Skip to content

QUICK_REFERENCE.md

Latest commit

 

History

History
277 lines (207 loc) · 5.11 KB

QUICK_REFERENCE.md

File metadata and controls

277 lines (207 loc) · 5.11 KB

VersaAI - Quick Reference Card

🚀 Quick Commands

Start Services

# Code Assistant CLI
python versaai_cli.py

# Editor Bridge Server
python start_editor_bridge.py

# Test Bridge
python test_editor_bridge.py

Download Models

# Single model
python -m versaai.cli.download_models --model deepseek-1.3b

# Recommended set
python -m versaai.cli.download_models --preset recommended

# All models
python -m versaai.cli.download_models --preset all

💬 CLI Commands

Command Description
/help Show help
/models List models
/switch MODEL Change model
/load FILE Load file context
/clear Clear history
/save FILE Save conversation
/explain Explain code
/refactor Refactor code
/test Generate tests
/debug Debug help
/exit Quit

🤖 Available Models

Local Models (Free)

  • DeepSeek 1.3B - Fast (834MB, 2-4GB RAM)
  • DeepSeek 6.7B - Balanced (3.9GB, 8-12GB RAM)
  • StarCoder2 7B - Quality (4.1GB, 8-16GB RAM)
  • CodeLlama 7/13/34B - Best (4-19GB, 8-32GB+ RAM)
  • Qwen2.5 7/14B - Multilingual (4-8GB, 8-24GB RAM)

API Models (Paid)

  • GPT-4 - Best reasoning
  • GPT-3.5 - Fast & cheap
  • Claude 3 - Long context

🔌 WebSocket API

Connection

const ws = new WebSocket('ws://localhost:8765');

Message Types

// Code completion
{type: 'completion', context: {...}}

// Chat
{type: 'chat', message: '...', session_id: '...'}

// Explain code
{type: 'explain', code: '...', language: '...'}

// Refactor
{type: 'refactor', code: '...', language: '...'}

// Debug
{type: 'debug', code: '...', error_message: '...'}

// Generate tests
{type: 'test', code: '...', language: '...'}

// Health check
{type: 'ping'}

⚙️ Configuration

Config File

Location: ~/.versaai/config.yaml

models:
  default: deepseek-6.7b
  
preferences:
  temperature: 0.7
  max_tokens: 2048
  
editor_bridge:
  host: localhost
  port: 8765

Environment Variables

export OPENAI_API_KEY="sk-..."
export ANTHROPIC_API_KEY="sk-ant-..."
export VERSAAI_MODELS_DIR="~/.versaai/models"

📁 File Locations

Type Location
Models ~/.versaai/models/
Config ~/.versaai/config.yaml
Logs ~/.versaai/versaai.log
RAG Index ~/.versaai/rag_index/

🐛 Troubleshooting

Server won't start

# Check port
lsof -i :8765

# Kill process
kill -9 <PID>

# Use different port
python start_editor_bridge.py --port 8766

Out of memory

  • Use smaller model (1.3B)
  • Close other apps
  • Reduce max_tokens

Slow responses

  • Use DeepSeek 1.3B
  • Enable caching
  • Use GPU if available

No models found

# Download model
python -m versaai.cli.download_models --model deepseek-1.3b

# Verify
ls ~/.versaai/models/

📖 Documentation

Document Purpose
docs/USER_GUIDE.md Complete user manual
docs/CODE_EDITOR_INTEGRATION.md Integration guide
EDITOR_INTEGRATION_README.md Quick start
QUICKSTART.md 5-minute start

🔗 Python API

from versaai import VersaAI

# Initialize
ai = VersaAI(model='deepseek-6.7b')

# Chat
response = ai.chat("How do I use async in Python?")

# Generate code
code = ai.generate_code("Create a REST API with FastAPI")

# Explain code
explanation = ai.explain_code(code, language='python')

# Generate tests
tests = ai.generate_tests(code, language='python')

# Load file context
ai.load_file('main.py')
ai.chat("How can I optimize this?")

🎯 Common Tasks

Generate Function

You: Create a function to validate email addresses

VersaAI: [Generates complete function with validation]

Explain Code

You: /explain
[Paste code]

VersaAI: [Detailed explanation]

Optimize Code

You: /load myfile.py
You: Optimize the process_data function

VersaAI: [Optimization suggestions]

Generate Tests

You: /test
[Paste function]

VersaAI: [Complete unit tests]

Debug Error

You: /debug
[Paste code and error]

VersaAI: [Root cause + fix]

🚀 Getting Started (3 Steps)

  1. Install VersaAI

    cd VersaAI
pip install -e .

  2. Download Model

    python -m versaai.cli.download_models --model deepseek-1.3b

  3. Start Using

    # CLI
python versaai_cli.py

# Or Editor Bridge
python start_editor_bridge.py

💡 Pro Tips

  • Use auto mode for intelligent model selection
  • Load files for context-aware help: /load file.py
  • Cache responses - ask similar questions for speed
  • Try different models - use /switch to compare
  • Save conversations - /save for future reference
  • Local models for privacy, API for quality

📞 Support

  • Docs: docs/ directory
  • Examples: examples/ directory
  • Issues: GitHub Issues
  • Logs: ~/.versaai/versaai.log

🎓 Learn More

  • Complete Guide: docs/USER_GUIDE.md
  • Integration: docs/CODE_EDITOR_INTEGRATION.md
  • Quick Start: QUICKSTART.md

🚀 Happy Coding with VersaAI!