Merge pull request #300 from BCG-Gamma/dev/2.0.dev0

BUILD: release facet 2.0.dev0

azure-pipelines.yml

@@ -1,9 +1,9 @@
 trigger:
-  - 1.2.x
+  - 2.0.x
   - release/*
 
 pr:
-  - 1.2.x
+  - 2.0.x
   - release/*
 
 # set the build name
@@ -23,15 +23,15 @@ resources:
       type: github
       endpoint: BCG-Gamma
       name: BCG-Gamma/pytools
-      ref: 1.2.x
+      ref: 2.0.x
 
 variables:
   ${{ if not(startsWith(variables['Build.SourceBranch'], 'refs/pull/')) }}:
     branchName: $[ replace(variables['Build.SourceBranch'], 'refs/heads/', '') ]
   ${{ if startsWith(variables['Build.SourceBranch'], 'refs/pull/') }}:
     branchName: $[ replace(variables['System.PullRequest.SourceBranch'], 'refs/heads/', '') ]
-  source_is_release_branch: $[ startsWith(variables['branchName'], 'release') ]
-  source_is_develop_branch: $[ or(startsWith(variables['branchName'], 'develop'), startsWith(variables['branchName'], 'dev/')) ]
+  source_is_release_branch: $[ startsWith(variables['branchName'], 'release/') ]
+  source_is_develop_branch: $[ startsWith(variables['branchName'], 'dev/') ]
   is_scheduled: $[ eq(variables['Build.Reason'], 'Schedule') ]
   project_name: facet
   project_root: $(project_name)
@@ -97,7 +97,7 @@ stages:
                 cd $(System.DefaultWorkingDirectory)
                 files_changed=$(git diff $(Build.SourceVersion)^! --name-only)
                 echo "Files changed since last commit: ${files_changed}"
-                n_files_changed=$(git diff $(Build.SourceVersion)^! --name-only | grep -i -E 'meta.yaml|pyproject.toml|azure-pipelines.yml|tox.ini'  | wc -l | xargs)
+                n_files_changed=$(git diff $(Build.SourceVersion)^! --name-only | grep -i -E 'meta\.yaml|pyproject\.toml|azure-pipelines\.yml|tox\.ini|make\.py'  | wc -l | xargs)
                 if [ ${n_files_changed} -gt 0 ]
                 then
                 build_changed=1
@@ -210,7 +210,7 @@ stages:
         - script: dir $(Build.SourcesDirectory)
 
         - script: |
-            conda install -y -c anaconda conda-build~=3.20.5 conda-verify toml=0.10.* flit=3.0.*
+            conda install -y -c anaconda conda-build~=3.20.5 conda-verify toml=0.10.* flit=3.0.* packaging~=20.9
           displayName: 'Install conda-build, flit, toml'
           condition: eq(variables['BUILD_SYSTEM'], 'conda')
 
@@ -297,7 +297,7 @@ stages:
         - script: dir $(Build.SourcesDirectory)
 
         - script: |
-            conda install -y -c anaconda conda-build~=3.20.5 conda-verify toml=0.10.* flit=3.0.*
+            conda install -y -c anaconda conda-build~=3.20.5 conda-verify toml=0.10.* flit=3.0.* packaging~=20.9
           displayName: 'Install conda-build, flit, toml'
           condition: eq(variables['BUILD_SYSTEM'], 'conda')
 
@@ -398,7 +398,7 @@ stages:
               condition: ne(variables.branchName, 'develop')
               script: |
                 set -eux
-                python -m pip install "toml==0.10.*"
+                python -m pip install toml~=0.10.2 packaging~=20.9
                 cd $(System.DefaultWorkingDirectory)/pytools
                 python <<EOF
                 from os import environ
@@ -461,13 +461,14 @@ stages:
               script: |
                 set -eux
                 echo "Getting version"
-                pip install packaging
+                pip install packaging~=20.9
                 cd $(System.DefaultWorkingDirectory)/$(project_root)/src
                 export PYTHONPATH=$(System.DefaultWorkingDirectory)/pytools/sphinx/base
                 version=$(python -c "import make_base; print(make_base.get_package_version())")
                 echo "Current version: $version"
-                echo "Detecting pre-release ('rc' in version)"
+                echo "Detecting pre-release ('dev' or 'rc' in version)"
                 prerelease=False
+                [[ $version == *dev* ]] && prerelease=True && echo "Development release identified"
                 [[ $version == *rc* ]] && prerelease=True && echo "Pre-release identified"
                 echo "##vso[task.setvariable variable=current_version]$version"
                 echo "##vso[task.setvariable variable=is_prerelease]$prerelease"
@@ -541,7 +542,7 @@ stages:
                     conda install -c conda-forge -c bcg_gamma $(package_name)
 
