Apply some basic tests

test/Project.toml

@@ -6,6 +6,7 @@ Accessors = "7d9f7c33-5ae7-4f3b-8dc6-eff91059b697"
 Bijectors = "76274a88-744f-5084-9051-94815aaf08c4"
 Combinatorics = "861a8166-3701-5b0c-9a16-15d98fcdc6aa"
 Compat = "34da2185-b29b-5c13-b0c7-acf172513d20"
+DifferentiationInterface = "a0c0ee7d-e4b9-4e03-894e-1c5f64a51d63"
 Distributed = "8ba89e20-285c-5b6f-9357-94700520ee1b"
 Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
 DistributionsAD = "ced4e74d-a319-5a8a-b0ac-84af2272839c"

test/logdensityfunction.jl

@@ -31,6 +31,18 @@ end
             θ = varinfo[:]
             @test LogDensityProblems.logdensity(logdensity, θ) ≈ logjoint(model, varinfo)
             @test LogDensityProblems.dimension(logdensity) == length(θ)
+
+            # Test a single backend on the generic
+            # ADgradient(::AbstractADType, ::LogDensityFunction) method. This really just
+            # checks that it runs at all.
+            if varinfo isa DynamicPPL.TypedVarInfo
+                ad = ADTypes.AutoMooncake(; config=nothing)
+                ∇ℓ = LogDensityProblemsAD.ADgradient(ad, logdensity)
+                @test isa(
+                    LogDensityProblems.logdensity_and_gradient(∇ℓ, θ),
+                    Tuple{Float64, Vector{Float64}},
+                )
+            end
         end
     end
 end

test/runtests.jl

@@ -4,6 +4,7 @@ using DynamicPPL
 using AbstractMCMC
 using AbstractPPL
 using Bijectors
+using DifferentiationInterface
 using Distributions
 using DistributionsAD
 using Documenter