Merge remote-tracking branch 'origin/main' into mm/workflows

Author: MatthewMiddlehurst

.github/workflows/periodic_tests.yml

@@ -220,8 +220,7 @@ jobs:
         with:
           python-version: "3.12"
 
-      - if: ${{ github.event_name != 'pull_request' || !contains(github.event.pull_request.labels.*.name, 'no numba cache') }}
-        name: Use numba cache to set env variables but not restore cache
+      - name: Use numba cache to set env variables but not restore cache
         uses: ./.github/actions/numba_cache
         with:
           cache_name: "doctests"
@@ -242,6 +241,39 @@ jobs:
       - name: Run tests
         run: python -m pytest -n logical --doctest-only --doctest-continue-on-failure
 
+  multithreaded-estimators:
+    runs-on: ubuntu-24.04
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Setup Python 3.12
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Use numba cache to set env variables but not restore cache
+        uses: ./.github/actions/numba_cache
+        with:
+          cache_name: "multithreaded-estimators"
+          runner_os: ${{ runner.os }}
+          python_version: "3.12"
+          restore_cache: "false"
+
+      - name: Install aeon and dependencies
+        uses: nick-fields/retry@v3
+        with:
+          timeout_minutes: 30
+          max_attempts: 3
+          command: python -m pip install .[all_extras,dev]
+
+      - name: Show dependencies
+        run: python -m pip list
+
+      - name: Run tests
+        run: python -m pytest aeon/testing/tests/ --enablethreading true -k "check_estimator_multithreading"
+
   codecov:
     runs-on: ubuntu-24.04
 

.github/workflows/pr_labelled.yml

@@ -15,7 +15,7 @@ jobs:
 
     steps:
       - name: Create app token
-        uses: actions/create-github-app-token@v1
+        uses: actions/create-github-app-token@v2
         id: app-token
         with:
           app-id: ${{ vars.PR_APP_ID }}

.github/workflows/pr_pytest.yml

@@ -130,6 +130,36 @@ jobs:
       - name: Run tests
         run: python -m pytest -n logical --doctest-only --doctest-continue-on-failure
 
+  multithreaded-estimators:
+    runs-on: ubuntu-24.04
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Setup Python 3.12
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - if: ${{ github.event_name != 'pull_request' || !contains(github.event.pull_request.labels.*.name, 'no numba cache') }}
+        name: Restore numba cache
+        uses: ./.github/actions/numba_cache
+        with:
+          cache_name: "multithreaded-estimators"
+          runner_os: ${{ runner.os }}
+          python_version: "3.12"
+
+      - name: Install aeon and dependencies
+        uses: ./.github/actions/cpu_all_extras
+        with:
+          additional_extras: "dev"
+
+      - name: Show dependencies
+        run: python -m pip list
+
+      - name: Run tests
+        run: python -m pytest aeon/testing/tests/ --enablethreading true -k "check_estimator_multithreading"
 
   codecov:
     # run the code coverage job if a PR has the 'codecov actions' label

.pre-commit-config.yaml

@@ -29,7 +29,7 @@ repos:
         args: [ "--create", "--python-folders", "aeon" ]
 
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.12.0
+    rev: v0.12.1
     hooks:
       - id: ruff
         args: [ "--fix"]

aeon/anomaly_detection/series/distance_based/_lof.py

@@ -8,6 +8,7 @@
 import numpy as np
 
 from aeon.anomaly_detection.series._pyodadapter import PyODAdapter
+from aeon.utils.validation import check_n_jobs
 from aeon.utils.validation._dependencies import _check_soft_dependencies
 
 
@@ -98,14 +99,16 @@ def __init__(
         # Set a default contamination value internally
         contamination = 0.1
 
+        self._n_jobs = check_n_jobs(n_jobs)
+
         model = PyOD_LOF(
             n_neighbors=n_neighbors,
             algorithm=algorithm,
             leaf_size=leaf_size,
             metric=metric,
             p=p,
             metric_params=metric_params,
-            n_jobs=n_jobs,
+            n_jobs=self._n_jobs,
             contamination=contamination,  # Only for PyOD LOF
             novelty=False,  # Initialize unsupervised LOF (novelty=False)
         )
@@ -116,6 +119,7 @@ def __init__(
         self.p = p
         self.metric_params = metric_params
         self.n_jobs = n_jobs
+
         super().__init__(pyod_model=model, window_size=window_size, stride=stride)
 
     def _fit(self, X: np.ndarray, y: Union[np.ndarray, None] = None) -> None:

aeon/anomaly_detection/series/distance_based/_rockad.py

@@ -13,6 +13,7 @@
 
 from aeon.anomaly_detection.series.base import BaseSeriesAnomalyDetector
 from aeon.transformations.collection.convolution_based import Rocket
+from aeon.utils.validation import check_n_jobs
 from aeon.utils.windowing import reverse_windowing, sliding_windows
 
 
@@ -163,10 +164,12 @@ def _check_params(self, X: np.ndarray) -> None:
             )
 
     def _inner_fit(self, X: np.ndarray) -> None:
+        self._n_jobs = check_n_jobs(self.n_jobs)
+
         self.rocket_transformer_ = Rocket(
             n_kernels=self.n_kernels,
             normalise=self.normalise,
-            n_jobs=self.n_jobs,
+            n_jobs=self._n_jobs,
             random_state=self.random_state,
         )
         # X: (n_windows, window_size)
@@ -199,7 +202,7 @@ def _inner_fit(self, X: np.ndarray) -> None:
             # Initialize estimator
             estimator = NearestNeighbors(
                 n_neighbors=self.n_neighbors,
-                n_jobs=self.n_jobs,
+                n_jobs=self._n_jobs,
                 metric=self.metric,
                 algorithm="kd_tree",
             )

aeon/anomaly_detection/series/distribution_based/_copod.py

@@ -8,6 +8,7 @@
 import numpy as np
 
 from aeon.anomaly_detection.series._pyodadapter import PyODAdapter
+from aeon.utils.validation import check_n_jobs
 from aeon.utils.validation._dependencies import _check_soft_dependencies
 
