Skip to content

fix(tools): support Python 3.10+ pipe union syntax in function parameter parser#4806

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
Raman369AI wants to merge 1 commit into from
Closed

fix(tools): support Python 3.10+ pipe union syntax in function parameter parser #4806

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 5 additions & 3 deletions src/google/adk/tools/_function_parameter_parse_util.py
View file
Open in desktop
Original file line numberDiff line numberDiff line change
Expand Up@@ -287,8 +287,10 @@ def _parse_schema_from_parameter(
schema.default = param.default
_raise_if_schema_unsupported(variant, schema)
return schema
if isinstance(param.annotation, _GenericAlias) or isinstance(
param.annotation, typing_types.GenericAlias
if (
isinstance(param.annotation, _GenericAlias)
or isinstance(param.annotation, typing_types.GenericAlias)
or isinstance(param.annotation, typing_types.UnionType)
):
origin = get_origin(param.annotation)
args = get_args(param.annotation)
Expand DownExpand Up@@ -330,7 +332,7 @@ def _parse_schema_from_parameter(
schema.default = param.default
_raise_if_schema_unsupported(variant, schema)
return schema
if origin is Union:
if origin in (Union, typing_types.UnionType):

@gemini-code-assistgemini-code-assistBotMar 13, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

While this change is correct to include UnionType, the subsequent logic for handling Optional[T] types (which results in len(schema.any_of) == 1) is a bit fragile and could be improved for robustness.

The current implementation at lines 369-371 only copies the type from the wrapped schema, and relies on a special-case handling for list types at lines 352-362 to copy the items. This means that for other complex Optional types, like Optional[pydantic.BaseModel], properties will be lost during the unwrapping.

Consider refactoring the unwrapping logic to be more generic and robust. For example:

# at lines 369-371iflen(schema.any_of) ==1:
# This is an Optional[T] type, so unwrap it.unwrapped_schema=schema.any_of[0]
schema.type=unwrapped_schema.typeschema.items=unwrapped_schema.itemsschema.properties=unwrapped_schema.propertiesschema.enum=unwrapped_schema.enumschema.any_of=None

This would handle various Optional types correctly and would also allow removing the special-case logic for lists.

schema.any_of = []
schema.type = types.Type.OBJECT
unique_types = set()
Expand Down
22 changes: 22 additions & 0 deletions tests/unittests/tools/test_function_tool.py
View file
Open in desktop
Original file line numberDiff line numberDiff line change
Expand Up@@ -529,3 +529,25 @@ async def async_tool(query: str, context: Context) -> dict:

assert result["query"] == "hello"
assert result["context_type"] == "Context"


def test_get_declaration_with_pipe_union_optional_list():
"""Test that pipe-union syntax (list[str] | None) parses correctly."""

def func_with_pipe_union(
required: list[str],
optional_list: list[str] | None = None,
optional_str: str | None = None,
) -> dict:
"""Function using Python 3.10+ pipe union syntax."""
return {}

tool = FunctionTool(func_with_pipe_union)
decl = tool._get_declaration()
assert decl is not None
params = decl.parameters.properties
assert params["required"].type.name == "ARRAY"
assert params["optional_list"].type.name == "ARRAY"
assert params["optional_list"].nullable is True
assert params["optional_str"].type.name == "STRING"
assert params["optional_str"].nullable is True