JyotiPy

Pure-Python Vedic (Jyotish) astrology library. Computes sidereal planetary positions, ascendant, houses, divisional charts, dashas, panchanga, transits, Ashtakavarga, Shadbala, and a conservative set of yogas with zero C extensions or external ephemeris files.

Built as a permissive-license alternative to pyswisseph. All computations use PyMeeus for planetary positions (VSOP87 + ELP2000-82) followed by pure-Python ayanamsa, house, dasha, and strength-system logic. Supports Python 3.8+.

Installation

pip install jyotipy

Or install from source:

git clone https://github.com/AnandShah10/jyotipy.git
cd jyotipy
pip install -e .

Quick Start

from datetime import datetime
from jyotipy import BirthChart, Graha

chart = BirthChart(
    dt=datetime(1990, 6, 15, 14, 32),
    utc_offset_hours=5.5,
    latitude=23.03, longitude=72.58,   # Ahmedabad
)

print(chart.summary())          # sign, degree, nakshatra, pada per graha
print(chart.ascendant)          # sidereal Lagna longitude
print(chart.houses("whole_sign"))
print(chart.varga("D9"))        # Navamsha
print(chart.mahadashas()[:3])   # first 3 Vimshottari mahadashas
print(chart.yogas())
print(chart.panchanga())

New in 0.2

chart.reduced_ashtakavarga(Graha.SUN)                # Bhinnashtakavarga after Trikona + Ekadhipatya Sodhana
chart.transits(datetime(2026, 9, 1, 12, 0), 5.5)     # Gochara + Vedha-aware transit report
chart.sade_sati(datetime(2026, 9, 1, 12, 0), 5.5)    # Sade Sati / Kantaka Sani / Ashtama Sani status
chart.dasha_bhukti_activation(datetime(2026, 9, 1))  # which period lord dominates, via Shadbala
chart.full_shadbala()                                # all six Shadbala components, all 7 planets, Yuddha Bala applied
chart.houses("placidus")                             # or "sripati"

See CHANGELOG.md for the full list of what shipped in 0.2.

What's implemented (v0.2)

Accuracy notes -- read this before using it for real work

• Planetary positions: PyMeeus implements full VSOP87 (planets) and ELP2000-82 (Moon), validated in this project against Meeus's own published worked examples. Cross-check a real chart against JHora or astro.com before relying on this for client work.
• Ayanamsa: Lahiri/KP-Newcomb/True Chitrapaksha are cross-checked against multiple independent public sources and numerically validated against published reference values. Raman is now independently sourced from its own base epoch (397 CE) and precession rate, confirmed by 7 independent sources including B.V. Raman's grandson.
• Divisional charts: all 16 vargas (D1-D60) are implemented and cross-sourced. D60 specifically has a documented, unresolved disagreement between two classical methods in the wider literature; this implementation uses the one modern software (JHora) actually uses -- see varga.py for the alternate reading.
• Houses: Whole Sign and Equal House are exact by construction. Porphyry, Placidus, and Sripati are all implemented, but Placidus in particular required unusually extensive debugging -- see houses.py's module comments for the full story if you want to understand exactly what's verified (boundary conditions to numerical precision, universal house-numbering convention against real Swiss Ephemeris output) versus what has no external numeric example to check against (the final 12-cusp output end-to-end). Sripati has one known quirk: for some ordinary, non-polar charts, its House 1 cusp doesn't sit exactly at the true Ascendant, because the real reference algorithm (transcribed directly from Swiss Ephemeris source) includes a swap condition that triggers more broadly than its own documentation suggests.
• Ashtakavarga: the base bindu tables and both reduction techniques (Trikona/Ekadhipatya Sodhana) are implemented. The base tables are validated against a fully worked B.V. Raman example (all 7 planets plus Sarvashtakavarga matched exactly). The two reductions don't have an external worked example to check against; confidence rests on exhaustive coverage of every documented rule branch in the test suite instead.
• Shadbala: all six classical components (Sthana, Dig, Kala, Chesta, Naisargika, Drik) are now implemented. Two narrow open items remain: Yuddha Bala's victor rule doesn't implement a secondary "Venus always wins" exception some sources describe, and Drik Bala's formula was reconstructed from a worked reference grid (not a primary text) -- both documented in their respective modules (yuddhabala.py, drikbala.py).
• Transit/Gochara: the base good/bad-house and Vedha tables are cross-validated against multiple sources and a full worked example. Dasha/Bhukti activation weighting layers Shadbala-based strength judgment on top, per B.V. Raman's own stated method, but doesn't attempt his complete predictive synthesis (house lordships, functional benefic/malefic status, classical yogas).
• Vimshottari year length: uses 365.25 days/year, a common but not universal convention. Check what a reference tool uses if you need bit-for-bit agreement with it.

Roadmap

11. Nadi astrology, Prashna (horary), Muhurta (electional), Varshaphal/Tajika (annual charts) -- none started yet
12. A CLI for the library
13. Full documentation coverage as new modules are added (currently up to date as of 0.2.1)

License

MIT License (see the LICENSE file in the GitHub repository).

JyotiPy depends on PyMeeus (LGPLv3) as an unmodified runtime dependency. This does not change JyotiPy's own MIT license, but if you vendor or fork PyMeeus itself, the LGPL terms apply to that portion.

Full documentation: https://jyotipy.readthedocs.io

Source code, examples, and issue reports are at the GitHub repository.