analyzer/prompts.py

def get_code_analysis_prompt(code: str, language: str = "auto-detect", model: str = None) -> str:
    """Generate a focused prompt for practical code analysis."""
    return f"""
You are an expert code reviewer. Analyze this code and provide comprehensive feedback.

Code to analyze:
{code}

First, identify the programming language, then analyze the code. Provide a focused analysis with complete, readable sentences. Do NOT use markdown symbols like ### or ** in your response.

0. DETECTED_LANGUAGE: Respond with ONLY the programming language name (e.g., python, javascript, rust, cpp, java, go, php, ruby, swift, kotlin, typescript, csharp, html, css). Make sure to detect the language accurately based on the actual code.

1. QUALITY_SCORE: Rate 0-100 (consider bugs, readability, maintainability). Respond with ONLY the number.

2. SUMMARY: One complete sentence describing what this code does

3. BUG_DETECTION: 
   - List actual bugs or logical errors found
   - Include potential crashes or exceptions
   - Mention edge cases not handled
   (Write complete sentences, skip if none found)

4. CODE_QUALITY_ISSUES:
   - Poor naming or structure problems
   - Code readability issues  
   - Maintainability concerns
   (Focus on practical fixes, write complete sentences)

5. SECURITY_VULNERABILITIES:
   - Injection risks (SQL, XSS, etc.)
   - Insecure data handling
   - Authentication/authorization flaws
   (Only include actual security risks, write complete sentences)

6. QUICK_FIXES: 
   - Top 3 specific improvements with examples
   - Focus on high-impact, easy changes
   - Write complete actionable sentences

Format each section as clear, complete sentences. Be specific and actionable. Skip sections if no issues found.
"""

def get_github_analysis_prompt(repo_structure: str, main_files: str) -> str:
    """Generate prompt for GitHub repository analysis."""
    return f"""
Analyze this GitHub repository structure and key files. Provide clear, complete analysis without using markdown symbols.

Repository Structure:
{repo_structure}

Main Files Content:
{main_files}

Provide analysis focusing on:

1. PROJECT_OVERVIEW: Write one clear sentence about what this project does

2. ARCHITECTURE_QUALITY: 
   - Project structure assessment (write complete sentences)
   - Code organization quality (write complete sentences)
   - Missing important files like tests, docs, etc. (write complete sentences)

3. CRITICAL_ISSUES:
   - Security vulnerabilities across files (write complete sentences)
   - Major bugs or design flaws (write complete sentences)
   - Dependencies/configuration problems (write complete sentences)

4. IMPROVEMENT_PRIORITIES:
   - Top 5 specific things to fix first (write complete sentences)
   - Missing features or best practices (write complete sentences)
   - Code quality improvements needed (write complete sentences)

5. ONBOARDING_GUIDE: (This is a critical section)
   - Provide a step-by-step guide for a new developer to run this project.
   - List key commands for installation, testing, and running (e.g., `npm install`, `pytest`).
   - Infer any required environment variables (e.g., `DATABASE_URL`, `API_KEY`).
   - If setup instructions are in the README, summarize them here.

6. TECH_STACK_RATIONALE: (Provide your best assessment)
   - Identify the primary framework, language, and package manager.
   - Infer the likely reason for choosing this stack in one clear sentence.
   - Mention any interesting or unusual libraries and their purpose.

7. API_ENDPOINT_SUMMARY: (Scan all provided files for this)
   - List the main API endpoints found (e.g., `GET /api/users`, `POST /api/products`).
   - Provide a one-sentence guess for what each endpoint does based on the code.
   - If this is not a web service, you must explicitly state 'Not a web service'.

Write clear, complete sentences without markdown symbols. You must attempt to fill out every section. Be practical and focus on actionable feedback for the repository owner.
"""

def get_comparison_prompt(code: str, language: str = "auto-detect") -> str:
    """Generate a prompt for multi-model comparison."""
    return f"""
As an expert code reviewer, analyze this {language} code:
{code}

Provide a concise but thorough analysis covering:
- Main functionality
- Code quality (rate 0-100)
- Critical issues
- Top 3 improvements
- Security/performance concerns

Be specific and actionable in your feedback.
"""