Reference guide: NEST 2.x vs NEST 3.0

• This guide shows changes to functions or their output between PyNEST 2.x and PyNEST 3.0

• Functions not mentioned are unchanged

• Terms that changed for NEST 3.0 are marked in green

• Please note that NEST 3.0 no longer supports Python 2

See also

To see code examples of the key changes, check out our NEST 3.0 Features.

Suppress output on startup

NEST 2.x	NEST 3.0
export DELAY_PYNEST_INIT = 1 import nest nest.ll_api.init([“nest”, “–quiet”])	export PYNEST_QUIET = 1 import nest

Consistent use of allow_offgrid_times

NEST 2.x	NEST 3.0
nest.Create(“spike_generator”, params={“allow_offgrid_spikes”=True})	nest.Create(“spike_generator”, params={“allow_offgrid_times” =True})

Functions related to creation and retrieval of nodes

NEST 2.x	NEST 3.0
nest.Create(model, n=1, params= None) returns list	nest.Create(model, n=1, params= None) returns nest.NodeCollection
nest.GetLid(gid) returns list	nest.GetLocalNodeConnection( nest.NodeConnection) returns the MPI local nodes in a new nest.NodeCollection.

Functions related to connection

NEST 2.x	NEST 3.0
nest.GetConnections(list=None, list=None, synapse_model=None, synapse_label=None) returns numpy.array	nest.GetConnections( nest.NodeCollection =None, nest.NodeCollection =None, synapse_model=None, synapse_label=None) returns nest.SynapseCollection
nest.Connect(list, list, conn_spec =None, syn_spec=None, model=None)	nest.Connect(nest.NodeCollection, nest.NodeCollection, conn_spec= None, syn_spec=None, return_SynapseCollection= False) In syn_spec the synapse model is given by the synapse_model key, where before it was model
nest.DataConnect(pre, post)	Use nest.Connect(np.array, np.array, conn_spec=”one_to_one”)
nest.CGConnect(list, list, cg, parameter_map=params_map, model=’static_synapse’)	Use nest.Connect(nest.NodeCollection, nest.NodeCollection, {“rule”: “conngen”, “cg”: cg, “params_map”: params_map}, synapse_model =’static_synapse’)
nest.DisconnectOneToOne(int, int, syn_spec)	nest.Disconnect(nest.NodeCollection, nest.NodeCollection, syn_spec=’static_synapse’)
nest.Disconnect(list, list, conn_spec= ‘one_to_one’, syn_spec=’static_synapse’)	nest.Disconnect(nest.NodeCollection, nest.NodeCollection, conn_spec= ‘one_to_one’, syn_spec=’static_synapse’)

Functions related to subnets

The subnet model is removed in NEST 3.0!

NEST 2.x	NEST 3.0
nest.PrintNetwork(depth=1, subnet =None)	nest.PrintNodes()
nest.CurrentSubnet()
nest.ChangeSubnet(subnet)
nest.GetLeaves(subnet, properties =None, local_only=False)	nest.NodeCollection will contain all nodes
nest.GetNodes(subnets, properties =None, local_only=False)	GetNodes(properties={}, local_only=False) returns nest.NodeCollection
nest.GetChildren(subnets, properties =None, local_only=False)	nest.NodeCollection will contain all nodes
nest.GetNetwork(gid, depth)
nest.BeginSubnet(label=None, params =None)
nest.EndSubnet()
nest.LayoutNetwork(model, dim, label=None, params=None)	Use nest.Create(model, n=1, params=None, positions=None)

Functions related to setting and getting parameters

NEST 2.x	NEST 3.0
nest.SetStatus(list/tuple, params, val=None)	nest.SetStatus(nest.NodeCollection, params, val=None) Can also use nodes.set(params), nodes.parameter = value, conns.set(params) or conns.parameter = value
nest.GetStatus(list/tuple, keys=None)	nest.GetStatus(nest.NodeCollection, keys=None) Can also use nodes.get(keys=None), nodes.parameter, conns.get(keys=None) or conns.parameter

Functions related to spatially distributed nodes

Spatial structure, formerly provided by the Topology module, is now integrated into NEST and is no longer a separate module.

