Unpin shap and scipy (#4436)

* Unpin scipy and shap * Change for scipy compatibility * Changes for shap compatibility * Deps update check fixes * Theoretically fix nightly failure * release notes * Test fixes * removed -n 2 * removed -n 2 from git-test-other * maybe upping pytest-xdist fixes this? * windows no parallel-cpu and reorder linux test * python 3.10, 3.11 test * maybe fix test * fix dependency checker * Clarify make command names * Fix deps install logic --------- Co-authored-by: MichaelFu512 <[email protected]> Co-authored-by: Michael Fu <[email protected]>
alteryx · Jun 4, 2024 · 3d5bf45 · 3d5bf45
1 parent a6e14b5
commit 3d5bf45
Show file tree

Hide file tree

Showing 17 changed files with 63 additions and 55 deletions.
diff --git a/.github/meta.yaml b/.github/meta.yaml
@@ -27,14 +27,14 @@ outputs:
         - numpy >=1.22.0 
         - pandas >=1.5.0, <2.1.0
         - dask >=2022.2.0, !=2022.10.1
-        - scipy >=1.5.0, <1.12.0
+        - scipy >=1.5.0
         - scikit-learn >=1.3.2
         - scikit-optimize >=0.9.0
         - statsmodels >=0.12.2
         - colorama >=0.4.4
         - cloudpickle >=1.5.0
         - click >=8.0.0
-        - shap >=0.42.0, <0.45.0
+        - shap >=0.45.0
         - texttable >=1.6.2
         - woodwork >=0.22.0
         - featuretools >=1.16.0

diff --git a/.github/workflows/latest_dependency_checker.yaml b/.github/workflows/latest_dependency_checker.yaml
@@ -32,4 +32,4 @@ jobs:
           delete-branch: true
           base: main
           assignees: machineFL
-          reviewers: jeremyliweishih, chukarsten, MichaelFu512, eccabay, christopherbunn
+          reviewers: jeremyliweishih, thehomebrewnerd, MichaelFu512, eccabay, christopherbunn
diff --git a/.github/workflows/lint_tests.yaml b/.github/workflows/lint_tests.yaml
@@ -11,7 +11,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python_version: ["3.9", "3.10"]
+        python_version: ["3.9", "3.10", "3.11"]
     steps:
       - name: Checkout repository
         uses: actions/checkout@v3

diff --git a/.github/workflows/linux_unit_tests_with_latest_deps.yaml b/.github/workflows/linux_unit_tests_with_latest_deps.yaml
@@ -14,17 +14,8 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        include:
-          - python_version: "3.9"
-            command: 'git-test-automl'
-          - python_version: "3.9"
-            command: 'git-test-prophet'
-          - python_version: "3.9"
-            command: 'git-test-modelunderstanding'
-          - python_version: "3.9"
-            command: 'git-test-other'
-          - python_version: "3.9"
-            command: 'git-test-parallel'
+        python_version: ['3.9', '3.10', '3.11']
+        command: ['git-test-automl', 'git-test-modelunderstanding', 'git-test-parallel', 'git-test-prophet', 'git-test-other']
     steps:
       - name: Set up Python ${{ matrix.python_version }}
         uses: actions/setup-python@v4

diff --git a/.github/workflows/linux_unit_tests_with_minimum_deps.yaml b/.github/workflows/linux_unit_tests_with_minimum_deps.yaml
@@ -14,7 +14,7 @@ jobs:
       fail-fast: false
       matrix:
         python_version: ['3.9']
-        command: ['git-test-automl', 'git-test-modelunderstanding', 'git-test-other', 'git-test-parallel', 'git-test-prophet']
+        command: ['git-test-automl', 'git-test-modelunderstanding', 'git-test-parallel', 'git-test-prophet', 'git-test-other']
     steps:
       - name: Set up Python ${{ matrix.python_version }}
         uses: actions/setup-python@v4

