Finlearner v0.1.1 - Initial Release

v0.1.1 Patch Release (2026-02-01) 🐞

This is a patch release addressing a critical bug in the Deep Learning module and updating documentation.

🛠️ Fixes

• Deep Learning Models: Fixed an IndexError in the predict() method for LSTM, GRU, CNN-LSTM, Transformer, and Ensemble models caused by incorrect data slicing logic.
• Demo Notebook: Updated examples/notebooks/finlearner_demo.ipynb to correct API usage, improve plotting with aligned axes, and include MAE metrics.

Finlearner v0.1.0 - Initial Release

FinLearner v0.1.0 Release Notes

We are thrilled to announce the major release of FinLearner v0.1.0! This release marks a significant milestone, bringing a complete suite of financial analysis and deep learning tools, stable APIs, and ensuring full compatibility across platforms (including Windows).

Major Features

Portfolio Optimization Module

• Markowitz Mean-Variance: Optimize portfolios for Maximum Sharpe Ratio using Monte Carlo simulation.
• Black-Litterman Model: Integrate investor views with market equilibrium for robust asset allocation.
• Risk Parity: Construct portfolios with equal risk contribution from each asset.

Advanced Risk Metrics

• Value at Risk (VaR): Calculate VaR using Historical, Parametric (Gaussian), and Monte Carlo methods.
• Conditional VaR (CVaR): Assess tail risk with Expected Shortfall calculations.
• Drawdown Analysis: Calculate Maximum Drawdown and Calmar Ratio.
• Cornish-Fisher Expansion: Adjusted VaR for non-normal return distributions.

Machine Learning & Deep Learning

• Gradient Boosting Wrapper: Unified interface for XGBoost and LightGBM with automatic feature importance.
• Deep Learning Models: Production-ready implementations of LSTM, GRU, CNN-LSTM, Transformer, and Ensemble models.
• Anomaly Detection: Variational Autoencoder (VAE) for detecting anomalies in price patterns.

Options Pricing

• Black-Scholes-Merton: Price European Call/Put options and calculate all Greeks (Delta, Gamma, Vega, Theta, Rho).
• PINN: Physics-Informed Neural Networks for solving the Black-Scholes PDE.

Technical Analysis

• Comprehensive Indicators: Over 20+ indicators including RSI, MACD, Bollinger Bands, ATR, Ichimoku Cloud, and OBV.
• One-Shot Enrichment: TechnicalIndicators.add_all() to instantly generate a feature-rich dataset.

Enhancements & Fixes

• Windows Compatibility: Fixed encoding issues (Unicode/Emoji support) for Windows terminals.
• yfinance Compatibility: Updated data fetching logic to handle the new multi-index column structure in yfinance v0.2+.
• CLI Demo Scripts: Added 8 run-ready scripts in examples/examples-python/ covering every module.
• Documentation: Significantly expanded docs/THEORY.md explaining the mathematics behind every model.

Installation

pip install finlearner==0.1.0

Quick Start

Run the complete demo to see everything in action:

# Clone the repo first
python examples/examples-python/08_complete_demo.py

Contributors

Special thanks to everyone who contributed to testing and feedback for this release.

Finlearner v0.0.92 Initial Release

Release Notes

v0.0.92 (2026-01-29)

🎉 Highlights

This release brings comprehensive test coverage, a professional README redesign, and modernized CI/CD infrastructure.

---

✨ New Features

• Comprehensive Unit Test Suite — Added 39 tests covering all modules:

  • test_data.py — DataLoader tests with mocked yfinance
  • test_models.py — TimeSeriesPredictor tests with mocked Keras
  • test_portfolio.py — PortfolioOptimizer tests
  • test_plotting.py — Plotter visualization tests
  • test_utils.py — Utility function tests
  • conftest.py — Shared pytest fixtures

• Modern pyproject.toml — Replaced legacy setup with PEP 517/518 compliant build configuration

📖 Documentation

• Professional README — Complete redesign inspired by Hugging Face Transformers:
  • Centered logo and badges
  • Interactive quick start examples
  • Mermaid architecture diagram
  • Module reference table
  • Roadmap section
  • BibTeX citation for academic use

🔧 Infrastructure

• Separated CI/CD Workflows:

  • tests.yml — Runs on every push/PR with Python 3.9-3.12 matrix
  • publish.yml — Runs tests → builds → publishes to PyPI on version tags

• Improved .gitignore — Comprehensive patterns for Python projects

📦 Dependencies

No changes to runtime dependencies.

---

📋 Full Changelog

Tests Added:

• DataLoader: 5 tests
• TimeSeriesPredictor: 8 tests
• PortfolioOptimizer: 5 tests
• Plotter: 6 tests
• check_val: 8 tests
• BlackScholesMerton: 3 tests (existing)
• BlackScholesPINN: 2 tests (existing)
• TechnicalIndicators: 2 tests (existing)

Files Added:

• pyproject.toml
• tests/conftest.py
• tests/test_data.py
• tests/test_models.py
• tests/test_portfolio.py
• tests/test_plotting.py
• tests/test_utils.py
• .github/workflows/tests.yml
• .github/workflows/publish.yml

Files Removed:

• .github/workflows/ci_cd.yml (replaced by tests.yml + publish.yml)

Files Modified:

• README.md (complete redesign)
• .gitignore (expanded patterns)

---

🙏 Contributors

• Ankit Dutta (@ankitdutta428)