# buildifier: disable=load-on-top

workspace(name = "org_tensorflow")

# buildifier: disable=load-on-top

load("//third_party:repo.bzl", "tf_http_archive", "tf_mirror_urls")

tf_http_archive(
    name = "rules_shell",
    sha256 = "bc61ef94facc78e20a645726f64756e5e285a045037c7a61f65af2941f4c25e1",
    strip_prefix = "rules_shell-0.4.1",
    urls = tf_mirror_urls(

# Using Dataclasses { #using-dataclasses }

FastAPI is built on top of **Pydantic**, and I have been showing you how to use Pydantic models to declare requests and responses.

But FastAPI also supports using <a href="https://docs.python.org/3/library/dataclasses.html" class="external-link" target="_blank">`dataclasses`</a> the same way:

{* ../../docs_src/dataclasses_/tutorial001_py310.py hl[1,6:11,18:19] *}

    # More than one dependency could have the same field, it would show up as multiple
    # fields but it's the same one, so count them by name
    body_param_names_set = {field.name for field in fields}
    # A top level field has to be a single field, not multiple
    if len(body_param_names_set) > 1:
        return True
    first_field = fields[0]
    # If it explicitly specifies it is embedded, it has to be embedded

* Add support for setting a `default_response_class` in the `FastAPI` instance or in `include_router`. Initial PR [#467](https://github.com/tiangolo/fastapi/pull/467) by [@toppk](https://github.com/toppk).
* Add support for type annotations using strings and `from __future__ import annotations`. PR [#451](https://github.com/tiangolo/fastapi/pull/451) by [@dmontagu](https://github.com/dmontagu).

## 0.38.1

                [FastAPI docs for `lifespan`](https://fastapi.tiangolo.com/advanced/events/).
                """
            ),
        ] = None,
        # the generic to Lifespan[AppType] is the type of the top level application
        # which the router cannot know statically, so we use typing.Any
        lifespan: Annotated[
            Optional[Lifespan[Any]],
            Doc(
                """