diff --git a/.github/workflows/windows_unit_tests.yaml b/.github/workflows/windows_unit_tests.yaml
@@ -15,7 +15,7 @@ jobs:
       fail-fast: false
       matrix:
         python_version: ['3.9']
-        command: ['git-test-automl', 'git-test-modelunderstanding', 'git-test-other', 'git-test-parallel', 'git-test-prophet']
+        command: ['git-test-automl', 'git-test-modelunderstanding', 'git-test-parallel', 'git-test-prophet-no-parallel-cpu', 'git-test-other-no-parallel-cpu']
     steps:
       - name: Download Miniconda
         shell: pwsh
@@ -59,7 +59,7 @@ jobs:
           . $env:USERPROFILE\Miniconda3\shell\condabin\conda-hook.ps1
           conda activate curr_py
           conda install numba -q -y
-      - if: ${{ matrix.command == 'git-test-prophet' }}
+      - if: ${{ matrix.command == 'git-test-prophet-no-parallel-cpu' }}
         name: Install EvalML with test requirements and prophet
         shell: pwsh
         run: |
@@ -69,7 +69,7 @@ jobs:
           python -m pip install .[test]
           python -m pip install .[prophet]
           pip freeze
-      - if: ${{ matrix.command != 'git-test-prophet' }}
+      - if: ${{ matrix.command != 'git-test-prophet-no-parallel-cpu' }}
         name: Install EvalML with test requirements
         shell: pwsh
         run: |

diff --git a/Makefile b/Makefile
@@ -63,6 +63,14 @@ git-test-prophet:
 git-test-integration:
 	pytest evalml/tests/integration_tests -n 2 --durations 0 --timeout $(TIMEOUT) --cov=evalml --cov-config=./pyproject.toml --junitxml=test-reports/git-test-integration-junit.xml
 
+.PHONY: git-test-other-no-parallel-cpu
+git-test-other-no-parallel-cpu:
+	pytest evalml/tests --ignore evalml/tests/automl_tests/ --ignore evalml/tests/tuner_tests/ --ignore evalml/tests/model_understanding_tests/ --ignore evalml/tests/pipeline_tests/test_pipelines.py --ignore evalml/tests/component_tests/test_prophet_regressor.py --ignore evalml/tests/component_tests/test_components.py --ignore evalml/tests/component_tests/test_utils.py --ignore evalml/tests/integration_tests/ --durations 0 --timeout $(TIMEOUT) --cov=evalml --cov-config=./pyproject.toml --junitxml=test-reports/git-test-other-junit.xml
+	make doctests
+
+.PHONY: git-test-prophet-no-parallel-cpu
+git-test-prophet-no-parallel-cpu:
+	pytest evalml/tests/component_tests/test_prophet_regressor.py evalml/tests/component_tests/test_components.py evalml/tests/component_tests/test_utils.py evalml/tests/pipeline_tests/test_pipelines.py --durations 0 --timeout $(TIMEOUT) --cov=evalml --cov-config=./pyproject.toml --junitxml=test-reports/git-test-prophet-junit.xml
 
 .PHONY: installdeps
 installdeps:

diff --git a/core-requirements.txt b/core-requirements.txt
@@ -1,13 +1,13 @@
 numpy>=1.21.0
 pandas>=1.5.0, <2.1.0
-scipy>=1.5.0, <1.12.0
+scipy>=1.5.0
 scikit-learn>=1.3.2
 scikit-optimize>=0.9.0
 pyzmq>=20.0.0
 colorama>=0.4.4
 cloudpickle>=1.5.0
 click>=8.0.0
-shap>=0.42.0
+shap>=0.45.0
 statsmodels>=0.12.2
 texttable>=1.6.2
 woodwork>= 0.21.1

diff --git a/docs/source/release_notes.rst b/docs/source/release_notes.rst
@@ -9,12 +9,13 @@ Release Notes
         * Removed vowpalwabbit :pr:`4427`
         * Uncapped holidays :pr:`4428`
         * Unpinned kaleido :pr:`4423`
+        * Unpinned shap and scipy :pr:`4436`
     * Documentation Changes
     * Testing Changes
-        * Run airflow tests in Python 3.9 :pr:`4391`
-        * Remove iterative test from airflow runs :pr:`4424`
-        * Update GH actions to improve handling of potentially unsafe variables :pr:`4417`
-        * Fix install test :pr:`4423`
+        * Added ability to run airflow tests in Python 3.9 :pr:`4391`
+        * Removed iterative test from airflow runs :pr:`4424`
+        * Updated GH actions to improve handling of potentially unsafe variables :pr:`4417`
+        * Fixed install test :pr:`4423`
 