NEST 2.x	NEST 3.0
tp.CreateLayer(specs) returns tuple of int(s)	nest.Create(model, params=None, positions=nest.spatial.free/grid) returns nest.NodeCollection NOTE: Composite layers no longer possible.
tp.ConnectLayers(list, list, projections)	nest.Connect(nest.NodeCollection, nest.NodeCollection, conn_spec=None, syn_spec=None, return_SynapseCollection =False)
	spatial_NodeCollection.spatial returns Dictionary with spatial properties
tp.GetLayer(nodes) returns tuple	nest.NodeCollection will represent the spatially distributed nodes
tp.GetElement(layers, location) returns tuple	nest.NodeCollection will contain all nodes
tp.GetPosition(tuple) returns tuple of tuple(s)	nest.GetPosition(nest.NodeCollection) returns tuple or tuple of tuple(s)
tp.Displacement(from_arg, to_arg) from_arg: tuple/list of int(s) / tuple/list of tuples/lists of floats] to_arg: tuple/list of int(s) returns tuple	nest.Displacement(from_arg, to_arg) from_arg: nest.NodeCollection or tuple/list with tuple(s)/list(s) of floats to_arg: nest.NodeCollection returns tuple
tp.Distance(from_arg, to_arg) from_arg: [tuple/list of ints / tuple/list with tuples/lists of floats] to:arg: tuple/list of ints returns tuple	nest.Distance(from_arg, to_arg) from_arg: nest.NodeCollection or tuple/list with tuple(s)/list(s) of floats to_arg: nest.NodeCollection returns tuple
tp.FindNearestElement(tuple/list, locations, find_all=True) returns tuple	nest.FindNearestElement(nest.NodeCollection, locations, find_all=True) returns nest.NodeCollection
tp.DumpLayerNodes(tuple, outname)	nest.DumpLayerNodes(nest.NodeCollection, outname)
tp.DumpLayerConnections(tuple, synapse_model, outname)	nest.DumpLayerConnections( nest.NodeCollection, nest.NodeCollection, synapse_model, outname)
tp.FindCenterElement(tuple) returns tuple	nest.FindCenterElement(nest.NodeCollection) returns nest.NodeCollection
tp.GetTargetNodes(tuple, tuple, tgt_model=None, syn_model=None) returns tuple of list(s) of int(s)	nest.GetTargetNodes(nest.NodeCollection, nest.NodeCollection, syn_model=None) returns tuple of nest.NodeConnection
tp.GetTargetPositions(tuple, tuple, tgt_model=None, syn_model=None) returns tuple of tuple(s) of tuple(s) of floats	nest.GetTargetPositions(nest.NodeCollection, nest.NodeCollection, synapse_model=None) returns list of list(s) of tuple(s) of floats
tp.SelectNodesByMask(tuple, anchor, mask_obj) returns list	nest.SelectNodesByMaks(nest.NodeCollection, anchor, mask_obj) returns nest.NodeConnection
tp.PlotLayer(tuple, fig=None, nodecolor=’b’, nodesize=20) returns matplotlib.figure.Figure object	nest.PlotLayer(nest.NodeCollection, fig=None, nodecolor =’b’, nodesize=20) returns matplotlib.figure.Figure object
tp.PlotTargets(int, tuple, tgt_model= =None, syn_type=None, fig=None, mask=None, kernel=None, src_color= ‘red’, src_size=50, tgt_color= ‘blue’, tgt_size=20, mask_color =’red’, kernel_color=’red’) returns matplotlib.figure.Figure object	nest.PlotTargets(nest.NodeCollection, nest.NodeCollection, syn_type= None, fig=None, mask=None, kernel= None, src_color=’red’, src_size= 50, tgt_color=’blue’, tgt_size= ‘20, mask_color=’red’, kernel_color=’red’) returns matplotlib.figure.Figure object
tp.PlotKernel(ax, int, mask, kern=None, mask_color=’red’, kernel_color=’red’)	nest.PlotProbabilityParameter ( nest.NodeCollection, parameter=None, mask=None, edges=[-0.5, 0.5, -0.5, 0.5], shape=[100, 100], ax=None, prob_cmap =’Greens’, mask_color=’yellow’)
‘mask’: {‘volume’: {‘lower_left’: [-2., -1., -1.], ‘upper_right’: [2., 1., 1.]}}	‘mask’: {‘box’ {‘lower_left’: [-2., -1., -1.], ‘upper_right’: [2., 1., 1.]}}

Spatially distributed connection rules

NEST 2.x	NEST 3.0
convergent	pairwise_bernoulli and use_on_source=True
convergent and num_connections	fixed_indegree
divergent	pairwise_bernoulli
divergent and num_connections	fixed_outdegree

Functions related to simulation

NEST 2.x	NEST 3.0
nest.ResetNetwork()	Use nest.ResetKernel() instead
nest.GetKernelStatus(‘time’), nest.SetKernelStatus({‘time’: 0.})	nest.GetKernelStatus(‘biological_time’), nest.SetKernelStatus({‘biological_time’: 0.})

Note

The structural_plasticity_update_interval now reports time in ms instead of number of simulation steps.