               isDraft: false
-              isPreRelease: $(is_prerelease)
+              isPrerelease: $(is_prerelease)
               assets: |
                 $(System.ArtifactsDirectory)/tox_default/tox/$(package_name)-*.tar.gz
                 $(System.ArtifactsDirectory)/conda_default/conda/noarch/$(package_name)-*.tar.bz2

environment.yml

@@ -5,7 +5,7 @@ channels:
 dependencies:
   # run
   - boruta_py ~= 0.3
-  - gamma-pytools ~= 1.2, >= 1.2.1
+  - gamma-pytools >= 2dev0, < 3a
   - joblib ~= 1.0
   - lightgbm ~= 3.2
   - matplotlib ~= 3.3
@@ -14,7 +14,7 @@ dependencies:
   - python ~= 3.8
   - scikit-learn ~= 0.24.2
   - scipy ~= 1.5
-  - sklearndf ~= 1.2
+  - sklearndf >= 2dev0, < 3a
   # build/test
   - black = 20.8b1
   - conda-build ~= 3.20

pyproject.toml

@@ -15,14 +15,14 @@ license = "Apache Software License v2.0"
 
 requires = [
     # direct requirements of gamma-facet
-    "gamma-pytools  ~=1.2,>=1.2.1",
+    "gamma-pytools  ~=2.0dev0",
     "matplotlib     ~=3.0",
     "numpy          >=1.17,<2a",
     "packaging      >=20",
     "pandas         >=0.24,<2a",
     "scipy          ~=1.2",
     "shap           >=0.34,<0.40a",
-    "sklearndf      ~=1.2",
+    "sklearndf      ~=2.0dev0",
     # additional requirements of shap 0.38
     "ipython        >=7",
 ]
@@ -71,15 +71,15 @@ Repository = "https://github.com/BCG-Gamma/facet"
 
 [build.matrix.min]
 # direct requirements of gamma-facet
-gamma-pytools  = "~=1.2.1"
+gamma-pytools  = "~=2.0.dev0"
 matplotlib     = "~=3.0.3"
 numpy          = ">=1.17.5,<18a"
 packaging      = "~=20.9"
 pandas         = "~=0.24.2"
 python         = "~=3.6.13"
 scipy          = "~=1.2.1"
 shap           = "~=0.34.0"
-sklearndf      = "~=1.2.0"
+sklearndf      = "~=2.0.dev0"
 # additional minimum requirements of sklearndf
 boruta         = "~=0.3.0"
 lightgbm       = "~=3.0.0"
@@ -92,15 +92,15 @@ ipython        = "~=7.0"
 
 [build.matrix.max]
 # direct requirements of gamma-facet
-gamma-pytools  = "~=1.2,>=1.2.1"
+gamma-pytools  = ">=2dev0,<3a"
 matplotlib     = "~=3.4"
 numpy          = ">=1.20,<2a"
 packaging      = ">=20.9"
 pandas         = "~=1.2"
 python         = "~=3.8"
 scipy          = "~=1.5"
 shap           = "~=0.39.0"
-sklearndf      = "~=1.2"
+sklearndf      = ">=2dev0,<3a"
 # additional maximum requirements of sklearndf
 boruta         = "~=0.3"
 lightgbm       = "~=3.2"

src/facet/__init__.py

@@ -6,7 +6,7 @@
 """
 
 
-__version__ = "1.2.0"
+__version__ = "2.0.dev0"
 