 .. warning::
 

diff --git a/evalml/data_checks/target_distribution_data_check.py b/evalml/data_checks/target_distribution_data_check.py
@@ -147,7 +147,8 @@ def _detect_log_distribution_helper(y):
     normalization_test = shapiro if len(y) <= 5000 else jarque_bera
     normalization_test_string = "shapiro" if len(y) <= 5000 else "jarque_bera"
     # Check if a normal distribution is detected with p-value above 0.05
-    if normalization_test(y).pvalue >= 0.05:
+    pvalue = normalization_test(y).pvalue
+    if pvalue >= 0.05 or np.isnan(pvalue):
         return False, normalization_test_string, None
 
     y_new = round(y, 6)
@@ -161,6 +162,6 @@ def _detect_log_distribution_helper(y):
 
     # If the p-value of the log transformed target is greater than or equal to the p-value of the original target
     # with outliers dropped, then it would imply that the log transformed target has more of a normal distribution
-    if norm_test_log.pvalue >= norm_test_og.pvalue:
+    if round(norm_test_log.pvalue, 6) >= round(norm_test_og.pvalue, 6):
         return True, normalization_test_string, norm_test_og
     return False, normalization_test_string, norm_test_og
diff --git a/evalml/model_understanding/prediction_explanations/_algorithms.py b/evalml/model_understanding/prediction_explanations/_algorithms.py
@@ -144,15 +144,6 @@ def _compute_shap_values(pipeline, features, training_data=None):
         if ws:
             logger.debug(f"_compute_shap_values TreeExplainer: {ws[0].message}")
         shap_values = explainer.shap_values(features, check_additivity=False)
-        # shap only outputs values for positive class for Catboost/Xgboost binary estimators.
-        # this modifies the output to match the output format of other binary estimators.
-        # Ok to fill values of negative class with zeros since the negative class will get dropped
-        # in the UI anyways.
-        if estimator.model_family in {
-            ModelFamily.CATBOOST,
-            ModelFamily.XGBOOST,
-        } and is_binary(pipeline.problem_type):
-            shap_values = [np.zeros(shap_values.shape), shap_values]
     else:
         if training_data is None:
             raise ValueError(
@@ -189,16 +180,29 @@ def _compute_shap_values(pipeline, features, training_data=None):
         except IndexError:
             expected_value = explainer.expected_value
 
-    # classification problem
-    if isinstance(shap_values, list):
-        mappings = []
-        for class_shap_values in shap_values:
-            mappings.append(_create_dictionary(class_shap_values, feature_names))
-        return (mappings, expected_value)
     # regression problem
-    elif isinstance(shap_values, np.ndarray):
+    if is_regression(pipeline.problem_type):
         dic = _create_dictionary(shap_values, feature_names)
         return (dic, expected_value)
+    # classification problem
+    if len(shap_values.shape) == 3:
+        mappings = []
+        for class_shap_values in shap_values.T:
+            mappings.append(_create_dictionary(class_shap_values.T, feature_names))
+        return (mappings, expected_value)
+    # shap only outputs values for positive class for boosted binary estimators.
+    # this modifies the output to match the output format of other binary estimators.
+    # Ok to fill values of negative class with the positive class since the negative class
+    # will get dropped in the UI anyways.
+    if estimator.model_family in {
+        ModelFamily.CATBOOST,
+        ModelFamily.XGBOOST,
+        ModelFamily.LIGHTGBM,
+    } and is_binary(pipeline.problem_type):
+        mappings = []
+        for _ in range(2):
+            mappings.append(_create_dictionary(shap_values, feature_names))
+        return (mappings, expected_value)
     else:
         raise ValueError(f"Unknown shap_values datatype {str(type(shap_values))}!")
 

diff --git a/evalml/tests/automl_tests/test_automl.py b/evalml/tests/automl_tests/test_automl.py
@@ -1530,9 +1530,12 @@ def test_results_getter(AutoMLTestEnv, X_y_binary):
 
     assert automl.results["pipeline_results"][0]["mean_cv_score"] == 1.0
 
-    with pytest.raises(AttributeError, match="set attribute"):
+    with pytest.raises(AttributeError) as atr_error:
         automl.results = 2.0
 
