Python quick start

Install

pip install xpplab            # the engine (NumPy is the only dependency)
pip install "xpplab[all]"     # plus matplotlib and pandas

The wheel contains the compiled XPPAUT engine. Nothing needs a compiler, and no X11 server is involved.

Start from the desktop app

The desktop app writes each action as the xpplab code that does the same, in its command window, and File > Save session as Python script turns a session into a script that repeats it number for number. Explore with the mouse, then carry on in Python from the saved script.

Load a model and run it

Any XPPAUT .ode file works. You can also write the model inline:

import xpplab as xs

m = xs.Model.from_text("""
v' = v - v^3/3 - w + I
w' = eps*(v + a - b*w)
par I=0.5, a=0.7, b=0.8, eps=0.08
init v=-1, w=1
@ total=100, dt=0.05
done
""")

sol = m.run()               # integrate with the file's settings
sol.t                       # times (NumPy array)
sol["V"]                    # a variable (names ignore case)
sol.to_pandas()             # a DataFrame indexed by time
sol.plot("V", "W")          # a quick matplotlib plot

Change things

Parameters, initial values and numerical options behave like dictionaries:

m.parameters["I"] = 0.2
m.initial_conditions["v"] = 0.5
m.set(total=300, meth="stiff")    # anything an ODE-file @ line can set
m.run(I=0.4)                      # set, then run
m.apply_set("fast")               # a named 'set' from the file

Explore

eq = m.equilibrium()       # Newton's method from the initial values
eq.kind, eq.eigenvalues    # 'unstable focus', array([...])

m.phase_plane("v", "w")    # trajectory, nullclines, flow, equilibrium

Bifurcations

m.parameters["I"] = 0
d = m.auto.equilibria("I", parmax=2)   # follow the rest state in I
d.special()                            # [<HB #2: ... I=0.331>, <HB #3: ...>]
orbits = m.auto.continue_from(d.first("HB"), "periodic")
m.auto.diagram.plot("V")

Several models at once

Each Model holds its own copy of the engine, so models are independent:

a = xs.load("lecar.ode")
b = xs.load("lecar.ode")
a.parameters["iapp"] = 0.3     # b is unaffected

Use with xs.load(...) as m: or m.close() to free a model's memory promptly.