Public | Automated Build

Last pushed: 18 days ago
Short Description
Pype9 is a collection of Python pipelines for simulating networks of neuron models described in 9ML
Full Description


.. image::
.. image::
.. image::
:alt: Supported Python versions
.. image::
:alt: Latest Version
.. image::
:alt: Documentation Status

PYthon PipelinEs for 9ML (Pype9) is a collection of Python pipelines
for simulating networks of neuron models described in NineML_ with various
simulator backends.


Supported Simulators

Pype9 works with either or both of the following simulator backends

  • Neuron_ >= 7.5
  • NEST_ >= 2.14.0

Detailed instructions on how to install these simulators on different platforms
can be found in the Installation documentation_.

Unsupported NineML

NineML aims to be a comprehensive description language for neural simulation. This
means that it allows the expression of some uncommon configurations that are
difficult to implement in Neuron
and NEST. Work is planned to make the NEURON
and NEST pipelines in Pype9 support NineML
fully, however until then the
following restrictions apply to models that can be used with Pype9.

  • synapses must be linear
  • synapses can only have one variable that varies over a projection (e.g.
  • no recurrent analog connections between populations (e.g. gap junctions)
  • only one event send port per cell
  • names given to NineML_ elements are not escaped and therefore can clash with
    built-in keywords and some PyPe9 method names (e.g. 'lambda' is a reserved
    keyword in Python). Please avoid using names that clash with C++ or Python
    keywords (NB: This will be fixed in future versions).


Given a cell model described in NineML_ saved in
my_hodgkin_huxley.xml, the simulator pipeline can run from the command line:

.. code-block:: bash

$ pype9 simulate my_hodgkin_huxley.xml#hh_props neuron 100.0 0.01 \
--play isyn isyn.neo.pkl --record v v.neo.pkl --init_value v -65.0 mV

or in a Python script

.. code-block:: python

from pype9.simulator.neuron import cell, Simulation
from nineml import units as un

HodgkinHuxley = cell.MetaClass('my_hodgkin_huxley.xml#hh_class')
with Simulation(dt=0.01, seed=1234) as sim:
hh = HodgkinHuxley('my_hodgkin_huxley.xml#hh_props', v=-65.0
hh.record('v') *
v = hh.recording('v')

Pype9 also supports network models described in NineML via integration with PyNN

.. code-block:: bash

$ pype9 simulate brunel.xml nest 1000.0 0.01 \
--record Exc.spike_output Exc-nest.neo.pkl \
--record Inh.spike_output Inh-nest.neo.pkl \
--seed 12345


.. code-block:: python

from pype9.simulator.neuron import Network, Simulation
from nineml import units as un

with Simulation(dt=0.01, seed=1234) as sim:
brunel_ai = Network('brunel.xml#AI')
exc_spikes = brunel_ai.component_array('Exc').recording('spike_output')
inh_spikes = brunel_ai.component_array('Inh').recording('spike_output')

See Creating Simulations in Python_ in the Pype9 docs for more examples and pipelines.

In addition to the simulate command there is also a plot command for
conveniently plotting the results of the simulation with Matplotlib,
and a convert command to convert NineML
files between different serialization
formats (XML, YAML, JSON and HDF5) and NineML_ versions (1.0 and 2.0dev). See the
documentation for details.

:copyright: Copyright 20012-2016 by the Pype9 team, see AUTHORS.
:license: MIT, see LICENSE for details.

.. PyNN:
NeuralEnsemble Google Group:!forum/neuralensemble
.. Matplotlib:
Creating Simulations in Python:
.. _Installation documentation:
.. _NineML:
.. _NEST:
.. _Neuron:

Docker Pull Command
Source Repository