+    assert "set attribute" in str(atr_error.value) or "has no setter" in str(
+        atr_error.value,
+    )
     automl.results["pipeline_results"][0]["mean_cv_score"] = 2.0
     assert automl.results["pipeline_results"][0]["mean_cv_score"] == 1.0
 
@@ -4850,7 +4853,7 @@ def test_search_parameters_held_automl(
         max_batches=batches,
     )
     aml.search()
-    estimator_args = inspect.getargspec(RandomForestClassifier)
+    estimator_args = inspect.getfullargspec(RandomForestClassifier)
     # estimator_args[0] gives the parameter names, while [3] gives the associated values
     # estimator_args[0][i + 1] to skip 'self' in the estimator
     # we do len - 1 in order to skip the random seed, which isn't present in the row['parameters']

diff --git a/evalml/tests/dependency_update_check/latest_dependency_versions.txt b/evalml/tests/dependency_update_check/latest_dependency_versions.txt
@@ -27,9 +27,9 @@ pmdarima==2.0.4
 pyzmq==26.0.3
 scikit-learn==1.4.2
 scikit-optimize==0.10.1
-scipy==1.11.4
+scipy==1.13.1
 seaborn==0.13.2
-shap==0.44.1
+shap==0.45.1
 sktime==0.28.1
 statsmodels==0.14.2
 texttable==1.7.0

diff --git a/evalml/tests/dependency_update_check/minimum_requirements.txt b/evalml/tests/dependency_update_check/minimum_requirements.txt
@@ -27,7 +27,7 @@ scikit-learn==1.3.2
 scikit-optimize==0.9.0
 scipy==1.5.0
 seaborn==0.11.1
-shap==0.42.0
+shap==0.45.0
 sktime==0.21.0
 statsmodels==0.12.2
 texttable==1.6.2

diff --git a/evalml/tests/dependency_update_check/minimum_test_requirements.txt b/evalml/tests/dependency_update_check/minimum_test_requirements.txt
@@ -35,7 +35,7 @@ scikit-learn==1.3.2
 scikit-optimize==0.9.0
 scipy==1.5.0
 seaborn==0.11.1
-shap==0.42.0
+shap==0.45.0
 sktime==0.21.0
 statsmodels==0.12.2
 texttable==1.6.2

diff --git a/evalml/tests/model_understanding_tests/prediction_explanations_tests/test_algorithms.py b/evalml/tests/model_understanding_tests/prediction_explanations_tests/test_algorithms.py
@@ -332,7 +332,7 @@ def test_compute_shap_values_catches_shap_tree_warnings(
     def raise_warning_from_shap(estimator, feature_perturbation):
         warnings.warn("Shap raised a warning!")
         mock = MagicMock()
-        mock.shap_values.return_value = np.zeros(10)
+        mock.shap_values.return_value = np.zeros((1, 10, 2))
         return mock
 
     mock_tree_explainer.side_effect = raise_warning_from_shap

diff --git a/pyproject.toml b/pyproject.toml
@@ -31,14 +31,14 @@ requires-python = ">=3.9,<4"
 dependencies = [
     "numpy >= 1.22.0",
     "pandas >= 1.5.0, <2.1.0",
-    "scipy >= 1.5.0, < 1.12.0",
+    "scipy >= 1.5.0",
     "scikit-learn >= 1.3.2",
     "scikit-optimize >= 0.9.0",
     "pyzmq >= 20.0.0",
     "colorama >= 0.4.4",
     "cloudpickle >= 1.5.0",
     "click >= 8.0.0",
-    "shap >= 0.42.0, < 0.45.0",
+    "shap >= 0.45.0",
     "statsmodels >= 0.12.2",
     "texttable >= 1.6.2",
     "woodwork[dask] >= 0.22.0",
@@ -78,9 +78,9 @@ dependencies = [
 [project.optional-dependencies]
 test = [
     "pytest == 7.1.2",
-    "pytest-xdist == 2.1.0",
-    "pytest-timeout == 1.4.2",
-    "pytest-cov == 2.10.1",
+    "pytest-xdist >= 2.1.0",
+    "pytest-timeout >= 1.4.2",
+    "pytest-cov >= 2.10.1",
     "nbval == 0.9.3",
     "IPython >= 8.10.0, <8.12.1",
     "PyYAML == 6.0.1",