tests/main/conftest.py (1)

596-624: Consider updating run_main_url_and_assert to support default_output_model_type.

Unlike run_main_and_assert, this function doesn't accept or propagate default_output_model_type. It calls _run_main_url which now supports the parameter, but passes nothing. This means URL-based tests will not get the default injection.

If this is intentional, a brief comment would clarify. Otherwise, consider adding the parameter for consistency.

♻️ Proposed fix

 def run_main_url_and_assert(
     *,
     url: str,
     output_path: Path,
     input_file_type: InputFileTypeLiteral | None,
     assert_func: AssertFileContent,
     expected_file: str | Path,
     extra_args: Sequence[str] | None = None,
+    default_output_model_type: str | None = DataModelType.PydanticBaseModel.value,
     transform: Callable[[str], str] | None = None,
     force_exec_validation: bool = False,
 ) -> None:
     ...
     __tracebackhide__ = True
-    return_code = _run_main_url(url, output_path, input_file_type, extra_args=extra_args)
+    return_code = _run_main_url(
+        url, output_path, input_file_type,
+        extra_args=extra_args,
+        default_output_model_type=default_output_model_type,
+    )
     ...

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@tests/main/conftest.py` around lines 596 - 624, run_main_url_and_assert
currently doesn't accept or forward default_output_model_type to _run_main_url,
so URL-based tests won't receive the default injection; update the
run_main_url_and_assert signature to accept default_output_model_type:
DefaultOutputModelType | None = None (or the same type used elsewhere), pass
that value through to the _run_main_url call, and update the docstring to
document the new parameter (or add a clarifying comment if omission was
intentional). Make sure to use the same parameter name and type as
run_main_and_assert to keep behavior consistent across both helpers.

tests/test_input_model.py (1)

881-901: Consider adding default_output_model_type to run_multiple_input_models_and_assert.

Unlike run_input_model_and_assert, this helper function does not include the default_output_model_type parameter and injection logic. If this is intentional (e.g., multiple input models always require explicit output type), it would be helpful to add a brief comment explaining why. Otherwise, consider adding the parameter for consistency.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@tests/test_input_model.py` around lines 881 - 901, The helper
run_multiple_input_models_and_assert is missing the default_output_model_type
parameter and the logic to inject it into CLI args like
run_input_model_and_assert does; add an optional parameter
default_output_model_type: str | None = None to
run_multiple_input_models_and_assert and, when provided, append the
corresponding flag/value (e.g., "--default-output-model-type",
default_output_model_type) into the args list before calling main; if omitting
this is intentional, instead add a brief comment above
run_multiple_input_models_and_assert explaining why multiple input models
require explicit output type so future readers aren't confused.