吴恩达的翻译Agent项目，复现教程来了！

#执行任务
# 调用编译后的工作流，传入初始状态字典
result = app.invoke({
    "source_lang": "English", # 源语言为英语
    "target_lang": "中文",    # 目标语言为中文
    "source_text": """By using an LLM as the heart of the translation engine, this system is highly steerable. For example, by changing the prompts, it is easier using this workflow than a traditional machine translation (MT) system to:

Modify the output's style, such as formal/informal.
Specify how to handle idioms and special terms like names, technical terms, and acronyms. For example, including a glossary in the prompt lets you make sure particular terms (such as open source, H100 or GPU) are translated consistently.
Specify specific regional use of the language, or specific dialects, to serve a target audience. For example, Spanish spoken in Latin America is different from Spanish spoken in Spain; French spoken in Canada is different from how it is spoken in France."""
})

defone_chunk_initial_translation(
    source_lang: str, target_lang: str, source_text: str
) -> str:
    """
    Translate the entire text as one chunk using an LLM.

    Args:
        source_lang (str): The source language of the text.
        target_lang (str): The target language for translation.
        source_text (str): The text to be translated.

    Returns:
        str: The translated text.
    """

    system_message = f"You are an expert linguist, specializing in translation from {source_lang} to {target_lang}."

    translation_prompt = f"""This is an {source_lang} to {target_lang} translation, please provide the {target_lang} translation for this text. \
Do not provide any explanations or text apart from the translation.
{source_lang}: {source_text}

{target_lang}:"""

    translation = get_completion(translation_prompt, system_message=system_message)

    return translation

defone_chunk_reflect_on_translation(
    source_lang: str,
    target_lang: str,
    source_text: str,
    translation_1: str,
    country: str = "",
) -> str:
    """
    Use an LLM to reflect on the translation, treating the entire text as one chunk.

    Args:
        source_lang (str): The source language of the text.
        target_lang (str): The target language of the translation.
        source_text (str): The original text in the source language.
        translation_1 (str): The initial translation of the source text.
        country (str): Country specified for the target language.

    Returns:
        str: The LLM's reflection on the translation, providing constructive criticism and suggestions for improvement.
    """

    system_message = f"You are an expert linguist specializing in translation from {source_lang} to {target_lang}. \
You will be provided with a source text and its translation and your goal is to improve the translation."

    if country != "":
        reflection_prompt = f"""Your task is to carefully read a source text and a translation from {source_lang} to {target_lang}, and then give constructive criticism and helpful suggestions to improve the translation. \
The final style and tone of the translation should match the style of {target_lang} colloquially spoken in {country}.

The source text and initial translation, delimited by XML tags  and , are as follows:


{source_text}



{translation_1}


When writing suggestions, pay attention to whether there are ways to improve the translation's \n\
(i) accuracy (by correcting errors of addition, mistranslation, omission, or untranslated text),\n\
(ii) fluency (by applying {target_lang} grammar, spelling and punctuation rules, and ensuring there are no unnecessary repetitions),\n\
(iii) style (by ensuring the translations reflect the style of the source text and take into account any cultural context),\n\




    

(iv) terminology (by ensuring terminology use is consistent and reflects the source text domain; and by only ensuring you use equivalent idioms {target_lang}).\n\

Write a list of specific, helpful and constructive suggestions for improving the translation.
Each suggestion should address one specific part of the translation.
Output only the suggestions and nothing else."""

    else:
        reflection_prompt = f"""Your task is to carefully read a source text and a translation from {source_lang} to {target_lang}, and then give constructive criticisms and helpful suggestions to improve the translation. \

The source text and initial translation, delimited by XML tags  and , are as follows:


{source_text}



{translation_1}


When writing suggestions, pay attention to whether there are ways to improve the translation's \n\
(i) accuracy (by correcting errors of addition, mistranslation, omission, or untranslated text),\n\
(ii) fluency (by applying {target_lang} grammar, spelling and punctuation rules, and ensuring there are no unnecessary repetitions),\n\
(iii) style (by ensuring the translations reflect the style of the source text and take into account any cultural context),\n\
(iv) terminology (by ensuring terminology use is consistent and reflects the source text domain; and by only ensuring you use equivalent idioms {target_lang}).\n\

Write a list of specific, helpful and constructive suggestions for improving the translation.
Each suggestion should address one specific part of the translation.
Output only the suggestions and nothing else."""

    reflection = get_completion(reflection_prompt, system_message=system_message)
    return reflection

defone_chunk_improve_translation(
    source_lang: str,
    target_lang: str,
    source_text: str,
    translation_1: str,
    reflection: str,
) -> str:
    """
    Use the reflection to improve the translation, treating the entire text as one chunk.

    Args:
        source_lang (str): The source language of the text.
        target_lang (str): The target language for the translation.
        source_text (str): The original text in the source language.
        translation_1 (str): The initial translation of the source text.
        reflection (str): Expert suggestions and constructive criticism for improving the translation.

    Returns:
        str: The improved translation based on the expert suggestions.
    """

    system_message = f"You are an expert linguist, specializing in translation editing from {source_lang} to {target_lang}."

    prompt = f"""Your task is to carefully read, then edit, a translation from {source_lang} to {target_lang}, taking into
account a list of expert suggestions and constructive criticisms.

The source text, the initial translation, and the expert linguist suggestions are delimited by XML tags ,  and  \
as follows:


{source_text}



{translation_1}



{reflection}


Please take into account the expert suggestions when editing the translation. Edit the translation by ensuring:

(i) accuracy (by correcting errors of addition, mistranslation, omission, or untranslated text),
(ii) fluency (by applying {target_lang} grammar, spelling and punctuation rules and ensuring there are no unnecessary repetitions), \
(iii) style (by ensuring the translations reflect the style of the source text)
(iv) terminology (inappropriate for context, inconsistent use), or
(v) other errors.

Output only the new translation and nothing else."""

    translation_2 = get_completion(prompt, system_message)

    return translation_2

import json
import openai
from langgraph.graph import StateGraph, START, END
import os

# 设置 DeepSeek API 的相关参数
deep_seek_api_key = 'sk-21070************494b7e5db'# API 密钥
deep_seek_url = 'https://api.deepseek.com/v1'# API 请求的基地址
deep_seek_default_model = 'deepseek-reasoner'# 默认使用的模型名称

# 使用 OpenAI 的 Python 客户端库创建一个客户端实例，用于与 DeepSeek API 进行交互
client = openai.OpenAI(
    api_key=deep_seek_api_key,  # 设置 API 密钥
    base_url=deep_seek_url  # 设置 API 请求的基地址
)
default_model = deep_seek_default_model  # 设置默认模型

defget_completion(
    prompt: str,  # 提示文本
    system_message: str = "You are a helpful assistant.",  # 系统消息，默认值为 "You are a helpful assistant."
    model: str = default_model,  # 使用的模型，默认为 deep_seek_default_model
    temperature: float = 0.3,  # 生成文本的随机程度，值越低越确定
    json_mode: bool = False,  # 是否以 JSON 模式返回结果，默认为 False
):
    """
    向 DeepSeek API 发送请求并获取完成结果。

    参数:
        prompt (str): 提示文本，用于引导模型生成回答。
        system_message (str): 系统消息，用于设定模型的行为。
        model (str): 使用的模型名称。
        temperature (float): 控制生成文本的随机程度。
        json_mode (bool): 是否以 JSON 格式返回结果。

    返回:
        str: 模型生成的回答内容。
    """
    # 调用 OpenAI 客户端的 chat.completions.create 方法，发送请求
    response = client.chat.completions.create(
        model=model,  # 指定使用的模型