Added Aqua.jl to tests

Project.toml

@@ -7,7 +7,6 @@ version = "0.4.0"
 CUDA = "052768ef-5323-5732-b1bb-66c8b64840ba"
 ChainRulesCore = "d360d2e6-b24c-11e9-a2a3-2a2ae2dbcce4"
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
-DifferentialEquations = "0c46a032-eb83-5123-abaf-570d42b7fbaa"
 Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
 GraphNetCore = "7809f980-de1b-4f9a-8451-85f041491431"
 HDF5 = "f67ccb44-e63f-5c2f-98bd-6dc0ccc4ba2f"
@@ -16,7 +15,7 @@ JSON = "682c06a0-de6a-54ab-a142-c8b1cf79cde6"
 Lux = "b2108857-7c20-44ae-9111-449ecde12c47"
 LuxCUDA = "d0bbae9a-e099-4d5b-a835-1c6931763bda"
 Optimisers = "3bd65402-5787-11e9-1adc-39752487f4e2"
-Optimization = "7f7a1694-90dd-40f0-9382-eb1efda571ba"
+OrdinaryDiffEq = "1dea7af3-3e70-54e6-95c3-0bf5283fa5ed"
 Printf = "de0858da-6303-5e67-8744-51eddeeeb8d7"
 ProgressMeter = "92933f4c-e287-5a05-a399-4b506db050ca"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
@@ -29,10 +28,10 @@ Zygote = "e88e6eb3-aa80-5325-afca-941959d7151f"
 cuDNN = "02a925ec-e4fe-4b08-9a7e-0d78e3d38ccd"
 
 [compat]
+Aqua = "0.8"
 CUDA = "5"
 ChainRulesCore = "1.16.0 - 1"
 DataFrames = "1.6"
-DifferentialEquations = "7.11 - 7"
 Distributions = "0.25"
 GraphNetCore = "0.3"
 HDF5 = "0.17"
@@ -41,21 +40,23 @@ JSON = "0.21"
 Lux = "0.5"
 LuxCUDA = "0.3"
 Optimisers = "0.3"
-Optimization = "3.18"
+OrdinaryDiffEq = "6.85 - 6"
 Printf = "1"
 ProgressMeter = "1.7.0 - 1"
 Random = "1"
 SciMLBase = "2.7.0 - 2"
 SciMLSensitivity = "7.45 - 7"
 Statistics = "1"
 TFRecord = "0.4.2"
+Test = "1"
 Wandb = "0.5"
 Zygote = "0.6"
 cuDNN = "1.3"
 julia = "1.10"
 
 [extras]
+Aqua = "4c88cf16-eb10-579e-8560-4a9242c79595"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [targets]
-test = ["Test"]
+test = ["Aqua", "Test"]

README.md

@@ -3,6 +3,7 @@
 # MeshGraphNets.jl
 
 [![Docs](https://img.shields.io/badge/docs-dev-blue.svg)](https://una-auxme.github.io/MeshGraphNets.jl/dev)
+[![Aqua QA](https://raw.githubusercontent.com/JuliaTesting/Aqua.jl/master/badge.svg)](https://github.com/JuliaTesting/Aqua.jl)
 
 [*MeshGraphNets.jl*](https://github.com/una-auxme/MeshGraphNets.jl) is a software package for the Julia programming language that provides an implementation of the [MeshGraphNets](https://arxiv.org/abs/2010.03409) framework by [Google DeepMind](https://deepmind.google/) for simulating mesh-based physical systems via graph neural networks:
 

examples/cylinder_flow/cylinder_flow.jl

@@ -5,7 +5,7 @@
 
 using MeshGraphNets
 
-import DifferentialEquations: Euler, Tsit5
+import OrdinaryDiffEq: Euler, Tsit5
 import Optimisers: Adam
 
 ######################

src/MeshGraphNets.jl

@@ -13,14 +13,14 @@ using Optimisers
 using Wandb
 using Zygote
 
-import DifferentialEquations.OrdinaryDiffEq: OrdinaryDiffEqAlgorithm, Tsit5
+import OrdinaryDiffEq: OrdinaryDiffEqAlgorithm, Tsit5
 import ProgressMeter: Progress
 import SciMLBase: ODEProblem
 
 import Base: @kwdef
-import DifferentialEquations: solve, remake
 import HDF5: h5open, create_group, open_group
 import ProgressMeter: next!, update!, finish!
+import SciMLBase: solve, remake
 import Statistics: mean
 
 include("utils.jl")

src/dataset.jl

@@ -4,10 +4,11 @@
 #
 
 import Distributions: Normal
+import HDF5: Group
 import Random: MersenneTwister
 import TFRecord: Example
 
