✨ validate_or_build function introduced

tests/test_profile.py

@@ -1,9 +1,21 @@
+from typing import Callable, Type
+
 import polars as pl
+import pytest
 
 from wimsey import profile
 from wimsey import execution
 
 
+def raise_exception_patch(exception_type: Type[Exception]) -> Callable:
+    """Creates a patch that will throw an error"""
+
+    def raise_exception(*args, **kwargs) -> None:
+        raise exception_type
+
+    return raise_exception
+
+
 def test_starter_tests_from_sampling_returns_passing_test() -> None:
     df = pl.DataFrame(
         {
@@ -77,3 +89,16 @@ def test_save_tests_from_samples_creates_expected_and_runnable_file(tmp_path) ->
     )
     result = execution.test(df, str(tmp_path / "cool.json"))
     assert result.success
+
+
+def test_validate_or_build_falls_back_to_save_starter_tests_from_sampling_if_validate_crashes(
+    monkeypatch,
+) -> None:
+    monkeypatch.setattr(profile, "validate", raise_exception_patch(FileNotFoundError))
+    monkeypatch.setattr(
+        profile,
+        "save_starter_tests_from_sampling",
+        raise_exception_patch(ZeroDivisionError),
+    )
+    with pytest.raises(ZeroDivisionError):
+        profile.validate_or_build(None, None)

wimsey/_version.py

@@ -1 +1 @@
-__version__ = "0.5.0"
+__version__ = "0.6.0"

wimsey/execution.py

@@ -70,7 +70,9 @@ def test(
 
 
 def validate(
-    df: FrameT, contract: str | list[dict] | dict, storage_options: dict | None = None
+    df: FrameT,
+    contract: str | list[dict] | dict,
+    storage_options: dict | None = None,
 ) -> FrameT:
     """
     Carry out tests on dataframe, returning original dataframe if tests are

wimsey/profile.py

@@ -7,6 +7,7 @@
 from narwhals.stable.v1.typing import FrameT
 
 from wimsey.dataframe import profile_from_sampling, profile_from_samples
+from wimsey.execution import validate
 
 
 class _StarterTestStatus(Enum):
@@ -222,3 +223,33 @@ def _type_starter_tests(
             test |= {"be_one_of": list(types)}
         tests.append(test)
     return tests
+
+
+def validate_or_build(
+    df: FrameT,
+    contract: str,
+    samples: int = 100,
+    n: int | None = None,
+    fraction: int | None = None,
+    margin: float = 1,
+    storage_options: dict | None = None,
+) -> FrameT:
+    """
+    Will attempt to validate based on a given contract, but if that contract does not exist yet
+    will generate one from sampling the dataset.
+
+    Will fall back to starter_tests_from_sampling (a list samples is not possible with
+    only one dataframe), see *starter_tests_from_sampling* and *save_starter_tests_from_sampling*
+    for more details on use of keyword arguments aside from df, contract and storage_options.
+    """
+    try:
+        validate(df, contract=contract, storage_options=storage_options)
+    except FileNotFoundError:
+        save_starter_tests_from_sampling(
+            path=contract,
+            df=df,
+            samples=samples,
+            n=n,
+            fraction=fraction,
+            margin=margin,
+        )