InterpolatePy

InterpolatePy is a trajectory-planning and interpolation library for robotics, animation, and scientific computing. It includes scalar splines, B-spline curve tools, bounded motion profiles, quaternion interpolation, and 3D path utilities.

The package has a NumPy/SciPy implementation and an optional C++20 backend. Use the public names exported by interpolatepy; they select the available backend at import time.

Installation

python -m pip install InterpolatePy

InterpolatePy 3.2.1 requires Python 3.11 or newer, NumPy 1.26 or newer, SciPy 1.11 or newer, and Matplotlib 3.6 or newer.

Quick start

import numpy as np

from interpolatepy import CubicSpline
from interpolatepy import DoubleSTrajectory
from interpolatepy import StateParams
from interpolatepy import TrajectoryBounds

# A clamped cubic spline: v0 and vn are endpoint velocities.
spline = CubicSpline(
    [0.0, 1.0, 2.0, 3.0],
    [0.0, 1.5, -0.5, 2.0],
    v0=0.0,
    vn=0.0,
)
times = np.linspace(0.0, 3.0, 100)
positions = spline.evaluate(times)
velocities = spline.evaluate_velocity(times)

# A jerk-limited Double-S motion profile.
state = StateParams(q_0=0.0, q_1=10.0, v_0=0.0, v_1=0.0)
bounds = TrajectoryBounds(v_bound=5.0, a_bound=10.0, j_bound=30.0)
motion = DoubleSTrajectory(state, bounds)
sample_times = np.linspace(0.0, motion.get_duration(), 100)
q, qd, qdd, qddd = motion.evaluate_full(sample_times)

DoubleSTrajectory.evaluate() returns position only. Use evaluate_velocity(), evaluate_acceleration(), and evaluate_jerk() for one component, or evaluate_full() for all four.

What is included

Area	Public APIs	Typical use
Scalar splines	CubicSpline, CubicSmoothingSpline, CubicSplineWithAcceleration1, CubicSplineWithAcceleration2	Smooth scalar waypoint trajectories and noisy data
B-splines	BSpline, BSplineInterpolator, CubicBSplineInterpolation, ApproximationBSpline, SmoothingCubicBSpline	Parametric curves, interpolation, approximation, and smoothing
Motion profiles	DoubleSTrajectory, TrapezoidalTrajectory, PolynomialTrajectory, ParabolicBlendTrajectory	Bounded or boundary-conditioned scalar motion
Rotations	Quaternion, QuaternionSpline, SquadC2, LogQuaternionInterpolation, ModifiedLogQuaternionInterpolation	Orientation interpolation without Euler-angle singularities
Paths and utilities	LinearPath, CircularPath, Frenet-frame helpers, linear_traj, solve_tridiagonal	Geometric paths, moving frames, and numerical helpers

Degrees 3, 4, and 5 of BSplineInterpolator, LogQuaternionInterpolation, and ModifiedLogQuaternionInterpolation support as few as two waypoints in version 3.2.0.

See the algorithm selection guide, the full documentation, and the runnable examples.

Optional C++ backend

The package falls back to Python automatically when the extension is absent:

import interpolatepy

print(interpolatepy.HAS_CPP)

Set INTERPOLATEPY_NO_CPP=1 before importing the package to force the Python implementation. The standard Python package does not need a compiler. Building the native library or Python extension from a source checkout requires CMake 3.21+, a C++20 compiler, and network access for CMake's fetched dependencies; see the installation guide.

Development

This project uses uv:

git clone https://github.com/GiorgioMedico/InterpolatePy.git
cd InterpolatePy
uv sync
uv run pytest
uv run pre-commit run --all-files

Documentation dependencies are in a separate group:

uv sync --group docs
uv run mkdocs serve

Every Python program in examples/ can also be run directly, for example:

uv run python examples/double_s_ex.py

For the complete workflow, see Contributing.

License and citation

InterpolatePy is distributed under the MIT License.

@misc{InterpolatePy,
  author = {Giorgio Medico},
  title  = {InterpolatePy: Trajectory Planning and Interpolation for Python},
  year   = {2026},
  url    = {https://github.com/GiorgioMedico/InterpolatePy}
}