This class is a representation of the dictionary of connection properties given as an argument to the ConnectLayers function. More...

#include <connection_creator.h>

Classes
class	PoolWrapper_
	Wrapper for masked and unmasked pools. More...

Public Types
enum	ConnectionType { Target_driven, Source_driven, Convergent, Divergent }

Public Member Functions
	ConnectionCreator (DictionaryDatum dict)
	Construct a ConnectionCreator with the properties defined in the given dictionary. More...

template<int D>
void	connect (Layer< D > &source, Layer< D > &target)
	Connect two layers. More...

Private Member Functions
template<typename Iterator , int D>
void	connect_to_target_ (Iterator from, Iterator to, Node *tgt_ptr, const Position< D > &tgt_pos, thread tgt_thread, const Layer< D > &source)

template<int D>
void	target_driven_connect_ (Layer< D > &source, Layer< D > &target)

template<int D>
void	source_driven_connect_ (Layer< D > &source, Layer< D > &target)

template<int D>
void	convergent_connect_ (Layer< D > &source, Layer< D > &target)

template<int D>
void	divergent_connect_ (Layer< D > &source, Layer< D > &target)

void	connect_ (index s, Node *target, thread target_thread, double_t w, double_t d, index syn)

template<int D>
void	get_parameters_ (const Position< D > &pos, librandom::RngPtr rng, double &weight, double &delay)
	Calculate parameter values for this position. More...

Private Attributes
ConnectionType	type_

bool	allow_autapses_

bool	allow_multapses_

bool	allow_oversized_

Selector	source_filter_

Selector	target_filter_

index	number_of_connections_

lockPTR< AbstractMask >	mask_

lockPTR< Parameter >	kernel_

index	synapse_model_

lockPTR< Parameter >	weight_

lockPTR< Parameter >	delay_

Network &	net_

Detailed Description

This class is a representation of the dictionary of connection properties given as an argument to the ConnectLayers function.

The connect method is responsible for generating the connection according to the given parameters. This method is templated with the dimension of the layers, and is called via the Layer connect call using a visitor pattern. The connect method relays to another method (e.g., convergent_connect_) implementing the concrete connection algorithm. It would be more elegant if this was a base class for classes representing different connection algorithms with a virtual connect method, but it is not possible to have a virtual template method.

This class distinguishes between target driven and convergent connections, which are both called "convergent" in the Topology module documentation, and between source driven and divergent connections. The true convergent/divergent connections are those with a fixed number of connections (fan in/out). The only difference between source driven and target driven connections is which layer coordinates the mask and parameters are defined in.

Member Enumeration Documentation

enum nest::ConnectionCreator::ConnectionType

Enumerator
Target_driven
Source_driven
Convergent
Divergent

Constructor & Destructor Documentation

nest::ConnectionCreator::ConnectionCreator ( DictionaryDatum dict )

Construct a ConnectionCreator with the properties defined in the given dictionary.

Parameters for a ConnectionCreator are:

"connection_type": Either "convergent" or "convergent".
"allow_autapses": Boolean, true if autapses are allowed.
"allow_multapses": Boolean, true if multapses are allowed.
"allow_oversized": Boolean, true if oversized masks are allowed.
"number_of_connections": Integer, number of connections to make for each source or target.
"mask": Mask definition (dictionary or masktype).
"kernel": Kernel definition (dictionary, parametertype, or double).
"synapse_model": The synapse model to use.
"targets": Which targets (model or lid) to select (dictionary).
"sources": Which targets (model or lid) to select (dictionary).
"weights": Synaptic weight (dictionary, parametertype, or double).
"delays": Synaptic delays (dictionary, parametertype, or double).
other parameters are interpreted as synapse parameters, and may be defined by a dictionary, parametertype, or double.
Parameters

dict dictionary containing properties for the connections.

References nest::names::allow_autapses, allow_autapses_, nest::names::allow_multapses, allow_multapses_, allow_oversized_, nest::names::allow_oversized_mask, nest::names::connection_type, nest::names::convergent, Convergent, nest::TopologyModule::create_mask(), nest::TopologyModule::create_parameter(), nest::names::delay, delay_, nest::names::delays, nest::names::divergent, Divergent, nest::Network::get_connector_defaults(), nest::Network::get_synapsedict(), getValue< bool >(), nest::names::kernel, kernel_, Dictionary::lookup(), nest::names::mask, mask_, net_, nest::names::number_of_connections, number_of_connections_, Source_driven, source_filter_, nest::names::sources, nest::names::synapse_model, synapse_model_, Target_driven, target_filter_, nest::names::targets, type_, nest::names::weight, weight_, and nest::names::weights.

Member Function Documentation

template<int D>

void nest::ConnectionCreator::connect	(	Layer< D > &	source,
		Layer< D > &	target
	)

Connect two layers.

Parameters

source	source layer.
target	target layer.

References Convergent, convergent_connect_(), Divergent, divergent_connect_(), Source_driven, source_driven_connect_(), Target_driven, target_driven_connect_(), and type_.

Referenced by nest::Layer< D >::connect().

void nest::ConnectionCreator::connect_	(	index	s,
		Node *	target,
		thread	target_thread,
		double_t	w,
		double_t	d,
		index	syn
	)

inlineprivate

References nest::Network::connect(), nest::Node::get_gid(), nest::Network::get_thread_id(), nest::Network::is_local_gid(), and net_.

Referenced by connect_to_target_().

template<typename Iterator , int D>

void nest::ConnectionCreator::connect_to_target_	(	Iterator	from,
		Iterator	to,
		Node *	tgt_ptr,
		const Position< D > &	tgt_pos,
		thread	tgt_thread,
		const Layer< D > &	source
	)