artifacts

Project.toml

@@ -19,12 +19,9 @@ MLJFlux = "094fc8d1-fd35-5302-93ea-dabda2abf845"
 MLJModelInterface = "e80e1ace-859a-464e-9ed9-23947d8ae3ea"
 MLUtils = "f1d291b0-491e-4a28-83b9-f70985020b54"
 Parameters = "d96e819e-fc66-5662-9728-84c9c7592b0a"
-Plots = "91a5bcdd-55d7-5caf-9e0b-520d859cae80"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
-SliceMap = "82cb661a-3f19-5665-9e27-df437c7e54c8"
 Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
-StatsPlots = "f3b207a7-027a-5e70-b257-86293d7955fd"
 Term = "22787eb5-b846-44ae-b979-8e399b8463ab"
 
 [compat]
@@ -42,10 +39,7 @@ MLJFlux = "0.2.10"
 MLJModelInterface = "1.8.0"
 MLUtils = "0.4.3"
 Parameters = "0.12.3"
-Plots = "1.38.17"
-SliceMap = "0.2.7"
 StatsBase = "0.33.0 - 0.34.0"
-StatsPlots = "0.15.6"
 Term = "2.0.5"
 julia = "1.8, 1.9"
 

experiments/Project.toml

@@ -10,6 +10,7 @@ ECCCo = "0232c203-4013-4b0d-ad96-43e3e11ac3bf"
 Flux = "587475ba-b771-5e3f-ad9e-33799f191a9c"
 JointEnergyModels = "48c56d24-211d-4463-bbc0-7a701b291131"
 LazyArtifacts = "4af54fe1-eca0-43a8-85a7-787d91b784e3"
+LibGit2 = "76f85450-5226-5b5a-8eaa-529ad045b433"
 Logging = "56ddb016-857b-54e1-b83d-db4d58db5568"
 MLJBase = "a7f614a8-145f-11e9-1d2a-a57a1082229d"
 MLJEnsembles = "50ed68f4-41fd-4504-931a-ed422449fee0"

experiments/experiment.jl

@@ -49,7 +49,7 @@ end
 include("data/data.jl")
 include("models/models.jl")
 include("benchmarking/benchmarking.jl")
-include("post_processing.jl")
+include("post_processing/post_processing.jl")
 
 """
     train_models!(outcome::ExperimentOutcome, exp::Experiment)

experiments/post_processing/artifacts.jl

+using ghr_jll
+using LazyArtifacts
+using LibGit2
+using Pkg.Artifacts
+using Serialization
+
+"""
+    generate_artifacts(
+        datafiles=DEFAULT_OUTPUT_PATH;
+        artifact_name=nothing,
+        root=".",
+        artifact_toml=LazyArtifacts.find_artifacts_toml("."),
+        deploy=true,
+        tag=nothing
+    )
+
+Uploads results to github releases. If `deploy=true`, then the results will be uploaded to a github release. If `deploy=false`, then the results will be saved locally.
+"""
+function generate_artifacts(
+    datafiles=DEFAULT_OUTPUT_PATH;
+    artifact_name=nothing,
+    root=".",
+    artifact_toml=LazyArtifacts.find_artifacts_toml("."),
+    deploy=true,
+    tag=nothing
+)
+
+    # Artifact name:
+    if isnothing(artifact_name)
+        suffix = TIME_STAMPED ? "" : "_$(Dates.format(now(), "yyyy-mm-dd@HH:MM"))"
+        artifact_name = "artifacts_$(split(datafiles, "/")[end])$(suffix)"
+    end
+    # Artifact tag:
+    if isnothing(tag)
+        tag = replace(lowercase(artifact_name), " " => "-")
+    end
+
+    if deploy && !haskey(ENV, "GITHUB_TOKEN")
+        @warn "For automatic github deployment, need GITHUB_TOKEN. Not found in ENV, attemptimg global git config."
+    end
+
+    if deploy
+        # Where we will put our tarballs
+        tempdir = mktempdir()
+
+        # Try to detect where we should upload these weights to (or just override
+        # as shown in the commented-out line)
+        origin_url = get_git_remote_url(root)
+        deploy_repo = "$(basename(dirname(origin_url)))/$(basename(origin_url))"
+    end
+
+    # create_artifact() returns the content-hash of the artifact directory once we're finished creating it
+    hash = create_artifact() do artifact_dir
+        cp(datafiles, joinpath(artifact_dir, artifact_name))
+    end
+
+    # Spit tarballs to be hosted out to local temporary directory:
+    if deploy
+        tarball_hash = archive_artifact(hash, joinpath(tempdir, "$(artifact_name).tar.gz"))
+
+        # Calculate tarball url
+        tarball_url = "https://github.com/$(deploy_repo)/releases/download/$(tag)/$(artifact_name).tar.gz"
+
+        # Bind this to an Artifacts.toml file
+        @info("Binding $(artifact_name) in Artifacts.toml...")
+        bind_artifact!(
+            artifact_toml,
+            artifact_name,
+            hash;
+            download_info=[(tarball_url, tarball_hash)],
+            lazy=true,
+            force=true
+        )
+    end
+
+    if deploy
+        # Upload tarballs to a special github release
+        @info("Uploading tarballs to $(deploy_repo) tag `$(tag)`")
+
+        ghr() do ghr_exe
+            println(
+                readchomp(
+                    `$ghr_exe -replace -u $(dirname(deploy_repo)) -r $(basename(deploy_repo)) $(tag) $(tempdir)`,
+                ),
+            )
+        end
+
+        @info("Artifacts.toml file now contains all bound artifact names")
+    end
+end
+
+function get_git_remote_url(repo_path::String=".")
+    repo = LibGit2.GitRepo(repo_path)
+    origin = LibGit2.get(LibGit2.GitRemote, repo, "origin")
+    return LibGit2.url(origin)
+end

