pytermstructure 0.1.0

Educational Python library for interest rate term structure estimation

PyTermStructure

Educational Python library for interest rate term structure estimation

Inspired by Damir Filipović's "Interest Rate Models"
École Polytechnique Fédérale de Lausanne (EPFL)
Course Link

---

Version 0.1.0

NEW in v0.1.0: Bootstrap accuracy improvements

• Exact calendar date support with ACT/360 day-count
• Automatic curve densification with interpolated swap rates
• Three-phase bootstrap algorithm
• Accuracy: <1 bps on benchmark data (improved from ~13 bps)

Status: Educational implementation suitable for learning and research.

Recommendation: Use for educational purposes, research, and prototyping. For production systems requiring high precision, consider QuantLib or FinancePy.

---

Quick Start

Installation

pip install pytermstructure

Your First Term Structure (NEW: with exact dates!)

import pytermstructure as pts
from pytermstructure.core import MarketInstrument, InstrumentType
from datetime import datetime

# Create bootstrap with spot date
spot = datetime(2024, 1, 15)
bootstrap = pts.BootstrapMethod(verbose=True, spot_date=spot)

# Add LIBOR with exact date
bootstrap.add_instrument(MarketInstrument(
    instrument_type=InstrumentType.LIBOR,
    maturity=datetime(2024, 4, 15),  # 3 months
    quote=0.15,                       # 0.15%
    spot_date=spot
))

# Add swap with exact date
bootstrap.add_instrument(MarketInstrument(
    instrument_type=InstrumentType.SWAP,
    maturity=datetime(2026, 1, 15),  # 2 years
    quote=0.50,                       # 0.50%
    spot_date=spot
))

# Fit discount curve (automatic densification!)
discount_curve = bootstrap.fit()

# Get zero rates
zero_rates = bootstrap.get_zero_rates()

Getting Help

import pytermstructure as pts

# General help
pts.help()

# Method-specific help
pts.help("bootstrap")
pts.help("lorimier")
pts.help("accuracy")  # NEW!

---

Features

Methods Implemented

Method	Type	Accuracy	Use Case
Bootstrap	Exact	<1 bps	Educational, curve construction
Pseudoinverse	Exact	<1 bps	Smooth + exact pricing
Lorimier	Smooth	±3 bps	Smooth forward curves
PCA	Analysis	N/A	Risk management
Nelson-Siegel	Parametric	N/A	Quick approximation

Educational Quality

• Built-in help system with comprehensive documentation
• Based on academic course materials (Filipović, EPFL)
• Full typing support with type hints
• Validated with academic benchmark data
• Free Software (GNU GPLv3 license)
• Professional structure ready for contributions

---

Methods Overview

1. Bootstrap Method

Sequential construction from LIBOR → Futures → Swaps.

NEW in v0.1.0: Sub-basis-point accuracy with exact dates!

from datetime import datetime

spot = datetime(2024, 10, 3)
bootstrap = pts.BootstrapMethod(verbose=True, spot_date=spot)

bootstrap.add_instrument(pts.MarketInstrument(
    pts.InstrumentType.LIBOR, 
    datetime(2025, 1, 3), 
    0.15,
    spot_date=spot
))

bootstrap.add_instrument(pts.MarketInstrument(
    pts.InstrumentType.SWAP, 
    datetime(2026, 10, 3), 
    0.50,
    spot_date=spot
))

discount_curve = bootstrap.fit()

Accuracy: <1 bps on benchmark data (v0.1.0)

---

2. Lorimier Method

Smoothing splines with parameter α for smooth forward curves.

import numpy as np

# Swiss government bond yields
maturities = np.array([2, 3, 4, 5, 7, 10, 20, 30])
yields = np.array([-0.79, -0.73, -0.65, -0.55, 
                   -0.33, -0.04, 0.54, 0.73]) / 100

lorimier = pts.LorimierMethod(alpha=0.1)
discount_curve = lorimier.fit(yields, maturities)

# Interpolate at 6 years
y_6y = lorimier.get_yield_at(6.0)

Accuracy: ±3 bps on benchmark data

---

3. PCA Analysis

Principal component analysis of yield curve movements.

pca = pts.PCAAnalysis()
eigenvalues, eigenvectors, explained_var = pca.fit(yield_changes)

print(f"Level:     {explained_var[0]:.1f}%")
print(f"Slope:     {explained_var[1]:.1f}%")
print(f"Curvature: {explained_var[2]:.1f}%")

---

Installation

From PyPI

pip install pytermstructure

From Source

git clone https://github.com/MarcoGigante/pytermstructure.git
cd pytermstructure
pip install -e .

Test Installation

python -c "import pytermstructure as pts; pts.version(); pts.help()"

Expected output:

PyTermStructure 0.1.0 loaded. For help: pts.help()
PyTermStructure 0.1.0
Author: Marco Gigante
License: GPLv3

---

Running Examples

# Example 1: Bootstrap US market
python examples/example_bootstrap.py

# Example 2: Pseudoinverse
python examples/example_pseudoinverse.py

# Example 3: Lorimier Swiss bonds
python examples/example_lorimier.py

# Run all examples
python examples/practical_examples.py

---

What's New in v0.1.0

Bootstrap Enhancements

1. Exact Date Support
   Use real calendar dates instead of year fractions

2. Automatic Densification
   Automatically adds intermediate points with interpolated swap rates

3. Three-Phase Algorithm

   • Phase 1: Bootstrap market instruments (LIBOR, Futures, Swaps ≤20Y)
   • Phase 2: Add intermediate dates (21-29Y) with interpolated rates
   • Phase 3: Calculate long swaps (30Y+) using densified curve

4. ACT/360 Day-Count
   Exact day-count calculations for all periods

Technical Improvements

• Added python-dateutil dependency for date handling
• Enhanced MarketInstrument with maturity_date support
• Implemented _get_exact_swap_schedule() for ACT/360
• Removed unused dependencies (pandas, openpyxl)

---

Academic Reference

This library implements methods from:

Filipović, D. (2009). Term-Structure Models: A Graduate Course. Springer Finance.

Online Course: Interest Rate Models
École Polytechnique Fédérale de Lausanne (EPFL)

---

License

GNU General Public License v3.0 or later.

This ensures PyTermStructure remains free software forever.

See LICENSE for details.

---

Contributing

Contributions are welcome. Areas for improvement:

1. Enhanced numerical methods
2. Additional day-count conventions
3. Business day calendars
4. Curve analytics
5. Additional validation tests

Please ensure GPL-compatible contributions.

See CONTRIBUTING.md for guidelines.

---

Author

Marco Gigante
MSc in Quantitative Finance, University of Siena

Inspired by Prof. Damir Filipović's course at EPFL.

---

Acknowledgments

• Prof. Damir Filipović (EPFL)
• École Polytechnique Fédérale de Lausanne
• NumPy, SciPy communities
• Free Software Foundation

---

Support

• Built-in Help: pts.help()
• Issues: GitHub Issues
• Documentation: pytermstructure.readthedocs.io

---

Changelog

v0.1.0 (November 17, 2025)

• Bootstrap accuracy improvements (<1 bps on benchmark data)
• Exact calendar date support with ACT/360
• Automatic curve densification
• Three-phase bootstrap algorithm
• Enhanced documentation and help system

v0.0.1 (Initial Release)

• Bootstrap, Lorimier, PCA, Nelson-Siegel methods
• Built-in help system
• Educational implementation

See CHANGELOG.md for complete history.)