🚀 Alpha AI Trading Bot - Kompletna Instrukcja Wdrożenia

Projekt: alpha_data_scraper_ai | Architektura: Multi-broker + Claude AI
Broker wsparty: XTB (IKE) + yfinance | Raporty: Daily/Weekly/Monthly | Deploy: Docker 24/7

📋 QUICK START (Pick One)

Opcja	Opis	Czas	Dla kogo
A	Test Parser XTB	5 min	Chcę sprawdzić czy ładuje moje dane
B	Pełne uruchomienie lokalne	30 min	Chcę pracować z systemem na laptopie
C	Docker 24/7	45 min	Chcę to w chmurze / serwer
D	Wszystko na raz	N/A	Chcę mieć wszystko w ZIP

A️⃣ TEST PARSER - XTB Data Validation (5 min)

Krok 1: Dodaj swoje XLSX pliki

# Umieść pliki w ./data/
cp /path/to/IKE_54250698_2026-03-07_2026-04-07.xlsx ./data/
cp /path/to/PLN_51514835_2026-03-07_2026-04-07.xlsx ./data/

Krok 2: Uruchom test

python3 test_parser.py

Oczekiwany output:

============================================================
TEST PARSERA XTB – CLOSED POSITIONS + CASH OPERATIONS
============================================================

✅ Wczytano 47 pozycji:
   AAPL.US    |     5.0000 szt. | avg  152.30
   VWRL.L     |    20.0000 szt. | avg   25.40
   ...

✅ Cash Operations:
   Transakcji: 156
   Net cash flow: -45000.00 PLN
   Dywidendy brutto: 3200.50 PLN
   Podatek: 640.10 PLN
   Dywidendy netto: 2560.40 PLN

✅ Wartość portfela (yfinance): 950000.00 PLN

Co parser robi?

✅ Czyta XLSX z XTB (Closed Positions, Cash Operations)
✅ Normalizuje tickery (.US → brak, .PL → .WA, .UK → .L)
✅ Liczy dywidendy brutto/netto (19% Belka)
✅ Pobiera aktualne ceny z yfinance
✅ Eksportuje raport: reports/daily_YYYY-MM-DD.md

B️⃣ FULL LOCAL RUNTIME (30 min)

Krok 1: Setup environment

# Skopiuj .env
cp .env.example .env

# EDYTUJ .env:
nano .env

Wymagane zmienne w .env:

ANTHROPIC_API_KEY=sk-ant-xxxxxxxxxxxx          # Twój Claude API key
ACTIVE_BROKER=xtb
XTB_DATA_PATH=./data/
TZ=Europe/Warsaw

Opcjonalne:

TELEGRAM_BOT_TOKEN=123456:ABCdefgh            # Dla powiadomień
TELEGRAM_CHAT_ID=-1001234567890

Krok 2: Zainstaluj zależności

pip install -r requirements.txt

Krok 3: Uruchom testy

# Test 1: Parser XTB
python3 test_parser.py

# Test 2: Claude AI connection
python3 -c "from claude_ai import ClaudeAIIntegration; c = ClaudeAIIntegration(); print(c.test_connection())"

Krok 4: Uruchom raporty ręcznie

# Daily analysis
python3 daily_analysis.py

# Weekly analysis
python3 weekly_analysis.py

# Monthly dividend report
python3 monthly_dividend_report.py

Krok 5: Uruchom scheduler (tło)

python3 scheduler.py

Harmonogram:

Codziennie 08:30 → Daily risk analysis (Bridgewater)
Każdy poniedziałek 09:00 → Weekly cash flow report
1. dzień miesiąca 10:00 → Monthly dividend analysis (Harvard)

Krok 6: Webhook TradingView (port 8000)

W osobnym terminalu:

python3 webhook_server.py

Endpoint: http://localhost:8000/webhook (POST)
Health: http://localhost:8000/health (GET)

Struktura output'ów

reports/
├── daily_2026-04-07.md          # Risk analysis
├── weekly_2026-04-07_09-00.md   # Cash flow
└── monthly_dividend_2026-04-07.md  # Dividend strategy

C️⃣ DOCKER 24/7 PRODUCTION (45 min)

Wymagania

# Sprawdź czy masz Docker
docker --version    # Docker version 20.10+
docker-compose --version  # docker-compose 2.0+

# Jeśli nie masz - zainstaluj:
# Mac: brew install docker
# Linux: sudo apt install docker.io docker-compose
# Windows: https://docker.com/download

Krok 1: Setup .env

cp .env.example .env
nano .env
# Wpisz ANTHROPIC_API_KEY

Krok 2: Build & Run

./build_and_run.sh

Co się dzieje:

✅ Sprawdza Docker (docker-compose --version)
✅ Buduje images (docker-compose build)
✅ Uruchamia 2 serwisy:
- alpha-ai-scheduler (scheduler.py) — raporty co noc
- alpha-ai-webhook (webhook_server.py) — port 8000

Krok 3: Monitoruj logi

# Scheduler (raporty)
docker-compose logs -f alpha-ai-scheduler

# Webhook (TradingView alerts)
docker-compose logs -f alpha-ai-webhook

# Wszystko
docker-compose logs -f

Zarządzanie serwisem

# Status
docker-compose ps

# Restart
docker-compose restart

# Stop
docker-compose down

# Clean rebuild
docker-compose down --volumes
./build_and_run.sh

Skalowanie na serwer (np. AWS/DigitalOcean)

