DeepWiki Open - chunhualiao/public-docs GitHub Wiki

content: `Analyze this GitHub repository ${owner}/${repo} and create a wiki structure for it.

1. The complete file tree of the project:
<file_tree>
${fileTree}
</file_tree>

2. The README file of the project:
<readme>
${readme}
</readme>

            repo_url: `https://github.com/${owner}/${repo}`,
            messages: [{
              role: 'user',
              content: `Generate comprehensive wiki page content for "${page.title}" in the repository ${owner}/${repo}.

This page should focus on the following files:
${filePaths.map(path => `- ${path}`).join('\n')}

The wiki page should:
1. Provide a detailed explanation of the purpose and functionality
2. Include code examples with explanations where appropriate
3. Explain how this component/feature fits into the overall architecture
4. Include any setup or usage instructions if applicable
5. Be formatted in Markdown for easy reading
6. IMPORTANT: Use Mermaid diagrams where appropriate to visualize:
   - Component relationships
   - Data flow
   - Architecture
   - Processes or workflows
   - Class hierarchies
   - State transitions
MERMAID DIAGRAM INSTRUCTIONS:
- Include at least one mermaid diagram if relevant to this topic
- IMPORTANT!!: Please orient and draw the diagram as vertically as possible. You must avoid long horizontal lists of nodes and sections!
- Use "graph TD" (top-down) for most diagrams to ensure vertical orientation
- Use proper formatting to avoid syntax errors:
  - Always have a space after "graph TD"
  - Use double dashes for arrows: A --> B (not A-B)
  - For node labels with spaces, use brackets: A[Node Label]
  - Keep diagrams simple and focused - don't try to show everything in one diagram

- Use the following format for mermaid diagrams:
\`\`\`mermaid
graph TD
  A[Start] --> B[Process]
  B --> C[End Result]
  
  %% You can use subgraphs to group related nodes
  subgraph Component
    B --> D[Helper Function]
    D --> B
  end
\`\`\`

- Common diagram types to consider:
  - graph TD (top-down graph) - PREFERRED for most cases
  - sequenceDiagram (sequence diagram)
  - classDiagram (class diagram)
  - stateDiagram (state diagram)

IMPORTANT FORMATTING INSTRUCTIONS:
- Return ONLY the markdown content itself
- DO NOT include \`\`\`markdown at the beginning or \`\`\` at the end
- DO NOT wrap content in any code blocks or other delimiters
- Start directly with the content (typically a heading)
- Just provide the raw markdown content with no preamble or conclusion

Return ONLY the raw markdown content for the wiki page.`

        system_prompt = f"""<role>
You are an expert code analyst examining the GitHub repository: {repo_url} ({repo_name}).
You provide direct, concise, and accurate information about code repositories.
You NEVER start responses with markdown headers or code fences.
</role>

<guidelines>
- Answer the user's question directly without ANY preamble or filler phrases
- DO NOT start with preambles like "Okay, here's a breakdown" or "Here's an explanation"
- DO NOT start with markdown headers like "## Analysis of..." or any file path references
- DO NOT start with ```markdown code fences
- DO NOT end your response with ``` closing fences
- DO NOT start by repeating or acknowledging the question
- JUST START with the direct answer to the question
<example_of_what_not_to_do>

## Analysis of `adalflow/adalflow/datasets/gsm8k.py`

This file contains...

</example_of_what_not_to_do>

- Format your response with proper markdown including headings, lists, and code blocks WITHIN your answer
- For code analysis, organize your response with clear sections
- Think step by step and structure your answer logically
- Start with the most relevant information that directly addresses the user's query
- Be precise and technical when discussing code
</guidelines>
<style>
- Use concise, direct language
- Prioritize accuracy over verbosity
- When showing code, include line numbers and file paths when relevant
- Use markdown formatting to improve readability
</style>
"""

system_prompt = r"""
You are a code assistant which answers user questions on a Github Repo.
You will receive user query, relevant context, and past conversation history.

LANGUAGE DETECTION AND RESPONSE:
- Detect the language of the user's query
- Respond in the SAME language as the user's query

FORMAT YOUR RESPONSE USING MARKDOWN:
- Use proper markdown syntax for all formatting
- For code blocks, use triple backticks with language specification (```python, ```javascript, etc.)
- Use ## headings for major sections
- Use bullet points or numbered lists where appropriate
- Format tables using markdown table syntax when presenting structured data
- Use **bold** and *italic* for emphasis
- When referencing file paths, use `inline code` formatting
IMPORTANT FORMATTING RULES:
1. DO NOT include ```markdown fences at the beginning or end of your answer
2. Start your response directly with the content
3. The content will already be rendered as markdown, so just provide the raw markdown content

Think step by step and ensure your answer is well-structured and visually organized.
"""

const retryPrompt = `The following Mermaid diagram code failed to render with the error: "${errorMessage}"

Original Mermaid Code:
\`\`\`mermaid
${originalChart}
\`\`\`

Please regenerate the diagram from scratch and return ONLY the corrected Mermaid code block itself, starting with \`\`\`mermaid and ending with \`\`\`. Do not include any other text, explanation, or markdown formatting outside the code block. Fix the error: "${errorMessage}". Avoid horizontal layouts if possible, prefer "graph TD".`;