-import HDF5: h5open, Group, read_dataset
+import HDF5: read_dataset
 import JLD2: jldopen
 import JSON: parse
 import Random: seed!, make_seed, shuffle

src/graph.jl

@@ -69,11 +69,20 @@ Constructs a [FeatureGraph](@ref) based on the given arguments.
 - Resulting [FeatureGraph](@ref).
 """
 function build_graph(mgn::GraphNetwork, data, fields, datapoint::Integer, node_type, edge_features::AbstractArray{Float32, 2}, senders::AbstractArray{T, 1}, receivers::AbstractArray{T, 1}) where {T <: Integer}
+    # Removed generator in favor of removing Zygote.jl piracies (minimal increase of time and allocations)
+    # Can be reverted once Enzyme.jl is compatible
+    nt = mgn.n_norm["node_type"](node_type)
+    nf = similar(nt, 0, size(nt, 2))
+    for field in fields
+        nf = vcat(nf, mgn.n_norm[field](data[field][:, :, min(size(data[field], 3), datapoint)]))
+    end
+    nf = vcat(nf, nt)
     return FeatureGraph(
-        vcat(
-            [mgn.n_norm[field](data[field][:, :, min(size(data[field], 3), datapoint)]) for field in fields]...,
-            mgn.n_norm["node_type"](node_type)
-        ),
+        nf,
+        # vcat(
+        #     [mgn.n_norm[field](data[field][:, :, min(size(data[field], 3), datapoint)]) for field in fields]...,
+        #     mgn.n_norm["node_type"](node_type)
+        # ),
         mgn.e_norm(edge_features),
         senders,
         receivers

src/strategies.jl

@@ -3,8 +3,7 @@
 # Licensed under the MIT license. See LICENSE file in the project root for details.
 #
 
-import DifferentialEquations: ODEFunction
-import SciMLBase: AbstractSensitivityAlgorithm
+import SciMLBase: AbstractSensitivityAlgorithm, ODEFunction
 import SciMLSensitivity: InterpolatingAdjoint, ZygoteVJP
 
 #######################################################

src/utils.jl

@@ -4,7 +4,7 @@
 #
 
 import Printf: @sprintf
-import Statistics: mean, stdm
+import Statistics: stdm
 
 """
     der_minmax(path)
@@ -286,19 +286,3 @@ function clear_log(lines::Integer, move_up = true)
         clear_line()
     end
 end
-
-####################################################################
-# Overwritten for differentiation over GraphNetwork and normaliser #
-####################################################################
-
-Zygote.accum(x::Base.RefValue{Any}, y::NamedTuple{(:model, :ps, :st, :e_norm, :n_norm, :o_norm)}) = Zygote.accum(x[], y)
-
-Zygote.accum(x::NamedTuple{(:model, :ps, :st, :e_norm, :n_norm, :o_norm)}, y::Base.RefValue{Any}) = Zygote.accum(x, y[])
-
-Zygote.accum(x::Base.RefValue{Any}, y::NamedTuple{(:data_min, :data_max, :target_min, :target_max)}) = Zygote.accum(x[], y)
-
-Base.:+(x::Base.RefValue{Any}, y::NamedTuple{(:model, :ps, :st, :e_norm, :n_norm, :o_norm)}) = Zygote.accum(x[], y)
-
-Base.:+(x::NamedTuple{(:model, :ps, :st, :e_norm, :n_norm, :o_norm)}, y::Base.RefValue{Any}) = Zygote.accum(x, y[])
-
-Base.:+(x::Base.RefValue{Any}, y::NamedTuple{(:data_min, :data_max, :target_min, :target_max)}) = Zygote.accum(x[], y)

test/runtests.jl

@@ -5,7 +5,15 @@
 
 using MeshGraphNets
 using Test
+using Aqua
 
 @testset "MeshGraphNets.jl" begin
     # TODO
+    @testset "Aqua.jl" begin
+        # Ambiguities in external packages
+        @testset "Method ambiguity" begin
+            Aqua.test_ambiguities([MeshGraphNets])
+        end
+        Aqua.test_all(MeshGraphNets; ambiguities = false)
+    end
 end