It had the best performance benchmarks at the time (only surpassed by Starlette).

At first, it didn't have an automatic API documentation web UI, but I knew I could add Swagger UI to it.

    sponsors = mkdocs.utils.yaml_load(sponsors_data_path.read_text(encoding="utf-8"))
    if not (match_start and match_end):
        raise RuntimeError("Couldn't auto-generate sponsors section")
    if not match_pre:
        raise RuntimeError("Couldn't find pre section (<style>) in index.md")
    frontmatter_end = match_pre.end()
    pre_end = match_start.end()
    post_start = match_end.start()

This way, you don't have to know beforehand what the valid field/attribute names are (as would be the case with Pydantic models).

This would be useful if you want to receive keys that you don't already know.

---

Another useful case is when you want to have keys of another type (e.g., `int`).

That's what we are going to see here.

In this case, you would accept any `dict` as long as it has `int` keys with `float` values:

If you use a `requirements.txt` file you could specify the version with:

```txt
fastapi==0.45.0
```

that would mean that you would use exactly the version `0.45.0`.

Or you could also pin it with:

```txt
fastapi>=0.45.0,<0.46.0
```

that would mean that you would use the versions `0.45.0` or above, but less than `0.46.0`, for example, a version `0.45.2` would still be accepted.

</div>

#### Generate Client Code

To generate the client code you can use the command line application `openapi-ts` that would now be installed.

Because it is installed in the local project, you probably wouldn't be able to call that command directly, but you would put it on your `package.json` file.

It could look like this:

```JSON  hl_lines="7"
{
  "name": "frontend-app",
  "version": "1.0.0",
  "description": "",

So, the frontend (that runs in the browser) would try to reach `/openapi.json` and wouldn't be able to get the OpenAPI schema.

Because we have a proxy with a path prefix of `/api/v1` for our app, the frontend needs to fetch the OpenAPI schema at `/api/v1/openapi.json`.

```mermaid
graph LR

browser("Browser")

* **Uvicorn**:
    * Will have the best performance, as it doesn't have much extra code apart from the server itself.
    * You wouldn't write an application in Uvicorn directly. That would mean that your code would have to include more or less, at least, all the code provided by Starlette (or **FastAPI**). And if you did that, your final application would have the same overhead as having used a framework and minimizing your app code and bugs.

<font color="#4E9A06">INFO</font>:     Application startup complete.
<font color="#4E9A06">INFO</font>:     Uvicorn running on <b>http://0.0.0.0:8000</b> (Press CTRL+C to quit)
```

</div>

That would work for most of the cases. 😎

You could use that command for example to start your **FastAPI** app in a container, in a server, etc.

## ASGI Servers

Let's go a little deeper into the details.

    # For coverage
    # TODO: in theory this would allow declaring types that could be lists of bytes
    # to be read from files and other types, but I'm not even sure it's a good idea
    # to support it as a first class "feature"
    assert is_bytes_sequence_annotation(Union[List[str], List[bytes]])


def test_is_uploadfile_sequence_annotation():
    # For coverage

        return data
    ```

    Note that for OAuth2 the scope `items:read` is a single scope in an opaque string.
    You could have custom internal logic to separate it by colon caracters (`:`) or
    similar, and get the two parts `items` and `read`. Many applications do that to
    group and organize permissions, you could do it as well in your application, just
    know that that it is application specific, it's not part of the specification.
    """