Recipes¶
The arb::recipe
class documentation is below.
C++ best practices¶
Here we collect rules of thumb to keep in mind when making recipes in C++.
Stay thread safe
The load balancing and model construction are multithreaded, that is multiple threads query the recipe simultaneously. Hence calls to a recipe member should not have side effects, and should use lazy evaluation when possible (see Be lazy).
Recipe¶
-
class
recipe
¶ A description of a model, describing the cells and network, without any information about how the model is to be represented or executed.
All recipes derive from this abstract base class, defined in
src/recipe.hpp
.Recipes provide a cell-centric interface for describing a model. This means that model properties, such as connections, are queried using the global identifier (gid) of a cell. In the description below, the term gid is used as shorthand for “the cell with global identifier gid”.
Warning
All member functions must be thread safe, because the recipe is used by the multithreaded model building stage. In practice, this means that multiple threads should be able to call member functions of a recipe simultaneously. Model building is multithreaded to reduce model building times, so recipe implementations should avoid using locks and mutexes to introduce thread safety. See recipe best practices for more information.
Required Member Functions
The following member functions must be implemented by every recipe:
-
virtual cell_size_type
num_cells
() const = 0¶ The number of cells in the model.
-
virtual cell_kind
get_cell_kind
(cell_gid_type gid) const = 0¶ The kind of gid (see
arb::cell_kind
).
-
virtual util::unique_any
get_cell_description
(cell_gid_type gid) const = 0¶ A description of the cell gid, for example the morphology, synapses and ion channels required to build a multi-compartment neuron.
The type used to describe a cell depends on the kind of the cell. The interface for querying the kind and description of a cell are separate to allow the cell type to be provided without building a full cell description, which can be very expensive.
Optional Member Functions
-
virtual std::vector<cell_connection>
connections_on
(cell_gid_type gid) const¶ Returns a list of all the incoming connections for gid . Each connection
con
should have post-synaptic targetcon.dest.gid
that matches the argumentgid
, and a valid synapse idcon.dest.index
on gid. Seecell_connection
.By default returns an empty list.
-
virtual std::vector<gap_junction_connection>
gap_junctions_on
(cell_gid_type gid) const¶ Returns a list of all the gap junctions connected to gid. Each gap junction
gj
should have one of the two gap junction sitesgj.local.gid
orgj.peer.gid
matching the argumentgid
, and the corresponding synapse idgj.local.index
orgj.peer.index
should be valid on gid. Seegap_junction_connection
.By default returns an empty list.
-
virtual std::vector<event_generator>
event_generators
(cell_gid_type gid) const¶ Returns a list of all the event generators that are attached to gid.
By default returns an empty list.
-
virtual cell_size_type
num_sources
(cell_gid_type gid) const¶ Returns the number of spike sources on gid. This corresponds to the number of spike detectors on a multi-compartment cell. Typically there is one detector at the soma of the cell, however it is possible to attache multiple detectors at arbitrary locations.
By default returns 0.
-
virtual cell_size_type
num_targets
(cell_gid_type gid) const¶ The number of post-synaptic sites on gid, which corresponds to the number of synapses.
By default returns 0.
-
virtual cell_size_type
num_gap_junction_sites
(cell_gid_type gid) const¶ Returns the number of gap junction sites on gid.
By default returns 0.
-
virtual std::vector<probe_info>
get_probes
(cell_gid_type gid) const¶ Intended for use by cell group implementations to set up sampling data structures ahead of time and for putting in place any structures or information in the concrete cell implementations to allow monitoring.
Returns a vector containing (in order) all the probes on a given cell gid.
By default returns an empty vector.
-
virtual cell_size_type
Synapses¶
See Interconnectivity.
Probes¶
-
using
probe_tag
= int¶ Extra contextual information associated with a probe.