Skip to content

XPP Lab v0.1.2

Latest
Latest

Choose a tag to compare

View all tags
@github-actions github-actions released this 26 Oct 19:49
· 2 commits to main since this release
v0.1.2
8f8ec22
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

XPP Lab 0.1.2: XPPAUT 8.0's engine, giving the same numbers as XPPAUT, with a desktop app for Windows, macOS and Linux and a Python package. It needs no X server, compiler or Terminal.

New in 0.1.2

  • The app and Python together. Each action in the app now appears in the command window as the line of Python (the xpplab package) that does the same, and File > Save session as Python script saves them: the script repeats the session exactly. The command window also takes Python: click its >> prompt, and m.run() runs the open model in the app while its plots follow. On a computer without Python, Install Python downloads one for XPP Lab, with NumPy and Matplotlib (about 55 MB, once).
  • Drive the open app from outside. xpplab.app.connect() in a notebook or a script, or a coding agent through the MCP tools app_status, app_command and app_run, acts in the window you have open, and you watch the plots change.
  • The rest of XPPAUT's interface: equilibria over a parameter and every equilibrium in a box (Sing pts > Range and Monte Carlo); Range over two parameters; initial values from a file, a formula, Shift and several clicks; frozen and saved nullclines; the Kinescope, with animated GIFs and a film of the Range; statistics over many noisy runs (stocHastic > Compute). xpplab-batch is XPPAUT's -silent mode, and also runs the app's commands, AUTO included, without a window.
  • AUTO: diagrams save and load exactly (XPPAUT keeps 6 digits), and an orbit can be imported. Boundary-value and homoclinic starts are now tested, and a homoclinic start answers XPPAUT's info box. A periodic branch that ends on a homoclinic orbit no longer shows spurious limit and branch points on its tail.
  • Linearise (Run > Linearise): the A and B matrices at an equilibrium for the inputs you choose, the eigenvalues in the complex plane, and whether the inputs can steer the system. The app and Model.linearize give the same numbers.
  • Plots keep your zoom until a double-click or Fit.
  • Learn by example on the website: field tutorials in neuroscience, ecology, chemistry and mechanics. In Jupyter, models, runs, equilibria and AUTO diagrams show as tables.

Download and run

Your computer Download The first time you open it
Windows 10 or 11, 64-bit XPPLab-Windows-x64.zip Unzip it anywhere and run xpplab.exe. If Windows says "Windows protected your PC", click More info, then Run anyway.
macOS 11 or later, Apple silicon or Intel XPPLab-macOS.dmg Open it and drag the app to Applications. On macOS 15 and later, open the app once, then click Open Anyway in System Settings > Privacy & Security. On macOS 14 and earlier, Control-click the app and choose Open.
Linux, x86-64 (built on Ubuntu 22.04) XPPLab-Linux-x86_64.tar.gz tar xzf XPPLab-Linux-x86_64.tar.gz, then run XPPLab-Linux-x86_64/xpplab.

The warnings appear because the apps are not yet signed with a paid certificate. The checksums at the end let you check a download.

Try it: the Morris-Lecar model opens. Press Run, open the Phase Plane tab, tick Nullclines, then drag the iapp slider. The File menu opens the other 150 example models that come with XPPAUT, and F1 opens the manual and its tutorial.

Python

The same engine, with AUTO, from Python 3.9 or newer on any of the three systems:

pip install xpplab==0.1.2

Then see the Python guide.

What's inside

  • XPPAUT 8.0's own engine. Every example model gives the same output as unmodified XPPAUT, digit for digit. The only changes fix genuine bugs, listed in upstream-issues.
  • Phase plane:
    • nullclines and the direction field;
    • equilibria by stability, and the manifolds of saddles;
    • click-to-start trajectories.
  • Bifurcations: AUTO without the dialogs. A table of special points, and for each one the continuations that make sense.
  • Workspace: a slider for every parameter, with live re-simulation, plus parameter sweeps, Range, and boundary-value problems.
  • 3D view and the Animator: XPPAUT's .ani files, played over the run.
  • Analysis:
    • phase response curves and the H function;
    • spectra, histograms, correlations and the Liapunov exponent;
    • fitting to data.
  • Editor, command window and figure export. The editor's errors point at the line, and the command window takes XPPAUT's two-key commands, and Python. Figures export as PNG and SVG, data as CSV.
  • Everything as Python. Each action in the app is also a line of Python, and a session saves as a script that repeats it.

Found a problem, or a model that works in XPPAUT but not here? Please open an issue.

XPP Lab is free software under the GPL. It is an independent rebuild of G. Bard Ermentrout's XPPAUT, and is not endorsed by him; see THIRD-PARTY-NOTICES.md for the work of others that it contains.

Checksums (SHA-256)

cbe3e98632cca7caf47699d15b4654ef3ede87bdbb3dacd63720db75139d8771  XPPLab-Linux-x86_64.tar.gz
5ecb9e41e0a3c7bbfa220bfc4f234f14d67d68cafe650063f6b2b109a963662c  XPPLab-Windows-x64.zip
879005302d6603b92e074e716723712a079749f48539c3606fa8ee0fe1177314  XPPLab-macOS.dmg
52df4c0a828d98d9fa746fbba84f9b3996e2068040726b0e8d18b4e84b4b21b1  xpplab-0.1.2-py3-none-macosx_10_13_universal2.whl
e4a4631714563650dbd541a4beb3a7f4b9c3f7c2129cb1c2c2afc8ad527945c2  xpplab-0.1.2-py3-none-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl
99f44379d309851912df302fd48d88ce6b78ba230d052d6fed3f8834128717eb  xpplab-0.1.2-py3-none-win_amd64.whl

What's Changed

Full Changelog: v0.1.1...v0.1.2

Contributors

CesarPerdomoDiaz and limit-cyclist
Assets 9
Loading