Functions related to models

No Change

Functions related to random number generators

In NEST 2.x, you would have to set a global seed and individual seeds for each virtual process manually. In NEST 3.0, you set only a single rng_seed, which is used as a base for all other seeds.

NEST 2.x	NEST 3.0
nest.SetKernelStatus({’grng_seed’ : msd+N_vp})	nest.SetKernelStatus({‘rng_seed’: msd})
nest.SetKernelStatus({’rng_seeds’ : range( msd+N_vp+1, msd+2*N_vp+1)})
nest.ll_api.sli_func(‘rngdict keys’)	nest.GetKernelStatus(‘rng_types’)
nest.ll_api.sli_run(‘0 << /grng rngdict/MT19937 :: 101 CreateRNG >> SetStatus’)	nest.SetKernelStatus({‘rng_type’: ‘mt19937’, ‘rng_seed’: 101})

Functions related to parallel computing

No Change

Model names

• The synapse model stdp_nn_pre-centered_synapse has been renamed to stdp_nn_pre_centered_synapse for consistency with the naming convention.

Parameters

Parameters can now be used to set node and connection parameters.

Note

Check out the section on Parametrization for example usage

random

The random module contains random distributions that can be used to set node and connection parameters, as well as positions for spatially distributed nodes.

NEST 2.x	NEST 3.0
	nest.random.exponential(beta=1.0) returns nest.Parameter
	nest.random.lognormal(mean=0.0, std=1.0) returns nest.Parameter
	nest.random.normal(mean=0.0, std=1.0) returns nest.Parameter
	nest.random.uniform(min=0.0, max=1.0) returns nest.Parameter

spatial

The spatial module contains parameters related to spatial positions for the nodes.

NEST 2.x	NEST 3.0
	nest.spatial.distance.x nest.spatial.distance.y nest.spatial.distance.z returns nest.Parameter
	nest.spatial.distance returns nest.Parameter
	nest.spatial.free(pos, extent=None, edge_wrap=False, num_dimensions=None) returns nest.Parameter
	nest.spatial.grid(shape, center=None, extent=None, edge_wrap=False) returns nest.Parameter
	nest.spatial.pos.x nest.spatial.pos.y nest.spatial.pos.z returns nest.Parameter
	nest.spatial.source_pos.x nest.spatial.source_pos.y nest.spatial.source_pos.z returns nest.Parameter
	nest.spatial.target_pos.x nest.spatial.target_pos.y nest.spatial.target_pos.z returns nest.Parameter

math

The math module contains parameters for mathematical expressions. The mathematical expressions all take a nest.Parameter.

NEST 2.X	NEST 3.0
	nest.math.exp(nest.Parameter) returns nest.Parameter
	nest.math.sin(nest.Parameter) returns nest.Parameter
	nest.math.cos(nest.Parameter) returns nest.Parameter
	nest.math.min(nest.Parameter, value) returns nest.Parameter
	nest.math.max(nest.Parameter, value) returns nest.Parameter
	nest.math.redraw(nest.Parameter, min, max) returns nest.Parameter

logic

The logic module contains logical expressions between nest.Parameter’s.

NEST 2.x	NEST 3.0
	nest.logic.conditional(condition, param_if_true, param_if_false) returns nest.Parameter

spatial_distributions

The spatial_distributions module contains random distributions that take a spatial parameter as input and applies the distribution on the parameter. They are used for spatially distributed nodes.

NEST 2.x	NEST 3.0
	nest.spatial_distributions.exponential(nest.Parameter, beta=1.0) returns nest.Parameter
	nest.spatial_distributions.gaussian(nest.Parameter, mean=0.0, std=1.0) returns nest.Parameter
	nest.spatial_distributions.gaussian2D(nest.Parameter, nest.Parameter, mean_x=0.0, mean_y=0.0, std_x=1.0, std_y=1.0, rho=0.0) returns nest.Parameter
	nest.spatial_distributions.gamma(nest.Parameter, kappa=1.0 theta=1.0) returns nest.Parameter