FACETS CodeJam Workshop #2

5th-8th May 2008, Château du CNRS, Gif-sur-Yvette, France

Château CNRS, Gif sur Yvette


The first CodeJam focused on adding or improving Python support for different neuroscience simulators, and on the PyNN common simulator API. The focus for the second CodeJam was expanded to include the next stages in the simulation workflow - analysis and visualisation of simulation results, and management of simulation projects to promote reproducibility and reliability.

The general format of the workshop was to dedicate the mornings to invited and contributed talks on topics relating to simulation and collaborative software development in computational and systems neuroscience, leaving the afternoons free for informal discussions and code sprints.

Participants


Group photo from the FACETS CodeJam Workshop #2 More photographs...

Meeting Program

Download (PDF)

List of talks

Full talks

Subhasis Ray
MOOSE to PyMOOSE: Interfacing MOOSE with Python [PDF]
Romain Brette/Dan Goodman
Brian: a pure Python simulator [PDF]
Abigail Morrison
What's new with NEST
Michael Hines
Python + NEURON [PDF]
Stefan Wils
STEPS (STochastic Engine for Pathway Simulation)
Mikael Djurfeldt/Örjan Ekeberg
MUSIC [PDF]
Andrew Davison
What's new with PyNN [PDF]
Daniel Brüderle
PyNN and the FACETS hardware [PDF]
Jon Pierce
PsychoPy [PDF]
Padraig Gleeson
NeuroML and Python/HDF5 support in neuroConstruct
Eilif Muller
NeuroTools I: SpikeTrains and ParameterSets
Thierry Brizzi
NeuroTools II: HDF5
Samuel Garcia
OpenElectrophy: database storage for neural data [PDF]
Andrew Davison
Simulation project management with Sumatra [PDF]

Lightning talks

Moritz Helias
f2py: Python's interface to the world of number crunching [PDF]
Raphel Ritz
The Zope Component Architecture [HTML]
Bernhard Kaplan
Boost::Python [PDF]
Eric Müller
Git: a fast, distributed revision control system [PDF]
Dan Goodman
Sphinx: Python 3.0's documentation system [PDF]
Johannes Bill
The FACETS Graph Model [PDF]

Creative Commons License
Unless mentioned otherwise, all the downloadable talks are licenced under a Creative Commons Attribution-Noncommercial-No Derivative Works License.

Discussions and code sprints

Some highlights of the discussions and the code sprints:

  • Increased compatibility of neuroConstruct generated GENESIS code with MOOSE. (Padraig and Subhasis)
  • Began work on moving the neuroConstruct generated nrnpython code towards a more PyNN friendly format. (Padraig)
  • Discussions on using NetworkML XML/HDF5 format files as a storage format inside PyNN. (Padraig and Andrew)
  • Discussions on releasing the units module of Brian as a stand-alone package, "Piquant". (Dan, Raphael, Andrew, Eilif). Raphael offered support to figure out whether and how to get this package into SciPy. Update from Dan: Piquant is now available on SourceForge.
  • Started implementing the MUSIC interface in NEST and elaborated on a possible interface between NEST and MUSIC in order to make multiple-simulator-simulations with online spike exchange possible. Isolated potential problems between MUSIC and NEST (among them being threaded simulations, simulation time updates in multiples of the actual simulation time step, and simulator resetting) and started to think about possible solutions and necessary adaptations of NEST and of the MUSIC library. (Moritz, Mikael, Örjan and Susanne)
  • Beginnings of a PyNN interface to MOOSE. Implemented the setup(), create() and end() functions and the HH_cond_exp standard model. (Subhasis and Andrew)
  • Work on a microformat parser and a user interface for registering URLs to be harvested for the as-yet-unpublished "INCF Neuroinformatics Community Index". (Raphael)
  • Improved the stp/stdp support for the pyNN.hardware.stage1 module (Daniel, Bernhard, Johannes, Eric, Andrew)
  • Solved some open questions regarding the hardware neuron model implementation for nest2 (Abigail, Moritz, Daniel, Bernhard, Johannes, Eric)
  • Discussion of using pytables and hdf5 for the huge amounts of data generated by the hardware (Thierry, Daniel, Bernhard, Johannes, Eric)
  • Planned future NEST-hardware co-simulation experiments (Jens, Daniel, Bernhard, Johannes, Eric)
  • Began planning a Python interface to MUSIC, and how PyNN and MUSIC might interact. (Mikael and Andrew)
  • Copied the latest changes from the neuralensemble branch of the NEURON code base to the main NEURON Subversion server at Yale. (Michael, Eilif, Andrew)
  • Discussions on the most important extensions to effectively complete the Python + NEURON interoperability (Michael, Andrew, Eilif)
  • The most important thing about every CodeJam: we got to know many nice people from various communities, most of them working with python (Daniel, speaking for all the participants).

Support

The meeting organizers gratefully acknowledge the support of the European Union through the FACETS Project (grant no. IST-2005-15879).