Merge pull request #70 from cesmix-mit/main

update branch using main

.github/workflows/Documentation.yml

@@ -11,11 +11,9 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v2
-      - run: sudo apt-get update && sudo apt-get install -y xorg-dev mesa-utils xvfb libgl1 freeglut3-dev libxrandr-dev libxinerama-dev libxcursor-dev libxi-dev libxext-dev
       - uses: julia-actions/setup-julia@v1
         with:
-          version: "1.9"
-      - uses: julia-actions/cache@v1
+          version: "1.10"
       - name: add CESMIX registry
         run: |
           julia -e '
@@ -42,7 +40,7 @@ jobs:
             doctest(PotentialLearning)
           '
       - name: generate docs
-        run: DISPLAY=:0 xvfb-run -s '-screen 0 1024x768x24' julia --project=docs docs/make.jl
+        run: julia --project=docs docs/make.jl
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
           DOCUMENTER_KEY: ${{ secrets.DOCUMENTER_KEY }}

README.md

@@ -1,6 +1,6 @@
 ## [WIP] PotentialLearning.jl
 
-An open source Julia library for active learning of interatomic potentials in atomistic simulations of materials. It incorporates elements of bayesian inference, machine learning, differentiable programming,  software composability, and high-performance computing. This package is part of a software suite developed for the [CESMIX](https://computing.mit.edu/cesmix/) project.
+Developing Optimization Workflows for Fast and Accurate Interatomic Potentials. This package is part of a software suite developed for the [CESMIX](https://computing.mit.edu/cesmix/) project.
 
 <!--<a href="https://cesmix-mit.github.io/PotentialLearning.jl/stable">
 <img alt="Stable documentation" src="https://img.shields.io/badge/documentation-stable%20release-blue?style=flat-square">

docs/Project.toml

@@ -1,11 +1,11 @@
 [deps]
 AtomsBase = "a963bdd2-2df7-4f54-a1ee-49d51e6be12a"
-CairoMakie = "13f3f980-e62b-5c42-98c6-ff1f3baf88f0"
 Documenter = "e30172f5-a6a5-5a46-863b-614d45cd2de4"
 DocumenterCitations = "daee34ce-89f3-4625-b898-19384cb65244"
 InteratomicPotentials = "a9efe35a-c65d-452d-b8a8-82646cd5cb04"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 Literate = "98b081ad-f1c9-55d3-8b20-4c87d4299306"
+Plots = "91a5bcdd-55d7-5caf-9e0b-520d859cae80"
 PotentialLearning = "82b0a93c-c2e3-44bc-a418-f0f89b0ae5c2"
 Unitful = "1986cc42-f94f-5a68-af5c-568840ba703d"
 UnitfulAtomic = "a7773ee8-282e-5fa2-be4e-bd808c38a91a"

docs/make.jl

@@ -21,7 +21,8 @@ const EXAMPLES_DIR = joinpath(@__DIR__, "..", "examples")
 const OUTPUT_DIR   = joinpath(@__DIR__, "src/generated")
 
 examples = [
-    "Compute ACE descriptors, subsample, and fit ACE" => "Na/fit-dpp-ace-na.jl"
+    "Subsample Na dataset with DPP and fit with ACE" => "DPP-ACE-Na/fit-dpp-ace-na.jl",
+    "Load Ar+Lennard-Jones dataset and postprocess" => "LJ-Ar/lennard-jones-ar.jl"
 ]
 
