Fix CI (#1668)

Author: samuelcolvin

.github/workflows/ci.yml

@@ -390,7 +390,7 @@ jobs:
     steps:
       - uses: actions/setup-python@v5
         with:
-          python-version: '3.12'
+          python-version: "3.12"
       - name: Install dependencies
         run: pip install tweepy==4.14.0
       - name: Send tweet

docs/evals.md

@@ -156,7 +156,7 @@ _(This example is complete, it can be run "as is")_
 
 In this example we evaluate a method for generating recipes based on customer orders.
 
-```python {title="judge_recipes.py"}
+```python {title="judge_recipes.py" test="skip"}
 from __future__ import annotations
 
 from typing import Any
@@ -263,7 +263,7 @@ _(This example is complete, it can be run "as is")_
 
 Datasets can be saved to and loaded from YAML or JSON files.
 
-```python {title="save_load_dataset_example.py"}
+```python {title="save_load_dataset_example.py" test="skip"}
 from pathlib import Path
 
 from judge_recipes import CustomerOrder, Recipe, recipe_dataset
@@ -695,7 +695,7 @@ You can send these traces to any OpenTelemetry-compatible backend, including [Py
 
 All you need to do is configure Logfire via `logfire.configure`:
 
-```python {title="logfire_integration.py"}
+```python {title="logfire_integration.py" test="skip"}
 import logfire
 from judge_recipes import recipe_dataset, transform_recipe
 

pyproject.toml

@@ -114,7 +114,7 @@ include = [
 extend-select = [
     "Q",
     "RUF100",
-    "RUF018",  # https://docs.astral.sh/ruff/rules/assignment-in-assert/
+    "RUF018", # https://docs.astral.sh/ruff/rules/assignment-in-assert/
     "C90",
     "UP",
     "I",
@@ -193,7 +193,7 @@ filterwarnings = [
     "ignore:websockets.server.WebSocketServerProtocol is deprecated:DeprecationWarning",
     # random resource warnings; I suspect these are coming from vendor SDKs when running examples..
     "ignore:unclosed <socket:ResourceWarning",
-    "ignore:unclosed event loop:ResourceWarning"
+    "ignore:unclosed event loop:ResourceWarning",
 ]
 
 # https://coverage.readthedocs.io/en/latest/config.html#run

tests/test_examples.py

@@ -74,6 +74,15 @@ def find_filter_examples() -> Iterable[ParameterSet]:
             yield pytest.param(ex, id=test_id)
 
 
+@pytest.fixture
+def reset_cwd():
+    original_cwd = os.getcwd()
+    try:
+        yield
+    finally:
+        os.chdir(original_cwd)
+
+
 @pytest.mark.parametrize('example', find_filter_examples())
 def test_docs_examples(  # noqa: C901
     example: CodeExample,
@@ -83,6 +92,7 @@ def test_docs_examples(  # noqa: C901
     allow_model_requests: None,
     env: TestEnv,
     tmp_path: Path,
+    reset_cwd: None,
 ):
     mocker.patch('pydantic_ai.agent.models.infer_model', side_effect=mock_infer_model)
     mocker.patch('pydantic_ai._utils.group_by_temporal', side_effect=mock_group_by_temporal)
@@ -134,8 +144,6 @@ def print(self, *args: Any, **kwargs: Any) -> None:
         if sys.version_info < python_version_info:
             pytest.skip(f'Python version {python_version} required')
 
-    cwd = Path.cwd()
-
     if opt_test.startswith('skip') and opt_lint.startswith('skip'):
         pytest.skip('both running code and lint skipped')
 
@@ -189,7 +197,6 @@ def print(self, *args: Any, **kwargs: Any) -> None:
         else:
             module_dict = eval_example.run_print_check(example, call=call_name, module_globals=test_globals)
 
-        os.chdir(cwd)
         if title := opt_title:
             if title.endswith('.py'):
                 module_name = title[:-3]