Skip to content

Latest commit

 

History

History
150 lines (101 loc) · 3.55 KB

QUICK_START.md

File metadata and controls

150 lines (101 loc) · 3.55 KB

🚀 Quick Start Guide - ReCode with Local LLMs

✅ Status

All errors have been fixed! The application is now fully compatible with local LLMs (Ollama).

🔧 What Was Fixed

  1. Undefined AI Configuration Variables - Fixed in api/solution/index.js
  2. Poor Error Handling - Enhanced in api/solution/index.js and api/usage/index.js
  3. Health Check Endpoint - Completely rewritten in api/health.js
  4. Local LLM Compatibility - Verified all models are accessible

📋 Prerequisites

✅ Ollama is running on port 11434 ✅ Models installed:

  • qwen2.5-coder:7b
  • qwen2.5:7b
  • deepseek-r1:7b
  • nomic-embed-text:latest

🎯 How to Start

Option 1: Using Vercel Dev (Recommended)

npm run dev:vercel

Option 2: Using Vite Only (Frontend Only)

npm run dev

Note: The API endpoints require Vercel dev server to work properly.

🧪 Testing

Once the server is running, open your browser to:

http://localhost:3000

Then:

  1. Click "Get Solution"
  2. Enter a problem name (e.g., "Two Sum")
  3. Select language (e.g., "Python")
  4. Click "Generate Solution"

Expected Behavior:

  • Solution should generate within 10-30 seconds
  • Browser console should show: [LLM] 🟢 Config: Task=coding Provider=Ollama Model=qwen2.5-coder:7b
  • No 500 errors should appear

🐛 Troubleshooting

"AI service unavailable" error:

# Check if Ollama is running
curl http://127.0.0.1:11434/api/tags

# If not running, start it:
ollama serve

"Database connection failed" error:

  • Check your .env file has correct MONGO_URI
  • Verify internet connection (if using MongoDB Atlas)
  • Check MongoDB Atlas IP whitelist includes your IP

WebSocket errors in console:

  • These are harmless HMR (Hot Module Reload) messages
  • They don't affect API functionality
  • Can be safely ignored

📊 Monitoring

Check Health Status:

Invoke-RestMethod -Uri "http://localhost:3000/api/health" | ConvertTo-Json -Depth 10

Check Usage Limits:

Invoke-RestMethod -Uri "http://localhost:3000/api/usage" | ConvertTo-Json

📝 Configuration

Current .env settings for local LLM:

AI_PROVIDER=ollama
OLLAMA_BASE_URL=http://127.0.0.1:11434/v1
OLLAMA_MODEL_REASONING=deepseek-r1:7b
OLLAMA_MODEL_CODING=qwen2.5-coder:7b
OLLAMA_MODEL_EXPLANATION=qwen2.5:7b

🎓 Model Usage

  • Reasoning (DSA Logic): deepseek-r1:7b - Used for complex algorithmic thinking
  • Coding (Generation): qwen2.5-coder:7b - Used for generating code solutions
  • Explanation (Simple Text): qwen2.5:7b - Used for explanations and summaries

✨ Key Improvements

  1. Better Error Messages:

    • Instead of: "Failed to generate solution"
    • Now: "AI service unavailable. Please ensure Ollama is running if using local models."

  2. Detailed Health Checks:

    • MongoDB connection status
    • Ollama accessibility and model availability
    • Redis status (optional)

  3. Improved Logging:

    • All API calls now log provider, model, and task type
    • Stack traces for debugging
    • Development mode shows detailed errors

📖 Documentation

  • Full Fix Details: See FIXES_SUMMARY.md
  • API Endpoints: Check api/ folder
  • Model Configuration: See api/_lib/aiConfig.js

🤝 Support

If you encounter any issues:

  1. Check browser console (F12)
  2. Check terminal where vercel dev is running
  3. Run health check: http://localhost:3000/api/health
  4. Verify Ollama is running: ollama list

You're all set! Start the server and enjoy using local LLMs! 🎉