 for (_, example_path) in examples

docs/src/index.md

@@ -1,23 +1,20 @@
 # [WIP] PotentialLearning.jl
 
-An open source Julia library to enhance active learning of interatomic potentials in atomistic simulations of materials. It incorporates elements of bayesian inference, machine learning, differentiable programming,  software composability, and high-performance computing. This package is part of a software suite developed for the [CESMIX](https://computing.mit.edu/cesmix/) project.
-
-## Specific goals
-
-- **Intelligent subsampling** of atomistic configurations via [DPP](https://github.com/dahtah/Determinantal.jl) and [clustering](https://docs.google.com/document/d/1SWAanEWQkpsbr2lqetMO3uvdX_QK-Z7dwrgPaM1Dl0o/edit)-based algorithms.
-- **Optimization of interatomic potentials**
-    - Parallel optimization of hyperparameters and coefficients via [Hyperopt.jl](https://github.com/baggepinnen/Hyperopt.jl).
-    - Multi-objective optimization (Pareto fronts): force execution time vs fitting accuracy (e.g. MAE of energies and forces).
-- **Neuralization of linear interatomic potentials**
-    - Neural version of [Julia-ACE](https://github.com/ACEsuit/ACE1.jl) and [LAMMPS-POD](https://docs.lammps.org/pair_pod.html).
-    - Integration with [Flux](https://fluxml.ai/Flux.jl/stable/) ecosystem.
-- **Interatomic potential compression**
-    - Feature selection (e.g. [CUR](https://github.com/JuliaLinearAlgebra/LowRankApprox.jl)) and dimensionality reduction (e.g [PCA](https://juliastats.org/MultivariateStats.jl/dev/pca/)) of atomistic descriptors.
-- **Interatomic potential fitting/training**
-    - Inference of parameter uncertainties in linear interatomic potentials.
-- **Quantity of Interest (QoI) sensitivity** analysis of interatomic potential parameters.
-- **Dimension reduction of QoI** through the theory of Active Subspaces.
-- **Atomistic configuration and DFT data management and post-processing**
+PotentialLerning.jl: **Developing optimization workflows for fast and accurate interatomic potentials**. This package is part of a software suite developed for the [CESMIX](https://computing.mit.edu/cesmix/) project.
+
+## Goals
+
+**Optimize your atomistic data: intelligent subsampling of large datasets to reduce DFT computations**
+- Intelligent subsampling of atomistic configurations using algorithms based on [DPP](https://github.com/dahtah/Determinantal.jl), [DBSCAN](https://docs.google.com/document/d/1SWAanEWQkpsbr2lqetMO3uvdX_QK-Z7dwrgPaM1Dl0o/edit), [CUR](https://github.com/JuliaLinearAlgebra/LowRankApprox.jl), etc.
+- Highly scalable parallel subsampling via hierarchical subsampling and distributed parallelism ([Dagger.jl](https://github.com/JuliaParallel/Dagger.jl)).
+- Optimal subsampler choosing via [Hyperopt.jl](https://github.com/baggepinnen/Hyperopt.jl).
+
+**Optimize your interatomic potential model: hyperparameters, coefficients, model compression, and model selection.**
+- Parallel optimization of hyperparameters, coefficients, and model selection via [Hyperopt.jl](https://github.com/baggepinnen/Hyperopt.jl); multi-objective optimization (Pareto fronts): force execution time vs fitting accuracy (e.g. MAE of energies and forces).
+- Model compression via feature selection (e.g. [CUR](https://github.com/JuliaLinearAlgebra/LowRankApprox.jl)) and dimensionality reduction (e.g [PCA](https://juliastats.org/MultivariateStats.jl/dev/pca/), Active Subspaces) of atomistic descriptors.
+- Fitting of linear potentials and inference of parameter uncertainties. Training of neural versions of [Julia-ACE](https://github.com/ACEsuit/ACE1.jl) and [LAMMPS-POD](https://docs.lammps.org/pair_pod.html).
+
+Additionally, this package provides utilities for atomistic configuration and DFT data management and post-processing.
   - Process input data so that it is ready for training. E.g. read XYZ file with atomic configurations, linearize energies and forces, split dataset into training and testing, normalize data, transfer data to GPU, define iterators, etc.
   - Post-processing: computation of different metrics (MAE, RSQ, COV, etc), saving results, and plotting.
 

examples/Ar/Project.toml → examples/DPP-ACE-Na/Project.toml

@@ -1,8 +1,8 @@
 [deps]
 AtomsBase = "a963bdd2-2df7-4f54-a1ee-49d51e6be12a"
-CairoMakie = "13f3f980-e62b-5c42-98c6-ff1f3baf88f0"
 InteratomicPotentials = "a9efe35a-c65d-452d-b8a8-82646cd5cb04"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
+Plots = "91a5bcdd-55d7-5caf-9e0b-520d859cae80"
 PotentialLearning = "82b0a93c-c2e3-44bc-a418-f0f89b0ae5c2"
 Unitful = "1986cc42-f94f-5a68-af5c-568840ba703d"
 UnitfulAtomic = "a7773ee8-282e-5fa2-be4e-bd808c38a91a"

examples/Na/fit-dpp-ace-na.jl → examples/DPP-ACE-Na/fit-dpp-ace-na.jl

@@ -1,12 +1,10 @@
-#push!(Base.LOAD_PATH, "../../")
-
 using Unitful, UnitfulAtomic
 using AtomsBase, InteratomicPotentials, PotentialLearning
-using LinearAlgebra, CairoMakie
+using LinearAlgebra, Plots
 
