all other docstrings (#1783)

* all other docstrings

* docstrings improvements

* remove args in class docstrings

* put california housing dataset

* fix lfw dataset

* fix lfw dataset

* fix 20 news group

* fix cache

* fix housing

* fix housing

Author: mike0sv

.github/scripts/download_scipy_datasets.py

@@ -1,15 +1,31 @@
 import os
+import shutil
+import zipfile
+from pathlib import Path
 
 from sklearn import datasets
 
 data_home = datasets.get_data_home()
 print("Path of download datasets to: ", data_home)
 
-print("Download California Housing dataset...")
-datasets.fetch_california_housing()
-print("Download LFW people dataset...")
+# print("Download California Housing dataset...")
+# datasets.fetch_california_housing()
+print("Unzip LFW people dataset...")
+
+root_dir = Path(__file__).resolve().parents[2]
+lfw_zip_path = root_dir / "test_data" / "lfw-dataset.zip"
+lfw_home_path = Path(data_home) / "lfw_home"
+if not lfw_home_path.exists():
+    with zipfile.ZipFile(lfw_zip_path, "r") as zip_ref:
+        zip_ref.extractall(data_home)
+
+
 datasets.fetch_lfw_people()
-print("Download 20 news group dataset...")
+print("Copying 20 news group dataset...")
+shutil.copy(root_dir / "test_data" / "20news-bydate_py3.pkz", data_home)
 datasets.fetch_20newsgroups()
+print("Copying California Housing...")
+shutil.copy(root_dir / "test_data" / "cal_housing_py3.pkz", data_home)
+datasets.fetch_california_housing()
 print("Download completed.")
 print(f"Content of datasets cache: {os.listdir(data_home)}")

.github/share-actions/get-scipy-datasets/action.yml

@@ -1,4 +1,8 @@
 name: Get bikes dataset cached
+inputs:
+  google_backend_sa_key:
+    description: 'GCP service account key (base64 encoded)'
+    required: true
 runs:
   using: "composite"
   steps:
@@ -15,6 +19,20 @@ runs:
       uses: astral-sh/setup-uv@v7
       with:
         python-version: "3.11"
+    - name: Save GCP SA key
+      id: save_gcp_key
+      env:
+        GOOGLE_BACKEND_SA_KEY: ${{ inputs.google_backend_sa_key }}
+      run: |
+        echo "${GOOGLE_BACKEND_SA_KEY}" | base64 -d > gcp-credentials.json
+        echo "gcp_credentials_path=$(pwd)/gcp-credentials.json" >> "$GITHUB_OUTPUT"
+      shell: bash
+    - name: DVC Pull
+      if: ${{ steps.cache-scipy-data.outputs.cache-hit != 'true' }}
+      env:
+        GOOGLE_APPLICATION_CREDENTIALS: ${{ steps.save_gcp_key.outputs.gcp_credentials_path }}
+      run: uv run --with dvc[gs] dvc pull
+      shell: bash
     - name: Download datasets
       if: ${{ steps.cache-scipy-data.outputs.cache-hit != 'true' }}
       env:

.github/workflows/main.yml

@@ -128,6 +128,8 @@ jobs:
         uses: ./.github/share-actions/get-bikes-dataset-cached
       - name: 🔍 Get scipy dataset cached
         uses: ./.github/share-actions/get-scipy-datasets
+        with:
+          google_backend_sa_key: ${{ secrets.GOOGLE_BACKEND_SA_KEY }}
 
   test-minimal:
     name: Test on minimal requirements

src/evidently/core/container.py

@@ -23,22 +23,53 @@
 
 
 class MetricContainer(AutoAliasMixin, EvidentlyBaseModel, abc.ABC):
+    """Base class for containers that generate multiple metrics.
+
+    Metric containers are used to programmatically create multiple related metrics,
+    such as generating the same metric for multiple columns or creating metric combinations.
+    Examples include `ColumnMetricGenerator` and preset classes like `DataDriftPreset`.
+    """
+
     __alias_type__: ClassVar[str] = "metric_container"
 
     class Config:
         is_base_type = True
 
     include_tests: bool = True
+    """Whether to include default tests for generated metrics."""
 
