Start · 01

Install¶

TSDynamics runs on Python ≥ 3.12. Installing it is a single command, and the compiled integration engine ships inside the wheel — you need no Rust toolchain, no C compiler, and no build step.

pip install tsdynamics

or, with uv:

uv add tsdynamics

What "no compiler needed" means¶

pip install tsdynamics pulls a prebuilt abi3 wheel. There is one wheel per platform and architecture — manylinux, musllinux, macOS (Intel and Apple silicon), and Windows — and each bundles the native Rust engine (tsdynamics._rust) alongside the pure-Python package. Because the wheel is built against CPython's stable ABI (abi3, tagged cp312), a single wheel covers every CPython ≥ 3.12.

Concretely, this means:

No toolchain. Installing and running the library never invokes a compiler.
No build step, nothing on disk. Every family — ODEs, delay equations, stochastic equations, and maps — lowers its symbolic equations to an in-process engine tape the first time you run it, in a fraction of a second, and the tape is compiled to native code by the built-in JIT. Both steps are in-process and memoised per system; there is no ahead-of-time compilation and nothing is written to disk. Editing a system's equations simply takes effect on the next run; there is no cache to wipe.
The same numbers everywhere. The tape runs on a Cranelift JIT by default, or on a Rust interpreter that is bit-for-bit identical — see backends — so results do not depend on which BLAS or Python build you happen to have.

Building from source

Only building from the source distribution — the sdist, the fallback for a platform outside the wheel matrix — needs a Rust toolchain, because the package build backend is maturin. The normal pip install path never reaches it.

Optional extras¶

The base install pulls in no plotting library — import tsdynamics stays lightweight and imports nothing heavy. The plotting backends are opt-in extras:

Extra	Installs	When you want it
`tsdynamics[viz]`	`matplotlib`	The reference renderer: static 2-D and 3-D figures, movies (mp4/gif)
`tsdynamics[interactive]`	`plotly`	The interactive backend: rotatable 3-D and self-contained HTML export
`tsdynamics[plot]`	`matplotlib`	A back-compatible alias of `viz`

Combine extras in the usual way:

pip install "tsdynamics[viz,interactive]"

The json and threejs export renderers need no extra at all — they serialize a plot description over the standard library. See Visualization for what each backend can draw.

Verify the install¶

import tsdynamics as ts

print(ts.__version__)
print(ts.registry.families())   # {'ode': 142, 'dde': 6, 'sde': 3, 'map': 26}

traj = ts.systems.Henon().run(steps=100, ic=[0.1, 0.1])
print(traj.y.shape)             # (100, 2)

ts.registry.families() reports the built-in catalogue by family. If

ts.systems.Lorenz().run(final_time=1.0).y.shape   # (51, 3)

also returns cleanly, the engine is wired up and you are ready to go.

Next¶

02 · First trajectory — instantiate the Lorenz system, run it, read its components, and measure its Lyapunov spectrum.