 # Load dataset
-path = joinpath(dirname(pathof(PotentialLearning)), "../examples/Na/")
-confs, thermo = load_data("$path/data/liquify_sodium.yaml", YAML(:Na, u"eV", u"Å"))
+path = joinpath(dirname(pathof(PotentialLearning)), "../examples/DPP-ACE-Na")
+confs, thermo = load_data("$path/../data/Na/liquify_sodium.yaml", YAML(:Na, u"eV", u"Å"))
 confs, thermo = confs[220:end], thermo[220:end]
 
 # Split dataset
@@ -53,16 +51,13 @@ println("MAE: $e_mae, RMSE: $e_rmse, RSQ: $e_rsq")
 # Plot energy error scatter
 e_err_train, e_err_test = (e_train_pred - e_train), (e_test_pred - e_test)
 dpp_inds2 = get_random_subset(dpp; batch_size = 20)
-size_inches = (12, 8)
-size_pt = 72 .* size_inches
-fig = Figure(resolution = size_pt, fontsize = 16)
-ax1 = Axis(fig[1, 1], xlabel = "Energy (eV/atom)", ylabel = "Error (eV/atom)")
-scatter!(ax1, e_train, e_err_train, label = "Training", markersize = 5.0)
-scatter!(ax1, e_test, e_err_test, label = "Test", markersize = 5.0)
-scatter!(ax1, e_train[dpp_inds2], e_err_train[dpp_inds2], markersize = 5.0,
-         color = :darkred, label = "DPP Samples")
-axislegend(ax1)
-save("$path/figures/energy_error_training_test_scatter.pdf", fig)
-display(fig)
-
+scatter( e_train, e_err_train, label = "Training", color = :blue,
+         markersize = 1.5, markerstrokewidth=0)
+scatter!(e_test, e_err_test, label = "Test", color = :yellow,
+         markersize = 1.5, markerstrokewidth=0)
+scatter!(e_train[dpp_inds2], e_err_train[dpp_inds2],
+         color = :darkred, label = "DPP Samples",
+         markersize = 2.5, markerstrokewidth=0)
+scatter!(xlabel = "Energy (eV/atom)", ylabel = "Error (eV/atom)", 
+         dpi = 1000, fontsize = 16)
 

examples/DFT-subsampling/subsampling_dpp.jl → examples/DPP-ACE-Si/fit-dpp-ace-si.jl

@@ -1,5 +1,3 @@
-push!(Base.LOAD_PATH, "../../")
-
 using LinearAlgebra, Random, InvertedIndices
 using Statistics, StatsBase, Distributions, Determinantal
 using Unitful, UnitfulAtomic
@@ -12,8 +10,9 @@ include("subsampling_utils.jl")
 # Load dataset -----------------------------------------------------------------
 elname = "Si"
 elspec = [:Si]
-inpath = "../Si-3Body-LAMMPS/"
-outpath = "./output/$elname/"
+path = joinpath(dirname(pathof(PotentialLearning)), "../examples/DPP-ACE-Si")
+inpath = "$path/../data/Si-3Body-LAMMPS/"
+outpath = "$path/output/$elname/"
 
 # Read all data
 file_arr = readext(inpath, "xyz")
@@ -47,8 +46,8 @@ ace = ACE(species = elspec,             # species
 
 # Update dataset by adding energy (local) descriptors --------------------------
 println("Computing local descriptors")
-@time e_descr = compute_local_descriptors(confs, ace)
-@time f_descr = compute_force_descriptors(confs, ace)
+e_descr = compute_local_descriptors(confs, ace)
+f_descr = compute_force_descriptors(confs, ace)
 JLD.save(outpath*"$(elname)_energy_descriptors.jld", "e_descr", e_descr)
 JLD.save(outpath*"$(elname)_force_descriptors.jld", "f_descr", f_descr)
 

examples/aHfO2/build_dpp.jl → examples/DPP-ACE-aHfO2/fit-dpp-ace-ahfo2.jl

@@ -7,9 +7,12 @@ using AtomsBase, InteratomicPotentials, PotentialLearning
 using JLD, CairoMakie
 
