bima 0.6.6

N-body simulation powered with Rust backend

N-body simulation library in Rust with Python bindings.

Features

• [-] Integrators:
  • Euler
  • Runge-Kutta 4
  • Bulirsch-Stoer
  • [x]Leapfrog (Symplectic)
  • Wisdom-Holman
• Adaptive time stepping.
• Collision detection and handling.
  • Solid body collisions.
  • Merging bodies.
• [-] Close encounter detection.
  • Softening.
  • Truncated.
  • Regularization.
• Force calculation.
  • Direct summation.
  • Barnes-Hut tree (octree).
  • Fast multipole method.
• Parallelization.
  • CPU multi-threading.
  • GPU acceleration.

Installation

pip install bima

Quick Start

from bima.body import Body
from bima import Config
import numpy as np
import bima
from matplotlib import pyplot as plt

v = np.sqrt(2)
arr = np.array([[1, -1, 0, 0, 0, -2/3 * v, 0], [2, 0.5, 0, 0, 0, 1/3 * v, 0]])
initial = bima.Initial.from_arr(arr)

config = Config(
    force=bima.ForceMethod.Direct,
    # integrator=bima.Integrator.Euler,
    integrator=bima.Integrator.LeapFrog,
    timestep=bima.TimestepMethod.Constant(0.1),
    # timestep=bima.TimestepMethod.Constant(1),
    close_encounter=bima.CloseEncounterMethod.Regularized,
)

sim = bima.Simulation(initial)

bodies = sim.in_memory.run(config, 100)
# bodies = sim.in_memory.run(config, 1)

# print(len(energy.t))


def plot(bodies: list[Body]):
    body0 = bodies[0]
    length = len(body0)
    sample_n = np.min([1000, length])
    skip = length//sample_n
    x0 = body0.x[::skip]
    y0 = body0.y[::skip]

    body1 = bodies[1]
    length = len(body1)
    skip = length//sample_n
    x1 = body1.x[::skip]
    y1 = body1.y[::skip]

    fig, ax = plt.subplots()
    ax.plot(x0, y0)
    ax.scatter(x0[-1], y0[-1])
    ax.plot(x1, y1)
    ax.scatter(x1[-1], y1[-1])
    ax.set_aspect("equal")
    plt.show()


# plot(bodies)

energy = bima.Energy.from_bodies(bodies)
e0 = energy.e[0]
plt.plot(energy.t, (e0-energy.e)/e0)
plt.show()

License

GNU General Public License v3.0. See LICENSE for more details.

Documentation

under construction