 
@@ -46,7 +47,9 @@ def __init__(self, n_jobs: int = 1, window_size: int = 10, stride: int = 1):
         _check_soft_dependencies(*self._tags["python_dependencies"])
         from pyod.models.copod import COPOD
 
-        model = COPOD(n_jobs=n_jobs)
+        self._n_jobs = check_n_jobs(n_jobs)
+
+        model = COPOD(n_jobs=self._n_jobs)
         self.n_jobs = n_jobs
         super().__init__(model, window_size=window_size, stride=stride)
 

aeon/anomaly_detection/series/outlier_detection/_iforest.py

@@ -8,6 +8,7 @@
 import numpy as np
 
 from aeon.anomaly_detection.series._pyodadapter import PyODAdapter
+from aeon.utils.validation import check_n_jobs
 from aeon.utils.validation._dependencies import _check_soft_dependencies
 
 
@@ -101,12 +102,14 @@ def __init__(
         _check_soft_dependencies(*self._tags["python_dependencies"])
         from pyod.models.iforest import IForest
 
+        self._n_jobs = check_n_jobs(n_jobs)
+
         model = IForest(
             n_estimators=n_estimators,
             max_samples=max_samples,
             max_features=max_features,
             bootstrap=bootstrap,
-            n_jobs=n_jobs,
+            n_jobs=self._n_jobs,
             random_state=random_state,
             verbose=verbose,
         )

aeon/base/_base_collection.py

@@ -17,7 +17,6 @@ class name: BaseCollectionEstimator
     fitted model/strategy   - by convention, any attributes ending in "_"
     fitted state flag       - is_fitted (property)
     fitted state inspection - check_is_fitted()
-
 """
 
 from abc import abstractmethod
@@ -30,7 +29,6 @@ class name: BaseCollectionEstimator
     resolve_equal_length_inner_type,
     resolve_unequal_length_inner_type,
 )
-from aeon.utils.validation import check_n_jobs
 from aeon.utils.validation.collection import (
     get_n_cases,
     get_n_channels,
@@ -65,7 +63,6 @@ class BaseCollectionEstimator(BaseAeonEstimator):
     @abstractmethod
     def __init__(self):
         self.metadata_ = {}  # metadata/properties of data seen in fit
-        self._n_jobs = 1
 
         super().__init__()
 
@@ -115,21 +112,12 @@ def _preprocess_collection(self, X, store_metadata=True):
         """
         if isinstance(X, list) and isinstance(X[0], np.ndarray):
             X = self._reshape_np_list(X)
+
         meta = self._check_X(X)
         if len(self.metadata_) == 0 and store_metadata:
             self.metadata_ = meta
 
-        X = self._convert_X(X)
-        # This usage of n_jobs is legacy, see issue #102
-        multithread = self.get_tag("capability:multithreading")
-        if multithread:
-            if hasattr(self, "n_jobs"):
-                self._n_jobs = check_n_jobs(self.n_jobs)
-            else:
-                raise AttributeError(
-                    "self.n_jobs must be set if capability:multithreading is True"
-                )
-        return X
+        return self._convert_X(X)
 
     def _check_X(self, X):
         """

aeon/base/_estimators/hybrid/base_rist.py

@@ -114,6 +114,7 @@ def __init__(
 
     def _fit(self, X, y) -> object:
         self.n_cases_, self.n_channels_, self.n_timepoints_ = X.shape
+        self._n_jobs = check_n_jobs(self.n_jobs)
 
         rng = check_random_state(self.random_state)
 
@@ -134,10 +135,9 @@ def _fit(self, X, y) -> object:
 
         self._estimator = _clone_estimator(self._estimator, rng)
 
-        n_jobs = check_n_jobs(self.n_jobs)
         m = getattr(self._estimator, "n_jobs", "missing")
         if m != "missing":
-            self._estimator.n_jobs = n_jobs
+            self._estimator.n_jobs = self._n_jobs
 
         if self.series_transformers == "default":
             self._series_transformers = [
@@ -168,7 +168,7 @@ def _fit(self, X, y) -> object:
             if st is not None:
                 m = getattr(st, "n_jobs", "missing")
                 if m != "missing":
-                    st.n_jobs = n_jobs
+                    st.n_jobs = self._n_jobs
 
                 s = st.fit_transform(X, y)
             else:
@@ -198,7 +198,7 @@ def _fit(self, X, y) -> object:
                     row_numba_max,
                     row_ppv,
                 ],
-                n_jobs=n_jobs,
+                n_jobs=self._n_jobs,
             )
             _set_random_states(ct, rng)
             self._transformers.append(ct)
@@ -214,7 +214,7 @@ def _fit(self, X, y) -> object:
                 n_shapelets = self.n_shapelets
 
             st = RandomDilatedShapeletTransform(
-                max_shapelets=n_shapelets, n_jobs=n_jobs
+                max_shapelets=n_shapelets, n_jobs=self._n_jobs
             )
             _set_random_states(st, rng)
             self._transformers.append(st)