# 1. SSH na serwer
ssh ubuntu@your-server.com

# 2. Klonuj repo
git clone https://github.com/yourusername/alpha_data_scraper_ai
cd alpha_data_scraper_ai

# 3. Setup .env
echo "ANTHROPIC_API_KEY=sk-ant-xxxx" > .env

# 4. Uruchom
./build_and_run.sh

# 5. Setup reverse proxy (nginx/caddy) na port 8000
# (dokumentacja w sekcji "Production Deployment")

D️⃣ FULL PROJECT ZIP

Zawarte w alpha_ai_trading_bot.zip:

alpha_data_scraper_ai/
├── brokers/                    # Multi-broker abstraction
│   ├── __init__.py
│   ├── broker_interface.py     # ABC interface
│   ├── xtb_broker.py           # XTB parser (XLSX reader)
│   ├── ibkr_broker.py          # Interactive Brokers
│   ├── alpaca_broker.py        # Alpaca API
│   ├── hybrid_broker.py        # XTB + Alpaca combo
│   └── broker_factory.py       # Factory pattern
│
├── config/
│   ├── brokers.yaml            # Broker config + paths
│   └── notifications.yaml      # Telegram/Discord
│
├── prompts.py                  # 10 Wall Street prompts
├── claude_ai.py                # Claude API integration
├── ai_engine.py                # Broker + AI pipeline
│
├── daily_analysis.py           # Risk analysis (Bridgewater)
├── weekly_analysis.py          # Cash flow report
├── monthly_dividend_report.py  # Dividend strategy (Harvard)
├── notifications.py            # Telegram/Discord sender
│
├── scheduler.py                # APScheduler (Poland tz)
├── webhook_server.py           # FastAPI + TradingView
├── test_parser.py              # XTB XLSX validator
│
├── requirements.txt            # Python dependencies
├── Dockerfile                  # Docker image
├── docker-compose.yml          # 2-container setup
├── build_and_run.sh           # Docker automation
│
├── .env.example                # Template (edit and save as .env)
├── .gitignore                  # Git ignore
├── README.md                   # This file
│
├── data/                       # Your XLSX files
│   ├── IKE_54250698_2026-03-07_2026-04-07.xlsx
│   └── PLN_51514835_2026-03-07_2026-04-07.xlsx
│
└── reports/                    # Output directory
    ├── daily_2026-04-07.md
    ├── weekly_2026-04-07_09-00.md
    └── monthly_dividend_2026-04-07.md

🔑 WALL STREET PROMPTS (Included)

System zawiera 10 profesjonalnych prompt'ów:

goldman_screener — Stock screening (P/E, growth)
morgan_dcf — DCF valuation model
bridgewater_risk — Portfolio risk analysis ⭐ (daily default)
harvard_dividend — Dividend strategy ⭐ (monthly)
citadel_technical — Technical analysis (TradingView)
mckinsey_macro — Macro economic factors
jpmorgan_earnings — Earnings analysis
blackrock_portfolio — Asset allocation
bain_competitive — Competitive analysis
renaissance_quant — Quantitative models

📊 DATA FLOW

XTB XLSX Files (./data/)
         ↓
   XTBBroker (parser)
         ↓
   get_positions() → yfinance prices
   get_cash_summary() → dividends + taxes (Belka 19%)
   get_portfolio_value()
         ↓
   AI Engine + Claude
         ↓
   Wall Street Prompts (Bridgewater, Harvard, etc.)
         ↓
   Reports (MD files)
         ↓
   Notifications (Telegram/Discord)

🚨 TROUBLESHOOTING

Parser nie czyta XLSX

# Sprawdź ścieżki
ls -la ./data/

# Sprawdź format
python3 -c "import openpyxl; wb = openpyxl.load_workbook('./data/IKE_54250698_2026-03-07_2026-04-07.xlsx'); print(wb.sheetnames)"
# Powinno pokazać: ['Closed Positions', 'Cash Operations', ...]

Claude API error

# Sprawdź API key
echo $ANTHROPIC_API_KEY

# Test connection
python3 -c "from claude_ai import ClaudeAIIntegration; c = ClaudeAIIntegration(); print(c.test_connection())"

Docker doesn't start

# Logi
docker-compose logs

# Rebuild
docker-compose down --volumes
docker-compose build --no-cache
docker-compose up

📞 WSPARCIE

API Docs: https://docs.anthropic.com
XTB Format: Sprawdź config/brokers.yaml
Issues: Check GitHub issues or contact support

📝 LICENSE & CONTEXT

Investor: Mati (Poland)
Accounts: XTB IKE (PLN, 19% Belka tax)
Created: Apr 2026
Architecture: Based on Grok blueprint + Claude enhancement

✅ CHECKLIST PRZED PRODUKCJĄ

.env zawiera ANTHROPIC_API_KEY
XLSX pliki w ./data/
Docker zainstalowany (dla C)
Telegram token (opcjonalne)
Cron/systemd dla scheduler (dla C)
Backup danych (co tydzień)
Monitoring alertów (Telegram/Discord)

You're ready! Pick option A, B, C, or D and go. 🚀

FilesExpand file tree

INSTRUKCJA_PL.md

Latest commit

History