* Bu dokümanın hedeflediğiniz dile sıfırdan yeni bir çevirisini yapın (örneğin `translate.py` içindeki `translate-page` komutu). Bu, çeviriyi `docs/{language code}/docs/_llm-test.md` altında oluşturur.
* Çeviride her şeyin yolunda olup olmadığını kontrol edin.
* Gerekirse dile özel prompt'u, genel prompt'u veya İngilizce dokümanı iyileştirin.
* Ardından çeviride kalan sorunları elle düzeltin; böylece iyi bir çeviri elde edin.

Use the following preferred translations when they apply in documentation prose:

- request (HTTP): リクエスト
- response (HTTP): レスポンス
- path operation: path operation (do not translate)

### `///` admonitions

1) Keep the admonition keyword in English (do not translate `note`, `tip`, etc.).
2) If a title is present, prefer these canonical titles:

- `/// note | 備考`
- `/// note | 技術詳細`
- `/// tip | 豆知識`
- `/// warning | 注意`

    transition: all 0.2s ease;
}

.suggestion-chip:hover {
    background-color: #f4f5f7;
    border-color: #0052cc;
    color: #0052cc;
    transform: translateY(-1px);
}

.suggestion-chip:active {
    transform: translateY(0);
}

/* ============================================
   Progress Indicator (Step display)
   ============================================ */
.progress-indicator {

    expected_content = Path(f"{data_path}/translated_doc_number_gt.md").read_text(
        "utf-8"
    )

    assert fixed_content == expected_content  # Translated doc remains unchanged
    assert "Error processing docs/lang/docs/doc.md" in result.output
    assert (
        "Number of headers with permalinks does not match the number "
        "in the original document (5 vs 4)"

### `///` admonitions

1) Keep the admonition keyword in English (do not translate `note`, `tip`, etc.).
2) Many Traditional Chinese docs currently omit titles in `///` blocks; that is OK.
3) If a generic title is present, prefer these canonical titles:

- `/// note | 注意`

Notes:

- `details` blocks exist; keep `/// details` as-is and translate only the title after `|`.
- Example canonical titles used in existing docs:

```

* Tenha um prompt específico do idioma – `docs/{language code}/llm-prompt.md`.
* Faça uma tradução nova deste documento para o seu idioma de destino (veja, por exemplo, o comando `translate-page` do `translate.py`). Isso criará a tradução em `docs/{language code}/docs/_llm-test.md`.
* Verifique se está tudo certo na tradução.
* Se necessário, melhore seu prompt específico do idioma, o prompt geral ou o documento em inglês.

    expected_content = Path(
        f"{data_path}/translated_doc_mermaid_translated.md"
    ).read_text("utf-8")

    assert fixed_content == expected_content  # Translated doc remains unchanged
    assert (
        "Skipping mermaid code block replacement (lines 41-44). This should be checked manually."
    ) in result.output


@pytest.mark.parametrize(
    "copy_test_files",

本文件用來測試用於翻譯文件的 <abbr title="Large Language Model - 大型語言模型">LLM</abbr> 是否理解 `scripts/translate.py` 中的 `general_prompt`，以及 `docs/{language code}/llm-prompt.md` 中的語言特定提示。語言特定提示會附加在 `general_prompt` 後面。

此處新增的測試會提供給所有語言特定提示的設計者參考。

使用方式：

* 準備語言特定提示 - `docs/{language code}/llm-prompt.md`。
* 針對本文件做一次全新的翻譯為你想要的目標語言（例如使用 `translate.py` 的 `translate-page` 指令）。這會在 `docs/{language code}/docs/_llm-test.md` 產生翻譯檔。
* 檢查翻譯是否正確。

    expected_content = Path(f"{data_path}/translated_doc_lines_number_gt.md").read_text(
        "utf-8"
    )

    assert fixed_content == expected_content  # Translated doc remains unchanged
    assert "Error processing docs/lang/docs/doc.md" in result.output
    assert (
        "Code block (lines 14-18) has different number of lines than the original block (5 vs 4)"
    ) in result.output

    expected_content = Path(f"{data_path}/translated_doc_wrong_lang_code.md").read_text(
        "utf-8"
    )

    assert fixed_content == expected_content  # Translated doc remains unchanged
    assert "Error processing docs/lang/docs/doc.md" in result.output
    assert (
        "Code block (lines 16-19) has different language than the original block ('yaml' vs 'toml')"
    ) in result.output