# Header

{* ../../docs_src/dependencies/tutorial013_an_py310.py ln[19:21] *}

Some text

{* ../../docs_src/bigger_applications/app_an_py310/internal/admin.py hl[3] title["app/internal/admin.py"] *}

Some more text

{* ../../docs_src/dependencies/tutorial013_an_py310.py ln[30:38] hl[31:33] *}

# Set python3.12 as the default python version.
# TF does not support python3.13.
RUN ln -sf /usr/bin/python3.12 /usr/bin/python3
RUN ln -sf /usr/bin/python3.12 /usr/bin/python
RUN ln -sf /usr/lib/python3.12 /usr/lib/tf_python

# Link the compat driver to the location if available.
RUN if [ -e "/usr/local/cuda/compat/libcuda.so.1" ]; then ln -s /usr/local/cuda/compat/libcuda.so.1 /usr/lib/x86_64-linux-gnu/libcuda.so.1; fi

{* ../../docs_src/stream_data/tutorial002_py310.py ln[6,19:20] hl[20] *}

接著在路徑操作函式中用 `response_class=PNGStreamingResponse` 使用這個新類別：

{* ../../docs_src/stream_data/tutorial002_py310.py ln[23:27] hl[23] *}

### 模擬檔案 { #simulate-a-file }

此範例中我們用 `io.BytesIO` 模擬檔案。它是只存在於記憶體中的類檔案物件，但提供相同的介面。

例如，我們可以像讀取一般檔案一樣，透過迭代來消耗其內容。

{* ../../docs_src/stream_data/tutorial002_py310.py ln[1:27] hl[3,12:13,25] *}

/// note | 技術細節

# Header

{* ../../docs_src/dependencies/tutorial013_an_py310.py ln[19:21] *}

Some text

{* ../../docs_src/bigger_applications/app_an_py310/internal/admin.py hl[3] title["app/internal/admin.py"] *}

Some more text

{* ../../docs_src/dependencies/tutorial013_an_py310.py ln[30:38] hl[31:33] *}

And even more text

	handleListener := func(idx int, ln net.Listener) {
		for {
			conn, err := ln.Accept()
			select {
			case listener.acceptCh <- acceptResult{conn, err, idx}:
			case <-listener.ctxDoneCh:
				return
			}
		}
	}

	// Start separate goroutine for each listener to handle connection.
	for idx, ln := range listener.listeners {
		go handleListener(idx, ln)
	}
}

FastAPI에서 JSON Lines를 스트리밍하려면, *경로 처리 함수*에서 `return`을 사용하는 대신 `yield`로 각 항목을 차례로 생성하면 됩니다.

{* ../../docs_src/stream_json_lines/tutorial001_py310.py ln[1:24] hl[24] *}

보내려는 각 JSON 항목의 타입이 `Item`(Pydantic 모델)이고 함수가 async라면, 반환 타입을 `AsyncIterable[Item]`로 선언할 수 있습니다:

{* ../../docs_src/stream_json_lines/tutorial001_py310.py ln[1:24] hl[9:11,22] *}

반환 타입을 선언하면 FastAPI가 이를 사용해 데이터를 **검증**하고, OpenAPI에 **문서화**하고, **필터링**하고, Pydantic으로 **직렬화**합니다.

/// tip

{* ../../docs_src/json_base64_bytes/tutorial001_py310.py ln[1:9,29:35] hl[9] *}

If you check the `/docs`, they will show that the field `data` expects base64 encoded bytes:

<div class="screenshot">
<img src="/img/tutorial/json-base64-bytes/image01.png">
</div>

{* ../../docs_src/server_sent_events/tutorial001_py310.py ln[1:25] hl[4,22] *}

Each yielded item is encoded as JSON and sent in the `data:` field of an SSE event.

If you declare the return type as `AsyncIterable[Item]`, FastAPI will use it to **validate**, **document**, and **serialize** the data using Pydantic.

{* ../../docs_src/server_sent_events/tutorial001_py310.py ln[1:25] hl[10:12,23] *}

/// tip

{* ../../docs_src/stream_json_lines/tutorial001_py310.py ln[1:24] hl[24] *}

If each JSON item you want to send back is of type `Item` (a Pydantic model) and it's an async function, you can declare the return type as `AsyncIterable[Item]`:

{* ../../docs_src/stream_json_lines/tutorial001_py310.py ln[1:24] hl[9:11,22] *}

{* ../../docs_src/json_base64_bytes/tutorial001_py310.py ln[1:2,12:16,29,38:41] hl[16] *}

## Pydantic `bytes` 用於輸入與輸出資料 { #pydantic-bytes-for-input-and-output-data }

當然，你也可以使用同一個以 base64 設定的模型，同時處理輸入（以 `val_json_bytes` 驗證）與輸出（以 `ser_json_bytes` 序列化）的 JSON 資料。