CompatHelper: bump compat for Turing to 0.38 for package ppl, (keep existing compat) (#425)

* CompatHelper: bump compat for Turing to 0.38 for package ppl, (keep existing compat)

* Update turing.jl

---------

Co-authored-by: CompatHelper Julia <[email protected]>
Co-authored-by: Hong Ge <[email protected]>

Author: 3 people

test/ppl/Project.toml

@@ -6,5 +6,5 @@ Turing = "fce5fe82-541a-59a6-adf8-730c64b5f9a0"
 [compat]
 AbstractGPs = "0.5"
 Distributions = "0.25"
-Turing = "0.24, 0.25 - 0.29, 0.30"
+Turing = "0.24, 0.25 - 0.29, 0.30, 0.38"
 julia = "1.6"

test/ppl/turing.jl

@@ -35,7 +35,7 @@
             f = GP(Matern32Kernel())
             u ~ f(X)
             λ = exp.(u)
-            return y .~ Poisson.(λ)
+            return y ~ product_distribution([Poisson(λ_i) for λ_i in λ])
         end
         m = latent_gp_regression(y, RowVecs(X))
         @test length(Turing.sample(m, Turing.NUTS(), 5)) == 5