For a 0.2.0 release

.github/codecov.yml

@@ -0,0 +1,6 @@
+coverage:
+  status:
+    project:
+      default:
+        threshold: 0.5%
+        target: auto

.github/workflows/CI.yml

@@ -28,11 +28,12 @@ jobs:
       - uses: actions/checkout@v2
       - uses: actions/setup-python@v4
         with:
-          python-version: '3.10'
+          python-version: '3.10.13'
       - name: Setup mlflow locally
         run: |
           pip install mlflow
-          python3 /opt/hostedtoolcache/Python/3.10.12/x64/bin/mlflow server --host 0.0.0.0 --port 5000 &
+          pip show mlflow 
+          python3 /opt/hostedtoolcache/Python/3.10.13/x64/bin/mlflow server --host 0.0.0.0 --port 5000 &
           sleep 5
       - uses: julia-actions/setup-julia@v1
         with:
@@ -51,6 +52,7 @@ jobs:
       - uses: julia-actions/julia-buildpkg@v1
       - uses: julia-actions/julia-runtest@v1
       - uses: julia-actions/julia-processcoverage@v1
-      - uses: codecov/codecov-action@v2
+      - uses: codecov/codecov-action@v3
         with:
           files: lcov.info
+

Project.toml

@@ -1,7 +1,7 @@
 name = "MLJFlow"
 uuid = "7b7b8358-b45c-48ea-a8ef-7ca328ad328f"
 authors = ["Jose Esparza <[email protected]>"]
-version = "0.1.1"
+version = "0.2.0"
 
 [deps]
 MLFlowClient = "64a0f543-368b-4a9a-827a-e71edb2a0b83"
@@ -10,15 +10,16 @@ MLJModelInterface = "e80e1ace-859a-464e-9ed9-23947d8ae3ea"
 
 [compat]
 MLFlowClient = "0.4.4"
-MLJBase = "0.21.14"
+MLJBase = "1"
 MLJModelInterface = "1.9.1"
 julia = "1.6"
 
 [extras]
 MLFlowClient = "64a0f543-368b-4a9a-827a-e71edb2a0b83"
 MLJDecisionTreeInterface = "c6f25543-311c-4c74-83dc-3ea6d1015661"
 MLJModels = "d491faf4-2d78-11e9-2867-c94bc002c0b7"
+StatisticalMeasures = "a19d573c-0a75-4610-95b3-7071388c7541"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [targets]
-test = ["Test", "MLFlowClient", "MLJModels", "MLJDecisionTreeInterface"]
+test = ["Test", "MLFlowClient", "MLJModels", "MLJDecisionTreeInterface", "StatisticalMeasures"]

README.md

@@ -1,4 +1,4 @@
-# MLJFlow
+# MLJFlow.jl
 
 | Branch | Build | Coverage |
 | :---: | :---: | :---: |
@@ -64,7 +64,7 @@ We assume MLJDecisionTreeClassifier is in the user's active Julia package
 environment.
 
 ```julia
-using MLJ
+using MLJ # Requires MLJ.jl version 0.19.3 or higher
 ```
 
 We first define a logger, providing the address of our running MLflow. The experiment

src/MLJFlow.jl

@@ -1,7 +1,6 @@
 module MLJFlow
 
-using MLJBase: info, name, Model,
-    Machine
+using MLJBase:  Model, Machine, name
 using MLJModelInterface: flat_params
 using MLFlowClient: MLFlow, logparam, logmetric,
     createrun, MLFlowRun, updaterun,

src/base.jl

@@ -3,7 +3,10 @@ function log_evaluation(logger::MLFlowLogger, performance_evaluation)
         artifact_location=logger.artifact_location)
     run = createrun(logger.service, experiment;
         tags=[
-            Dict("key" => "resampling", "value" => string(performance_evaluation.resampling)),
+            Dict(
+                "key" => "resampling",
+                "value" => string(performance_evaluation.resampling)
+            ),
             Dict("key" => "repeats", "value" => string(performance_evaluation.repeats)),
             Dict("key" => "model type", "value" => name(performance_evaluation.model)),
         ]

src/service.jl

@@ -18,20 +18,57 @@ function logmodelparams(service::MLFlow, run::MLFlowRun, model::Model)
     end
 end
 
+const MLFLOW_CHAR_SET =
+    "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789_-. /"
+
+"""
+    good_name(measure)
+
+**Private method.**
+
+Returns a string representation of `measure` that can be used as a valid name in
+MLflow. Includes the value of the first hyperparameter, if there is one.
+
+```julia
+julia> good_name(macro_f1score)
+"MulticlassFScore-beta_1.0"
+
+"""
+function good_name(measure)
+    name = string(measure)
+    name = replace(name, ", …" => "")
+    name = replace(name, " = " => "_")
+    name = replace(name, "()" => "")
+    name = replace(name, ")" => "")
+    map(collect(name)) do char
+        char in ['(', ','] && return '-'
+        char == '=' && return '_'
+        char in MLFLOW_CHAR_SET && return char
+        " "
+    end |> join
+end
+
 """
     logmachinemeasures(service::MLFlow, run::MLFlowRun, model::Model)
 
 Extracts the parameters of a model and logs them to the MLFlow server.
 
 # Arguments
-- `service::MLFlow`: An MLFlow service. See [MLFlowClient.jl](https://juliaai.github.io/MLFlowClient.jl/dev/reference/#MLFlowClient.MLFlow)
-- `run::MLFlowRun`: An MLFlow run. See [MLFlowClient.jl](https://juliaai.github.io/MLFlowClient.jl/dev/reference/#MLFlowClient.MLFlowRun)
+
+- `service::MLFlow`: An MLFlow service. See
+  [MLFlowClient.jl](https://juliaai.github.io/MLFlowClient.jl/dev/reference/#MLFlowClient.MLFlow)
+
+- `run::MLFlowRun`: An MLFlow run. See
+  [MLFlowClient.jl](https://juliaai.github.io/MLFlowClient.jl/dev/reference/#MLFlowClient.MLFlowRun)
+
 - `measures`: A vector of measures.
+
 - `measurements`: A vector of measurements.
+
 """
 function logmachinemeasures(service::MLFlow, run::MLFlowRun, measures,
     measurements)
-    measure_names = measures .|> info .|> x -> x.name
+    measure_names = measures .|> good_name
     for (name, value) in zip(measure_names, measurements)
         logmetric(service, run, name, value)
     end

test/runtests.jl

@@ -6,6 +6,9 @@ using MLJBase
 using MLJModels
 using MLFlowClient
 using MLJModelInterface
+using StatisticalMeasures
 
 include("base.jl")
 include("types.jl")
+include("service.jl")
+

test/service.jl

@@ -0,0 +1,8 @@
+@testset "good_name" begin
+    @test MLJFlow.good_name(rms) == "RootMeanSquaredError"
+    @test MLJFlow.good_name(macro_f1score) == "MulticlassFScore-beta_1.0"
+    @test MLJFlow.good_name(log_score) == "LogScore-tol_2.22045e-16"
+    @test MLJFlow.good_name("??") == "  "
+end
+
+true