 #################### Importing Data ###################
+
+path = joinpath(dirname(pathof(PotentialLearning)), "../examples/DPP-ACE-aHfO2")
+
 # Import Raw Data
 energies, descriptors = JLD.load(
-    "examples/aHfO2/data/aHfO2_diverse_descriptors_3600.jld",
+    "$path/../data/aHfO2/aHfO2_diverse_descriptors_3600.jld",
     "energies",
     "descriptors",
 )

examples/Na/Project.toml → examples/LJ-Ar/Project.toml

@@ -1,8 +1,8 @@
 [deps]
 AtomsBase = "a963bdd2-2df7-4f54-a1ee-49d51e6be12a"
-CairoMakie = "13f3f980-e62b-5c42-98c6-ff1f3baf88f0"
 InteratomicPotentials = "a9efe35a-c65d-452d-b8a8-82646cd5cb04"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
+Plots = "91a5bcdd-55d7-5caf-9e0b-520d859cae80"
 PotentialLearning = "82b0a93c-c2e3-44bc-a418-f0f89b0ae5c2"
 Unitful = "1986cc42-f94f-5a68-af5c-568840ba703d"
 UnitfulAtomic = "a7773ee8-282e-5fa2-be4e-bd808c38a91a"

examples/LJ-Ar/lennard-jones-ar.jl

@@ -0,0 +1,34 @@
+using Unitful, UnitfulAtomic
+using AtomsBase, InteratomicPotentials, PotentialLearning
+using LinearAlgebra, Plots
+
+# Load dataset: Lennard-Jones + Argon
+path = joinpath(dirname(pathof(PotentialLearning)), "../examples/LJ-Ar")
+ds, thermo = load_data("$path/../data/LJ-AR/lj-ar.yaml", YAML(:Ar, u"eV", u"Å"))
+
+# Filter first configuration (zero energy)
+ds = ds[2:end]
+
+# Compute distance from origin, LJ energies, and time range
+systems = get_system.(ds)
+n_atoms = length(first(systems)) # Note: in this dataset all systems contain the same no. of atoms
+positions = position.(systems)
+dists_origin = map(x->ustrip.(norm.(x)), positions)
+energies = get_values.(get_energy.(ds))
+time_range = 0.5:0.5:5000
+
+# Plot distance from origin vs time
+p = plot(xlabel = "τ | ps",
+         ylabel = "Distance from origin | Å", 
+         dpi = 300, fontsize = 12)
+for i = 1:n_atoms
+    plot!(time_range, map(x->x[i], dists_origin), label="")
+end
+p
+
+# Plot LJ energies vs time
+plot(time_range, energies,
+     xlabel = "τ | ps",
+     ylabel = "Lennard Jones energy | eV",
+     dpi = 300, fontsize = 12)
+

examples/README.md

@@ -2,7 +2,7 @@
 
 Change the directory to the desired example folder. E.g.
 ```bash
-$ cd PotentialLearning.jl/examples/Na
+$ cd PotentialLearning.jl/examples/DPP-ACE-Na
 ```
 
 Open Julia REPL, activate ```Project.toml``` file in folder ```examples```, and chose the number of threads. E.g.

test/io/extxyz_test.jl

@@ -3,7 +3,7 @@ using Unitful, UnitfulAtomic
 
 energy_units = u"eV"
 distance_units = u"Å"
-ds = load_data("../examples/Si-3Body-LAMMPS/data.xyz", ExtXYZ(energy_units, distance_units));
+ds = load_data("../examples/data/Si-3Body-LAMMPS/data.xyz", ExtXYZ(energy_units, distance_units));
 
 @test length(ds) == 201
 @test typeof(ds) == DataSet

test/io/yaml_test.jl

@@ -4,7 +4,7 @@ using Unitful, UnitfulAtomic
 energy_units = u"eV"
 distance_units = u"Å"
 ds, t = load_data(
-    "../examples/Na/data/empirical_sodium_2d.yaml",
+    "../examples/data/Na/empirical_sodium_2d.yaml",
     YAML(:Na; energy_units = energy_units, distance_units = distance_units),
 );
 

test/subset_selector/subset_selector_tests.jl

@@ -28,7 +28,7 @@ r = RandomSelector(num_configs; batch_size = batch_size)
 # DBSCANSelector tests
 energy_units = u"eV"
 distance_units = u"Å"
-ds = load_data("../examples/Si-3Body-LAMMPS/data.xyz", ExtXYZ(energy_units, distance_units));
+ds = load_data("../examples/data/Si-3Body-LAMMPS/data.xyz", ExtXYZ(energy_units, distance_units));
 epsi, minpts, sample_size = 0.05, 5, batch_size
 dbscans = DBSCANSelector(   ds,
                             epsi,