Merge branch 'dev' into production

Author: sverchkov

.dependabot/config.yml

@@ -0,0 +1,8 @@
+version: 2
+updates:
+- package-ecosystem: pip
+  directory: "/"
+  schedule:
+    interval: daily
+  open-pull-requests-limit: 10
+  target-branch: dev

.github/dependabot.yml

@@ -0,0 +1,55 @@
+# This workflow will install Python dependencies, run tests and lint with a variety of Python versions
+# For more information see: https://help.github.com/actions/language-and-framework-guides/using-python-with-github-actions
+
+name: Python package
+
+on:
+  push:
+    branches: [ dev ]
+  pull_request:
+    branches: [ dev ]
+
+jobs:
+  build:
+
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version: [3.8, 3.9]
+        requirements-cmd:
+          - -r testing-requirements.txt -r requirements.txt
+        pytest-run-cmd:
+          - pytest
+        exclude:
+          - python-version: 3.9
+            requirements-cmd: -r testing-requirements.txt -r requirements.txt
+            pytest-run-cmd: pytest
+        include:
+          - python-version: 3.8
+            requirements-cmd: -r testing-requirements.txt -r oldest
+          - python-version: 3.9
+            requirements-cmd: -r testing-requirements.txt -r coverage-requirements.txt -r requirements.txt
+            pytest-run-cmd: |
+              pytest --cov=generalizedtrees --cov-report xml
+              codecov
+
+    steps:
+    - uses: actions/checkout@v2
+    - name: Set up Python ${{ matrix.python-version }}
+      uses: actions/setup-python@v2
+      with:
+        python-version: ${{ matrix.python-version }}
+    - name: Install dependencies
+      run: |
+        python -m pip install --upgrade pip
+        python -m pip install flake8 pytest
+        python -m pip install ${{ matrix.requirements-cmd }}
+    - name: Lint with flake8
+      run: |
+        # stop the build if there are Python syntax errors or undefined names
+        flake8 . --count --select=E9,F63,F7,F82 --show-source --statistics
+        # exit-zero treats all errors as warnings. The GitHub editor is 127 chars wide
+        flake8 . --count --exit-zero --max-complexity=10 --max-line-length=127 --statistics
+    - name: Test with pytest
+      run: |
+        ${{ matrix.pytest-run-cmd }}

.github/workflows/python-package.yml

@@ -1,2 +1,2 @@
-pytest-cov==2.10.1
+pytest-cov==2.11.1
 codecov==2.1.11

.travis.yml

@@ -122,6 +122,8 @@ class MofN(Constraint):
     class SearchOperator(Flag):
         INC_M = auto()
         INC_N = auto()
+        DEC_M = auto()
+        DEC_N = auto()
         INC_NM = INC_N | INC_M
 
     def __init__(self, m: int, constraints):
@@ -208,15 +210,19 @@ def neighboring_tests(
 
         for operator in search_operators:
 
-            new_m = constraint.m_to_satisfy + \
-                (1 if operator & MofN.SearchOperator.INC_M else 0)
+            new_m = constraint.m_to_satisfy + (
+                1 if operator & MofN.SearchOperator.INC_M else (
+                    -1 if operator & MofN.SearchOperator.DEC_M else 0))
 
             if operator & MofN.SearchOperator.INC_N:
                 for atom in constraint_candidates:
                     new_atoms = constraint.constraints + (atom,)
-
                     yield MofN(new_m, new_atoms)
-            
+
+            elif operator & MofN.SearchOperator.DEC_N:
+                atoms = list(constraint.constraints)
+                for atom in atoms:
+                    yield MofN(new_m, tuple(atoms - {atom}))
             else:
                 yield MofN(new_m, constraint.constraints)
 

coverage-requirements.txt

@@ -110,10 +110,25 @@ class DataWithOracleGivensLC(GivensLC):
     a predictor oracle.
     """
 
-    def process(self, data, oracle, *args, target_names=None, prelabel_data=True, **kwargs):
+    def process(self, data, oracle, *args, target_names=None, prelabel_data=True, feature_groups=None, **kwargs):
+        """
+        Processing input for the explanation setting:
+
+        :param: data - the n-by-d feature matrix
+        :param: oracle - the oracle, can either be a SKLearn-like classifier, in which case the predict function is
+            used, or a function that takes data and outputs a prediction (in matrix form).
+        :param: target_names - The names of the target classes/components.
+        :param: prelabel_data - Whether to run the oracle on the data first, before building the model
+        :param: feature_groups - Used in some explanation configurations - A list of lists of feature indeces
+            representing semantically meaningful feature groups
+        """
 
         if target_names is not None:
             self.target_names = target_names
+        
+        if feature_groups is not None:
+            # TODO: Validation
+            self.feature_groups = feature_groups
 
         # Parse data
         self.data_matrix, self.feature_names, self.feature_spec = parse_data(
@@ -127,7 +142,7 @@ def process(self, data, oracle, *args, target_names=None, prelabel_data=True, **
             logger.info(
                 'Inferring that oracle is a Scikit-Learn-like classifier '
                 'and using the "predict" method.')
-            self.oracle = lambda x: np.eye(len(oracle.classes_))[oracle.predict(x),]
+            self.oracle = lambda x: np.eye(len(oracle.classes_))[oracle.predict(x).astype('intp'),]
         else:
             logger.info('Treating oracle as a function')
             self.oracle = oracle

generalizedtrees/constraints.py

@@ -192,6 +192,10 @@ class GreedyBuilderLC:
 
     def __init__(self):
         self.splitter = DefaultSplitConstructorLC()
+    
+    def initialize(self, givens: GivensLC) -> None:
+        self.node_builder.initialize(givens)
+        self.splitter.initialize(givens)
 
     def build_tree(self) -> Tree:
 

generalizedtrees/givens.py

@@ -105,8 +105,7 @@ def fit(self, *args, **kwargs):
 
         # Set components
         self.predictor.initialize(self.givens)
-        self.builder.node_builder.initialize(self.givens)
-        self.split_generator.initialize(self.givens)
+        self.builder.initialize(self.givens)
 
         # Build tree
         self.tree = self.builder.build_tree()