+0.5
+```
+
 Find the shortest path from vertex 1 to vertex 3, taking weights into account:
 
 ```jldoctest tuto

src/abstractsimpleweightedgraph.jl

@@ -28,10 +28,14 @@ weighttype(::AbstractSimpleWeightedGraph{T,U}) where {T} where {U} = U
 
 """
     get_weight(g, u, v)
+    get_weight(g, e)
 
-Retrieve the weight of edge `(u, v)`.
+Retrieve the weight of edge `(u, v)` or `e`.
 """
 get_weight(g::AbstractSimpleWeightedGraph, u::Integer, v::Integer) = weights(g)[u, v]
+function get_weight(g::AbstractSimpleWeightedGraph, e::AbstractEdge{<:Integer})
+    return weights(g)[src(e), dst(e)]
+end
 
 ## Vertices
 

src/overrides.jl

@@ -194,7 +194,7 @@ function Graphs.induced_subgraph(
     for e in elist
         if has_edge(g, e)
             i, j = index_map[src(e)], index_map[dst(e)]
-            w = get_weight(g, dst(e), src(e))
+            w = get_weight(g, e)
             push!(I, j)  # storage is transposed!
             push!(J, i)
             push!(W, w)

test/simpleweightedgraph.jl

@@ -204,6 +204,9 @@ using SparseArrays
     @test sum(weights(g)) == 2 * ne(g) * 3
     @test @inferred(get_weight(g, 1, 2)) == 3
 
+    @test @inferred(get_weight(g, Edge(1, 2))) == 3
+    @test @inferred(get_weight(g, SimpleWeightedEdge(1, 2, 0.5))) == 3
+
     g = SimpleWeightedDiGraph(path_graph(5), 4.0)
     @test sum(weights(g)) == ne(g) * 4.0
 
@@ -294,12 +297,18 @@ using SparseArrays
 
             @test g[1, 2, Val{:weight}()] ≈ 5
             @test get_weight(g, 1, 2) ≈ 5
+            @test get_weight(g, Edge(1, 2)) ≈ 5
+            @test get_weight(g, SimpleWeightedEdge(1, 2)) ≈ 5
             if is_directed(G)
                 @test g[2, 1, Val{:weight}()] ≈ 0
                 @test get_weight(g, 2, 1) ≈ 0
+                @test get_weight(g, Edge(2, 1)) ≈ 0
+                @test get_weight(g, SimpleWeightedEdge(2, 1)) ≈ 0
             else
                 @test g[2, 1, Val{:weight}()] ≈ 5
                 @test get_weight(g, 2, 1) ≈ 5
+                @test get_weight(g, Edge(2, 1)) ≈ 5
+                @test get_weight(g, SimpleWeightedEdge(2, 1)) ≈ 5
             end
             m = adjacency_matrix(g)
             @test g[2, 1, Val{:weight}()] ≈ g.weights[1, 2]