experiments/post_processing.jl → experiments/post_processing/meta_data.jl

@@ -10,7 +10,7 @@ function meta(outcome::ExperimentOutcome; save_output::Bool=false)
     generator_params = meta_generators(outcome; save_output=save_output)
 
     return model_params, model_performance, generator_params
-    
+
 end
 
 """
@@ -40,7 +40,7 @@ function meta_model(outcome::ExperimentOutcome; save_output::Bool=false)
             :jem_sampling_steps => exp.sampling_steps,
         )
     )
-   
+
     if save_output
         save_path = joinpath(exp.params_path, "$(exp.save_name)_model_params.csv")
         @info "Saving model parameters to $(save_path)."
@@ -107,7 +107,7 @@ function meta_model_performance(outcome::ExperimentOutcome; measures::Union{Noth
         _perf.dataname .= exp.dataname
         model_performance = vcat(model_performance, _perf)
     end
-    
+
     @info "Model performance:"
     println(model_performance)
 
@@ -121,4 +121,5 @@ function meta_model_performance(outcome::ExperimentOutcome; measures::Union{Noth
     end
 
     return model_performance
-end
\ No newline at end of file
+end
+

experiments/post_processing/post_processing.jl

+include("meta_data.jl")
+include("artifacts.jl")
\ No newline at end of file

experiments/run_experiments.jl

@@ -42,3 +42,8 @@ end
 if USE_MPI
     MPI.Finalize()
 end
+
+if UPLOAD
+    @info "Uploading results."
+    generate_artifacts()
+end

experiments/setup_env.jl

@@ -61,16 +61,19 @@ artifact_toml = LazyArtifacts.find_artifacts_toml(".")
 _hash = artifact_hash(ARTIFACT_NAME, artifact_toml)
 const LATEST_ARTIFACT_PATH = joinpath(artifact_path(_hash), ARTIFACT_NAME)
 
+time_stamped = false
 if any(contains.(ARGS, "output_path"))
     @assert sum(contains.(ARGS, "output_path")) == 1 "Only one output path can be specified."
     _path = ARGS[findall(contains.(ARGS, "output_path"))][1] |> x -> replace(x, "output_path=" => "")
 else
     timestamp = Dates.format(now(), "yyyy-mm-dd@HH:MM")
+    time_stamped = true
     _path = "$(pwd())/results_$(timestamp)"
 end
 
 "Default output path."
 const DEFAULT_OUTPUT_PATH = _path
+const TIME_STAMPED = time_stamped
 
 ispath(DEFAULT_OUTPUT_PATH) || mkpath(DEFAULT_OUTPUT_PATH)
 
@@ -101,4 +104,8 @@ const CE_MEASURES = [
 ]
 
 "Test set proportion."
-const TEST_SIZE = 0.2
\ No newline at end of file
+const TEST_SIZE = 0.2
+
+if "upload" ∈ ARGS
+    const UPLOAD = true
+end
\ No newline at end of file

experiments/upload_artifacts.jl

+include("setup_env.jl");
+generate_artifacts()
\ No newline at end of file

src/model.jl

@@ -6,7 +6,6 @@ using MLJBase
 using MLJEnsembles
 using MLJFlux
 using MLUtils
-using SliceMap
 using Statistics
 
 const CompatibleAtomicModel = Union{<:MLJFlux.MLJFluxProbabilistic,MLJEnsembles.ProbabilisticEnsembleModel{<:MLJFlux.MLJFluxProbabilistic}}