ArticleAgent

from qwen_agent.agents.article_agent import ArticleAgent

agent = ArticleAgent(
    llm={'model': 'qwen-max-latest', 'model_type': 'qwen_dashscope'},
    system_message='You are a skilled article writer.',
    files=['./reference1.pdf', './reference2.docx'],
    rag_cfg={'lang': 'en', 'max_ref_token': 4000}
)

from qwen_agent.agents.article_agent import ArticleAgent

agent = ArticleAgent(
    llm={'model': 'qwen-max-latest', 'model_type': 'qwen_dashscope'},
    system_message='You are an expert technical writer.'
)

messages = [{
    'role': 'user',
    'content': 'Write a comprehensive article about machine learning basics'
}]

for response in agent.run(messages=messages, lang='en', full_article=True):
    print(response[-1]['content'])

from qwen_agent.agents.article_agent import ArticleAgent

agent = ArticleAgent(
    llm={'model': 'qwen-max-latest', 'model_type': 'qwen_dashscope'}
)

existing_content = """
# Introduction to Neural Networks

Neural networks are computational models inspired by the human brain...
"""

messages = [{
    'role': 'user',
    'content': f'Continue this article:\n\n{existing_content}\n\nAdd a section about training neural networks.'
}]

for response in agent.run(messages=messages, lang='en', full_article=False):
    print(response[-1]['content'])

from qwen_agent.agents.article_agent import ArticleAgent

# Load reference documents
agent = ArticleAgent(
    llm={'model': 'qwen-max-latest', 'model_type': 'qwen_dashscope'},
    files=[
        './research_paper1.pdf',
        './research_paper2.pdf',
        './dataset_documentation.pdf'
    ],
    rag_cfg={
        'lang': 'en',
        'max_ref_token': 8000,  # More context for research
        'search_method': 'hybrid'
    },
    system_message='You are a research writer. Cite sources when using information.'
)

messages = [{
    'role': 'user',
    'content': 'Write an article summarizing the key findings from the research papers'
}]

for response in agent.run(messages=messages, lang='en', full_article=True):
    content = response[-1]['content']
    if 'Search for relevant information' in content:
        print('Researching references...')
    else:
        print(content)

agent = ArticleAgent(
    llm={'model': 'qwen-max-latest', 'model_type': 'qwen_dashscope'},
    system_message="""
    You are a technical writer specializing in developer documentation.
    
    Guidelines:
    - Use clear, concise language
    - Include code examples where appropriate
    - Structure content with headers and sections
    - Use bullet points for lists
    - Explain technical concepts simply
    """
)

# English article
messages_en = [{'role': 'user', 'content': 'Write about AI safety'}]
for response in agent.run(messages=messages_en, lang='en', full_article=True):
    print(response[-1]['content'])

# Chinese article  
messages_zh = [{'role': 'user', 'content': '写一篇关于AI安全的文章'}]
for response in agent.run(messages=messages_zh, lang='zh', full_article=True):
    print(response[-1]['content'])

agent = ArticleAgent(
    llm={'model': 'qwen-max-latest', 'model_type': 'qwen_dashscope'},
    function_list=['code_interpreter'],  # For code examples
    system_message='You are a developer advocate writing technical blog posts.'
)

messages = [{
    'role': 'user',
    'content': 'Write a blog post about async/await in Python with code examples'
}]

for response in agent.run(messages=messages, full_article=True):
    print(response[-1]['content'])

agent = ArticleAgent(
    llm={'model': 'qwen-max-latest', 'model_type': 'qwen_dashscope'},
    files=['./research/*.pdf'],
    rag_cfg={'lang': 'en', 'max_ref_token': 12000}
)

messages = [{
    'role': 'user',
    'content': 'Write a literature review summarizing these research papers'
}]

for response in agent.run(messages=messages, full_article=True):
    print(response[-1]['content'])

outline = """
# Guide to Docker

## 1. Introduction
## 2. Installation  
## 3. Basic Commands
## 4. Dockerfile Basics
## 5. Docker Compose
"""

messages = [{
    'role': 'user',
    'content': f'Expand this outline into a full guide:\n{outline}'
}]

for response in agent.run(messages=messages, full_article=True):
    print(response[-1]['content'])