AI Portfolio Analyzer for Indian Stocks
An AI-driven workflow to monitor, analyze, and optimize your Indian stock portfolio using free tools and local LLM inference via Ollama.

---

🚀 Features

• Portfolio Performance Analysis: Calculates total and annualized returns, volatility, Sharpe ratio, and max drawdown.
• Stock Contributions: Analyzes individual stock returns, weights, correlation to market, beta, and key financial metrics.
• Optimization Suggestions: Recommendations for diversification, underperformance, and risk-adjusted rebalancing.
• News Integration: Fetches and categorizes recent news via RSS (Google News) or NewsAPI (optional), then summarizes via local LLM.
• Decision Engine: Generates data-driven, reasoning-backed portfolio advice using your local Ollama LLM.
• Daily Reporting & Scheduling: Produces Markdown reports and visualizations; supports on-demand runs or cron-like scheduling.
• PDF Report Generation: Creates comprehensive daily reports in PDF format with performance metrics, stock returns, news summaries, and optimization suggestions. Additionally, generates a text summary for messaging.

---

🛠 Tech Stack & Free Services

• Python 3.8+
• yfinance: Stock price & financial metrics (free)
• feedparser: RSS news scraping (free)
• NewsAPI.org (optional; free tier)
• Ollama: Local LLM inference (e.g., llama3)
• schedule: Job scheduling in Python
• fpdf: PDF generation for reports
• matplotlib: Data visualization
• Environment Variables: Managed via .env and python-dotenv

---

⚙️ Installation

1. Clone the repository

   git clone https://github.com/mokshablr/ai-stock-portfolio-manager.git
   cd ai-stock-portfolio-manager

2. Create & activate a virtual environment

   python -m venv venv
   source venv/bin/activate  # Linux/macOS
   venv\\Scripts\\activate  # Windows

3. Install dependencies

   pip install -r requirements.txt

4. Configure environment variables
   Copy the example and fill in any API keys:

   cp .env.example .env
   # Edit .env: set NEWS_API_KEY, OLLAMA_API_ENDPOINT, etc.

5. Run Ollama local model
   Ensure Ollama is installed and running:

   ollama run llama3

---

📁 Repository Structure

├── data_cache/                # Cached stock & news data
├── portfolio_ai/              # Workflow outputs & history
│   └── run_history.json
├── analysis_results/          # Performance & contribution JSONs, CSVs, plots
├── decision_reports/          # LLM-generated decision Markdown files
├── news_reports/              # Categorized news & summaries
├── generated_reports/         # Generated PDF and Markdown reports
├── src/                       # Main source files (optional grouping)
│   ├── data_collector.py      # Fetches stock & news data
│   ├── portfolio_analyser.py  # Calculates performance & suggestions
│   ├── news_aggregator.py     # Categorizes & summarizes news
│   ├── llm_connector.py       # Interfaces with local LLM & DecisionEngine
│   ├── report_generator.py    # Generates PDF and text reports (portfolio performance, suggestions, news summaries)
│   ├── portfolio_ai_workflow.py # Orchestrates full workflow
│   └── main.py                # CLI entry point
├── .env.example               # Example environment variables
├── requirements.txt           # Python dependencies
└── README.md                  # This file

---

🚀 Usage

1. On-Demand Run

python main.py --portfolio path/to/portfolio.json

This outputs summary in console and writes detailed reports in respective directories.

2. Schedule Daily Runs

python main.py --schedule --time 18:00

This runs the full workflow every day at 18:00 local time.

---

📄 Configuration

Your portfolio.json file should either be created manually in the following format or generated automatically using sample_portfolio_generator.py:

{
  "stocks": ["RELIANCE", "INFY", "TCS"],
  "holdings": {"RELIANCE": 10, "INFY": 5, "TCS": 8},
  "avg_costs": {"RELIANCE": 2400.50, "INFY": 1450.00, "TCS": 3300.00},
  "company_names": {"RELIANCE": "Reliance Industries Limited", ...}
}

Ensure .env contains:

NEWS_API_KEY=your_newsapi_key
OLLAMA_API_ENDPOINT=your_ollama_endpoint

---

🔍 Example Output

• Analysis Results: analysis_results/portfolio_performance.csv, performance_metrics.json, stock_contributions.json
• Visualizations: analysis_results/portfolio_composition.png, portfolio_performance.png, stock_returns.png
• News Summary: news_reports/news_summary_YYYY-MM-DD.md
• Decision Report: decision_reports/portfolio_decisions_YYYY-MM-DD.md
• Generated Reports: generated_reports/portfolio_report_YYYY-MM-DD.pdf, generated_reports/summary_YYYY-MM-DD.md

---

🤝 Contributing

Contributions, issues, and feature requests are welcome! Feel free to open an issue or submit a pull request.