     def __init__(self, include_tests: bool = True, **data):
+        """Initialize a metric container.
+
+        Args:
+        * `include_tests`: If `True`, generated metrics will include default tests.
+        """
         self.include_tests = include_tests
         super().__init__(**data)
 
     @abc.abstractmethod
     def generate_metrics(self, context: "Context") -> Sequence[MetricOrContainer]:
+        """Generate metrics based on the container configuration.
+
+        Args:
+        * `context`: `Context` containing datasets and configuration.
+
+        Returns:
+        * Sequence of `Metric` or `MetricContainer` objects to compute.
+        """
         raise NotImplementedError()
 
     def metrics(self, context: "Context") -> List[MetricOrContainer]:
+        """Get all metrics generated by this container.
+
+        Results are cached in the context to avoid regenerating on subsequent calls.
+
+        Args:
+        * `context`: `Context` containing datasets and configuration.
+
+        Returns:
+        * List of `Metric` or `MetricContainer` objects.
+        """
         metric_container_fp = self.get_fingerprint()
         metrics = context.metrics_container(metric_container_fp)
         if metrics is None:
@@ -51,9 +82,33 @@ def render(
         context: "Context",
         child_widgets: Optional[List[Tuple[Optional[MetricId], List[BaseWidgetInfo]]]] = None,
     ) -> List[BaseWidgetInfo]:
+        """Render visualization widgets for this container.
+
+        Combines widgets from all child metrics/containers.
+
+        Args:
+        * `context`: `Context` containing datasets and configuration.
+        * `child_widgets`: Optional list of (metric_id, widgets) tuples from child metrics.
+
+        Returns:
+        * List of `BaseWidgetInfo` objects for visualization.
+        """
         return list(itertools.chain(*[widget[1] for widget in (child_widgets or [])]))
 
     def list_metrics(self, context: "Context") -> Generator[Metric, None, None]:
+        """Iterate over all leaf metrics in this container.
+
+        Recursively yields all `Metric` objects, flattening nested containers.
+
+        Args:
+        * `context`: `Context` containing datasets and configuration.
+
+        Yields:
+        * `Metric` objects from this container and nested containers.
+
+        Raises:
+        * `ValueError`: If metrics haven't been generated yet.
+        """
         metrics = context.metrics_container(self.get_fingerprint())
         if metrics is None:
             raise ValueError("Metrics weren't composed in container")
@@ -66,6 +121,14 @@ def list_metrics(self, context: "Context") -> Generator[Metric, None, None]:
                 raise ValueError(f"invalid metric type {type(item)}")
 
     def _get_tests(self, tests):
+        """Get tests list, handling None and include_tests flag.
+
+        Args:
+        * `tests`: Optional list of tests.
+
+        Returns:
+        * Converted tests list, or None if default tests should be used, or empty list if tests disabled.
+        """
         if tests is not None:
             return convert_tests(tests)
         if self.include_tests:
@@ -77,8 +140,21 @@ def _get_tests(self, tests):
 
 
 class ColumnMetricContainer(MetricContainer, abc.ABC):
+    """Base class for metric containers that operate on a specific column.
+
+    Simplifies container implementation for containers that generate metrics
+    for a single column. Subclasses only need to implement `generate_metrics()`.
+    """
+
     column: str
+    """Name of the column to generate metrics for."""
 
     def __init__(self, column: str, include_tests: bool = True):
+        """Initialize a column metric container.
+
+        Args:
+        * `column`: Name of the column to generate metrics for.
+        * `include_tests`: If `True`, generated metrics will include default tests.
+        """
         self.column = column
         super().__init__(include_tests=include_tests)