Fix critical setup blockers and add setup documentation

SETUP.md

@@ -0,0 +1,197 @@
+# Moon Dev AI Agents - Setup Guide
+
+## Quick Start (Critical Steps)
+
+### 1. Activate Conda Environment
+```bash
+conda activate tflow
+```
+
+### 2. Install Dependencies
+```bash
+pip install -r requirements.txt
+```
+
+This will install all required packages including:
+- pandas, numpy (data handling)
+- anthropic, openai, groq, deepseek (AI models)
+- python-dotenv (environment variables)
+- termcolor (console colors)
+- backtesting, pandas-ta, ta-lib (trading/backtesting)
+- And 20+ more packages
+
+### 3. Create .env File
+```bash
+cp .env_example .env
+```
+
+Then edit `.env` and add your actual API keys:
+
+**CRITICAL - Required for basic operation:**
+- `ANTHROPIC_KEY` - Claude API (for AI agents)
+- `BIRDEYE_API_KEY` - Solana token data
+- `SOLANA_PRIVATE_KEY` - Your wallet (Base58 format)
+- `RPC_ENDPOINT` - Helius or other Solana RPC
+
+**Optional - For specific features:**
+- `OPENAI_KEY` - GPT models
+- `GROQ_API_KEY` - Fast inference
+- `DEEPSEEK_KEY` - Reasoning models
+- `MOONDEV_API_KEY` - Custom signals
+- `COINGECKO_API_KEY` - Token metadata
+- `YOUTUBE_API_KEY` - YouTube integration
+- `TWILIO_*` - Phone agent
+- `TWITTER_*` - Tweet agent
+
+### 4. Configure Trading Settings
+
+Edit `src/config.py` to set:
+- `MONITORED_TOKENS` - Which tokens to trade
+- `usd_size` - Position size (default: $25)
+- `MAX_LOSS_USD` - Circuit breaker (default: $25)
+- `MINIMUM_BALANCE_USD` - Minimum balance (default: $50)
+- Wallet address (line 32) - Replace placeholder
+
+### 5. Run the System
+
+**Option A: Run main orchestrator**
+```bash
+python src/main.py
+```
+
+**Option B: Run individual agents**
+```bash
+python src/agents/trading_agent.py
+python src/agents/risk_agent.py
+python src/agents/rbi_agent.py
+python src/agents/sentiment_agent.py
+# ... any agent can run standalone
+```
+
+**Option C: Backtest strategies**
+```bash
+python src/agents/rbi_agent.py
+# Provide YouTube URL, PDF, or strategy description
+```
+
+## What Was Fixed
+
+This setup includes recent critical fixes:
+
+1. **Fixed import error in main.py** (line 12)
+   - Changed `from config import *` → `from src.config import *`
+   - Now consistent with all agents
+
+2. **Fixed Linux-incompatible path in config.py** (line 101)
+   - Changed `/Volumes/Moon 26/OBS` → `os.path.expanduser('~/Videos/OBS')`
+   - Now works cross-platform (Mac/Linux/Windows)
+
+## System Architecture
+
+```
+src/
+├── main.py              # Orchestrator (runs multiple agents in loop)
+├── config.py            # Global configuration
+├── nice_funcs.py        # Shared trading utilities (1,200+ functions)
+├── agents/              # 43 specialized AI agents
+│   ├── trading_agent.py      # LLM trading decisions
+│   ├── risk_agent.py         # Portfolio risk management
+│   ├── rbi_agent.py          # Research-Backtest-Implement
+│   ├── sentiment_agent.py    # Twitter sentiment
+│   └── ...38 more agents
+├── models/              # LLM provider abstraction
+│   └── model_factory.py      # Unified interface for Claude/GPT/Groq/etc
+├── strategies/          # User-defined strategies
+└── data/                # Agent outputs, analysis results
+```
+
+## Agent Activation
+
+By default, **all agents are OFF** in `main.py`. To enable agents:
+
