Skip to content

AI_CONFIG.md

Latest commit

 

History

History
113 lines (89 loc) · 2.91 KB

AI_CONFIG.md

File metadata and controls

113 lines (89 loc) · 2.91 KB

AI Configuration for Real-Time Code Review

Setup Instructions

1. Environment Variables

Create a .env file in the project root or set these environment variables:

# Set your preferred AI provider: gemini, openai, claude, or mock
export AI_PROVIDER=gemini

# API Keys (only set the one you're using)
export GEMINI_API_KEY=your_gemini_api_key_here
export OPENAI_API_KEY=your_openai_api_key_here
export CLAUDE_API_KEY=your_claude_api_key_here

# Optional: Override default models
export AI_MODEL=gemini-pro

2. Getting API Keys

Gemini (Recommended - Free tier available)

  1. Go to Google AI Studio
  2. Create a new API key
  3. Set AI_PROVIDER=gemini and GEMINI_API_KEY=your_key

OpenAI

  1. Go to OpenAI API Keys
  2. Create a new API key
  3. Set AI_PROVIDER=openai and OPENAI_API_KEY=your_key

Claude

  1. Go to Anthropic Console
  2. Create a new API key
  3. Set AI_PROVIDER=claude and CLAUDE_API_KEY=your_key

3. Development Mode

For testing without API keys, use mock AI:

export AI_PROVIDER=mock

This provides realistic-looking responses without making external API calls.

4. Starting the Server

cd web-ui
go run main.go

The AI features will be available at:

  • POST /api/ai/code-review - Real-time code analysis
  • POST /api/ai/interviewer-questions - Generate follow-up questions
  • POST /api/ai/code-hint - Context-aware hints

Features ✅ WORKING

Real-Time Code Review ✅

  • Overall Score: 0-100 rating of code quality
  • Issues Detection: Bugs, performance, style, logic issues
  • Suggestions: Optimization and best practice recommendations
  • Complexity Analysis: Time/space complexity evaluation
  • Interviewer Feedback: What a real interviewer would say
  • Security: All content is HTML-escaped for safety

Dynamic Interview Questions ✅

  • Context-aware questions based on the user's solution
  • Progressive difficulty based on user performance
  • Go-specific technical probing
  • Edge case exploration
  • Array of 5 relevant questions per request

Smart Hints System ✅

  • 4 levels of hints (subtle nudge → detailed explanation)
  • Context-aware based on current code
  • Educational approach that teaches concepts
  • Progressive hint buttons (Lv1 → Lv2 → Lv3 → Lv4)

API Examples

Code Review

POST /api/ai/code-review
{
  "challengeId": 1,
  "code": "func Sum(a, b int) int { return a + b }",
  "context": "Interview started 5 minutes ago"
}

Get Interview Questions

POST /api/ai/interviewer-questions
{
  "challengeId": 1, 
  "code": "func Sum(a, b int) int { return a + b }",
  "userProgress": "Completed basic solution"
}

Get Hint

POST /api/ai/code-hint
{
  "challengeId": 1,
  "code": "func Sum(a, b int) int { // stuck here }",
  "hintLevel": 2
}