The high-performance mathematical foundation for space situational awareness.

ASTRA-Core is a rigorous Python astrodynamics engine for aerospace engineers, researchers, and developers. It propagates large catalogs, screens conjunctions, estimates collision probability, and predicts ground passes—using both legacy TLE and modern OMM data end-to-end.

📖 API reference: Read the Docs

🧠 How the math works: KNOWMORE.md — TLEs, OMM, SGP4, spatial screening, P_c, Cowell forces, and what the models assume.

ASTRA-Core supports both the legacy TLE format and the CCSDS OMM (Orbit Mean-Elements Message). Physics APIs accept either via the unified SatelliteState type.

Tip: Prefer OMM when you care about drag, realistic cross-sections, or conjunction risk—the extra metadata flows into screening and Cowell without extra glue code.

Spacebook provides synthetic covariance and standard/XP-TLEs along with highly precise Space Weather metrics. Spacebook overrides CelesTrak SW defaults when active.

import astra

# Load XP-TLE catalog formatted transparently as SatelliteOMM with precision tags
xp_catalog = astra.fetch_xp_tle_catalog()

import astra

tles = astra.fetch_celestrak_group("starlink")
omms = astra.fetch_celestrak_group_omm("starlink")

Requests identify the client as ASTRA-Core/<version>.

Never hardcode passwords. Set once per machine:

# Windows (Command Prompt — restart shell after setx)
setx SPACETRACK_USER [email protected]
setx SPACETRACK_PASS yourpassword

# Linux / macOS (~/.bashrc or ~/.zshrc)
export [email protected]
export SPACETRACK_PASS=yourpassword

import astra

starlinks = astra.fetch_spacetrack_group("starlink")
catalog = astra.fetch_spacetrack_active()

If credentials are missing, ASTRA raises a clear error with setup hints.

• Spacebook Integration: Direct streaming of Spacebook XP-TLEs, true observational covariance matrices, and live Space Weather priorities—bypassing heuristic estimation models for flight-grade accuracy.
• Dual format (TLE + OMM): One API surface for parsing, propagation, filtering, and conjunctions.
• SGP4 at scale: Vectorized propagation (propagate_many, generators) with UT1-aware handling where ephemeris data are available.
• Cowell propagation: Dormand–Prince integration with J₂–J₄, empirical drag (space weather), Sun/Moon third-body gravity (JPL DE421), high-fidelity solar radiation pressure with conical Earth shadow (continuous penumbra modeling), and 7-DOF finite burns with mass flow.
• Conjunction screening: KD-tree prefilter over time steps (~14.8x speedup), spline refinement for TCA, Spacebook EOP coordinate mapping, and dynamic effective radius from metadata when available.
• Collision probability: Analytical (Chan/Foster lineage), exact 2D Gaussian Quadrature (dblquad), and 6D Monte Carlo paths when full covariances are supplied; Seamless integration with Spacebook synthetic covariance matrices.
• Catalog ingestion: CelesTrak and Space-Track helpers plus local OMM files.
• Pass prediction: TEME → ground observer pipeline (ENU), coarse grid + refinement for AOS/TCA/LOS.
• Optional 3D plots: Interactive Plotly figures via the [viz] extra—core install stays lean for servers and CI.

Default (core physics, no Plotly):

pip install astra-core-engine

With 3D trajectory plotting:

pip install "astra-core-engine[viz]"

From source (development + tests):

git clone https://github.com/ISHANTARE/ASTRA.git
cd ASTRA
pip install -e ".[test]"

Requires Python 3.10+. Core dependencies include NumPy, SciPy, Skyfield, SGP4, Requests, Numba, and defusedxml.

ASTRA-Core implements widely used models suitable for research, education, integration prototypes, and operations-style workflows when you understand the assumptions. It is not a certified conjunction or mission-closure product by itself—validate against your own requirements and reference tools if needed.

Strict mode: astra.set_strict_mode(True) or astra.config.ASTRA_STRICT_MODE = True makes many missing-data paths raise instead of warn-and-continue—recommended when building tools that must not guess.

More detail: KNOWMORE.md and the Limitations page on Read the Docs.

import astra
import numpy as np

active_catalog = astra.fetch_celestrak_active()
objects = [astra.make_debris_object(tle) for tle in active_catalog]
leo_only = astra.filter_altitude(objects, min_km=200, max_km=2000)

sources = [obj.source for obj in leo_only]
times_jd = leo_only[0].source.epoch_jd + np.arange(0, 120, 5.0) / 1440.0
trajectories = astra.propagate_many(sources, times_jd)

events = astra.find_conjunctions(
    trajectories,
    times_jd=times_jd,
    elements_map={obj.source.norad_id: obj for obj in leo_only},
    threshold_km=5.0,
)

import astra
import numpy as np

omm_catalog = astra.fetch_celestrak_active_omm()
# Or: omm_catalog = astra.load_omm_file("catalog.json")

objects = [astra.make_debris_object(omm) for omm in omm_catalog]
leo_only = astra.filter_altitude(objects, min_km=200, max_km=2000)

sources = [obj.source for obj in leo_only]
times_jd = leo_only[0].source.epoch_jd + np.arange(0, 120, 5.0) / 1440.0
trajectories = astra.propagate_many(sources, times_jd)

events = astra.find_conjunctions(
    trajectories,
    times_jd=times_jd,
    elements_map={obj.source.norad_id: obj for obj in leo_only},
    threshold_km=5.0,
)
print(f"Found {len(events)} conjunction events.")

import astra

catalog = astra.fetch_spacetrack_active()
print(f"Loaded {len(catalog)} satellites.")

from astra import plot_trajectories

fig = plot_trajectories({"25544": positions_array})

Functions are available from the astra namespace.

• parse_omm_json(text) → list[SatelliteOMM]
• parse_omm_record(dict) → SatelliteOMM
• load_omm_file(path) → list[SatelliteOMM]
• validate_omm(dict) → bool

• load_tle_catalog(lines) → list[SatelliteTLE]
• parse_tle(name, l1, l2) → SatelliteTLE
• validate_tle(name, l1, l2) → bool

• make_debris_object(source) — SatelliteTLE or SatelliteOMM
• filter_altitude, filter_region, filter_time_window, apply_filters, catalog_statistics

• propagate_orbit, propagate_many, propagate_many_generator, propagate_trajectory, ground_track
• propagate_cowell — numerical Cowell + DragConfig

• find_conjunctions, closest_approach, distance_3d
• compute_collision_probability, compute_collision_probability_mc
• estimate_covariance, propagate_covariance_stm, rotate_covariance_rtn_to_eci
• parse_cdm_xml

• get_space_weather, load_space_weather, atmospheric_density_empirical
• sun_position_de, moon_position_de, etc.

• passes_over_location, visible_from_location

• convert_time, vincenty_distance, orbit_period, orbital_elements
• set_strict_mode, astra.config.ASTRA_STRICT_MODE

Release notes: CHANGELOG.md.

@software{Tare_ASTRA_2026,
  author = {Tare, Ishan},
  title = {ASTRA: Autonomous Space Traffic Risk Analyzer},
  year = {2026},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/ISHANTARE/ASTRA}},
  version = {3.4.0}
}

Ishan Tare

© 2026 ASTRA Project · MIT License