 __logo__ = (
     r"""

src/facet/crossfit/_crossfit.py

@@ -480,7 +480,7 @@ def on_run(self) -> None:
                 if do_fit:
                     crossfit._reset_fit()
 
-            def collate(self, job_results: List[FitResult]) -> Optional[np.ndarray]:
+            def aggregate(self, job_results: List[FitResult]) -> Optional[np.ndarray]:
                 models, scores = zip(*job_results)
 
                 if do_fit:

src/facet/inspection/_shap.py

@@ -275,23 +275,21 @@ def _make_explainer(_model: T_LearnerPipelineDF) -> BaseExplainer:
 
         else:
             shap_df_per_split = JobRunner.from_parallelizable(self).run_jobs(
-                *(
-                    Job.delayed(self._get_shap_for_split)(
-                        model,
-                        sample,
-                        _make_explainer(model),
-                        self.feature_index_,
-                        self._convert_raw_shap_to_df,
-                        self.get_multi_output_type(),
-                        self._get_multi_output_names(model=model, sample=sample),
-                    )
-                    for model, sample in zip(
-                        crossfit.models(),
-                        (
-                            sample.subsample(iloc=oob_split)
-                            for _, oob_split in crossfit.splits()
-                        ),
-                    )
+                Job.delayed(self._get_shap_for_split)(
+                    model,
+                    sample,
+                    _make_explainer(model),
+                    self.feature_index_,
+                    self._convert_raw_shap_to_df,
+                    self.get_multi_output_type(),
+                    self._get_multi_output_names(model=model, sample=sample),
+                )
+                for model, sample in zip(
+                    crossfit.models(),
+                    (
+                        sample.subsample(iloc=oob_split)
+                        for _, oob_split in crossfit.splits()
+                    ),
                 )
             )
 

src/facet/selection/_selection.py

@@ -518,7 +518,7 @@ def _rank_learners(
         )
 
         pipeline_scorings: List[np.ndarray] = list(
-            JobRunner.from_parallelizable(self).run_queues(*queues)
+            JobRunner.from_parallelizable(self).run_queues(queues)
         )
 
         for crossfit, pipeline_parameters, pipeline_scoring in zip(

src/facet/simulation/_simulation.py

@@ -383,12 +383,10 @@ def simulate_actuals(self) -> pd.Series:
         y_mean = self.expected_output()
 
         result: List[float] = JobRunner.from_parallelizable(self).run_jobs(
-            *(
-                Job.delayed(self._simulate_actuals)(
-                    model, subsample.features, y_mean, self._simulate
-                )
-                for model, subsample in self._get_simulations()
+            Job.delayed(self._simulate_actuals)(
+                model, subsample.features, y_mean, self._simulate
             )
+            for model, subsample in self._get_simulations()
         )
 
         return pd.Series(
@@ -455,16 +453,14 @@ def _simulate_feature_with_values(
         simulation_means_and_sems_per_split: List[
             Tuple[Sequence[float], Sequence[float]]
         ] = JobRunner.from_parallelizable(self).run_jobs(
-            *(
-                Job.delayed(UnivariateUpliftSimulator._simulate_values_for_split)(
-                    model=model,
-                    subsample=subsample,
-                    feature_name=feature_name,
-                    simulated_values=simulation_values,
-                    simulate_fn=self._simulate,
-                )
-                for (model, subsample) in self._get_simulations()
+            Job.delayed(UnivariateUpliftSimulator._simulate_values_for_split)(
+                model=model,
+                subsample=subsample,
+                feature_name=feature_name,
+                simulated_values=simulation_values,
+                simulate_fn=self._simulate,
             )
+            for (model, subsample) in self._get_simulations()
         )
 
         index_name: str

src/facet/simulation/viz/_style.py

@@ -201,7 +201,7 @@ def _make_sub_axes() -> Axes:
 
             def _x_axis_height() -> float:
                 _, axis_below_size_pixels = main_ax.get_xaxis().get_text_heights(
-                    self.renderer
+                    self.get_renderer()
                 )
                 ((_, y0), (_, y1)) = main_ax.transData.inverted().transform(
                     ((0, 0), (0, axis_below_size_pixels))