Adding example to README and new test to ensure logging

README.md

@@ -28,3 +28,22 @@ The entire workload is divided into three different repositories:
 - [x] Polished compatibility with composed models
 - [ ] Polished compatibility with tuned models
 - [ ] Polished compatibility with iterative models
+
+## Example
+```julia
+# We first define a logger instance, providing the mlflow server address.
+# The experiment name and artifact location are optional.
+logger = MLFlowLogger("http://localhost:5000";
+ experiment_name="MLJFlow tests",
+ artifact_location="./mlj-test")
+
+X, y = make_moons(100) # X is a 100x2 matrix, y is a 100-element vector
+
+# Writing a normal MLJ workflow
+DecisionTreeClassifier = @load DecisionTreeClassifier pkg=DecisionTree
+dtc_machine = machine(dtc, X, y)
+
+# Passing the logger to the machine is enough to enable mlflow logging
+e1 = evaluate!(dtc_machine, resampling=CV(),
+ measures=[LogLoss(), Accuracy()], verbosity=1, logger=logger)
+```

src/types.jl

@@ -8,7 +8,6 @@ multiple methods in MLJBase.
 
 To use this logger, you need to have a MLFlow server running. For more
 information, see [MLFlow documentation](https://www.mlflow.org/docs/latest/quickstart.html).
-If it is not running, an informative error will be thrown.
 
 Depending on the MLFlow server configuration, the `baseuri` can be a local
 server or a remote server. The `experiment_name` is used to identify the

test/base.jl

@@ -6,9 +6,9 @@
  X, y = make_moons(100)
  DecisionTreeClassifier = @load DecisionTreeClassifier pkg=DecisionTree
 
- dtc = DecisionTreeClassifier()
- dtc_machine = machine(dtc, X, y)
- e1 = evaluate!(dtc_machine, resampling=CV(),
+ pipe = Standardizer() |> DecisionTreeClassifier()
+ mach = machine(pipe, X, y)
+ e1 = evaluate!(mach, resampling=CV(),
  measures=[LogLoss(), Accuracy()], verbosity=1, logger=logger)
 
  @testset "log_evaluation" begin
@@ -17,8 +17,15 @@
  @test typeof(runs[1]) == MLFlowRun
  end
 
+ @testset "ensuring logging" begin
+ runs = searchruns(logger.client,
+ getexperiment(logger.client, logger.experiment_name))
+ @test issetequal(keys(runs[1].data.params),
+ String.([keys(MLJModelInterface.flat_params(pipe))...]))
+ end
+
  @testset "save" begin
- run = MLJBase.save(logger, dtc_machine)
+ run = MLJBase.save(logger, mach)
  @test typeof(run) == MLFlowRun
  @test listartifacts(logger.client, run) |> length == 1
  end