+Edit `src/main.py` line 29-34:
+```python
+ACTIVE_AGENTS = {
+    'risk': True,       # Enable risk management
+    'trading': True,    # Enable LLM trading
+    'strategy': False,  # Strategy-based trading
+    'copybot': False,   # CopyBot monitoring
+    'sentiment': False, # Twitter sentiment
+}
+```
+
+Or run agents directly:
+```bash
+python src/agents/whale_agent.py      # Whale activity tracker
+python src/agents/funding_agent.py    # Funding rate analysis
+python src/agents/clips_agent.py      # Video clipping automation
+```
+
+## Testing the Setup
+
+**Test 1: Check imports**
+```bash
+python -c "from src.config import *; print('Config loaded ✓')"
+python -c "from src.models.model_factory import ModelFactory; print('Models loaded ✓')"
+```
+
+**Test 2: Check API keys**
+```bash
+python -c "from dotenv import load_dotenv; import os; load_dotenv(); print('ANTHROPIC_KEY' in os.environ)"
+```
+
+**Test 3: Run a simple agent**
+```bash
+python src/agents/sentiment_agent.py
+```
+
+## Troubleshooting
+
+**Problem: ModuleNotFoundError**
+- Solution: Run `pip install -r requirements.txt` in `tflow` conda env
+
+**Problem: API key errors**
+- Solution: Check `.env` file exists and has valid keys
+
+**Problem: "No such file or directory: /Volumes/Moon 26/OBS"**
+- Solution: Already fixed! Update to latest config.py
+
+**Problem: Import error in main.py**
+- Solution: Already fixed! Update to latest main.py
+
+**Problem: ta-lib installation fails**
+- Solution Linux: `sudo apt-get install ta-lib`
+- Solution Mac: `brew install ta-lib`
+
+## Next Steps
+
+1. **Start with risk agent**: Validates your setup and checks portfolio
+2. **Try RBI agent**: Generate backtests from YouTube videos
+3. **Enable trading agents**: Start with paper trading (low position sizes)
+4. **Monitor performance**: Check `src/data/` for agent outputs
+5. **Join Discord**: Community support and weekly updates
+
+## Security Reminders
+
+- **NEVER commit .env file** (already in .gitignore)
+- **NEVER print API keys** in logs
+- **START WITH SMALL POSITIONS** ($25 default)
+- **USE CIRCUIT BREAKERS** (MAX_LOSS_USD setting)
+- **ROTATE KEYS** if accidentally exposed
+
+## Resources
+
+- YouTube: Moon Dev channel (weekly updates)
+- Discord: Community and support
+- GitHub: Issues and discussions
+- CLAUDE.md: Detailed development guide
+
+---
+
+**System Status After Setup:**
+- ✅ Dependencies installed
+- ✅ .env configured with API keys
+- ✅ config.py updated with wallet/settings
+- ✅ Import errors fixed
+- ✅ Paths Linux-compatible
+- 🚀 Ready to run!

src/config.py

@@ -98,7 +98,7 @@
 
 # Real-Time Clips Agent Settings 🎬
 REALTIME_CLIPS_ENABLED = True
-REALTIME_CLIPS_OBS_FOLDER = '/Volumes/Moon 26/OBS'  # Your OBS recording folder
+REALTIME_CLIPS_OBS_FOLDER = os.path.expanduser('~/Videos/OBS')  # OBS recording folder (cross-platform)
 REALTIME_CLIPS_AUTO_INTERVAL = 120  # Check every N seconds (120 = 2 minutes)
 REALTIME_CLIPS_LENGTH = 2  # Minutes to analyze per check
 REALTIME_CLIPS_AI_MODEL = 'groq'  # Model type: groq, openai, claude, deepseek, xai, ollama

src/main.py

@@ -9,7 +9,7 @@
 from dotenv import load_dotenv
 import time
 from datetime import datetime, timedelta
-from config import *
+from src.config import *
 
 # Add project root to Python path
 project_root = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))