def test_gating_signal_and_gating_projection_names(self):
T3 = pnl.TransferMechanism(name='T3')
T4 = pnl.TransferMechanism(name='T4', input_states=['First State','Second State'])
# GatingSignal with one GatingProjection
G1 = pnl.GatingMechanism(gating_signals=[T3])
assert G1.gating_signals[0].name == 'T3[InputState-0] GatingSignal'
assert G1.gating_signals[0].efferents[0].name == 'GatingProjection for T3[InputState-0]'
# GatingSignal with two GatingProjections to two States of same Mechanism
G2 = pnl.GatingMechanism(gating_signals=[{pnl.PROJECTIONS:[T4.input_states[0], T4.input_states[1]]}])
assert G2.gating_signals[0].name == 'T4[First State, Second State] GatingSignal'
assert G2.gating_signals[0].efferents[0].name == 'GatingProjection for T4[First State]'
assert G2.gating_signals[0].efferents[1].name == 'GatingProjection for T4[Second State]'
# GatingSignal with two GatingProjections to two States of different Mechanisms
G3 = pnl.GatingMechanism(gating_signals=[{pnl.PROJECTIONS:[T3, T4]}])
assert G3.gating_signals[0].name == 'GatingSignal-0 divergent GatingSignal'
assert G3.gating_signals[0].efferents[0].name == 'GatingProjection for T3[InputState-0]'
assert G3.gating_signals[0].efferents[1].name == 'GatingProjection for T4[First State]'
# GatingProjections to ProcessingMechanism from GatingSignals of existing GatingMechanism
T5 = pnl.TransferMechanism(name='T5',
input_states=[T3.output_states[pnl.RESULTS],
G3.gating_signals['GatingSignal-0 divergent GatingSignal']],
output_states=[G3.gating_signals['GatingSignal-0 divergent GatingSignal']])
def test_formats_for_gating_specification_of_input_and_output_states(self):
gating_spec_list = [
pnl.GATING, pnl.GATING_SIGNAL, pnl.GATING_PROJECTION,
pnl.GatingSignal,
pnl.GatingSignal(), pnl.GatingProjection, "GP_OBJECT",
pnl.GatingMechanism,
pnl.GatingMechanism(), (0.3, pnl.GATING), (0.3, pnl.GATING_SIGNAL),
(0.3, pnl.GATING_PROJECTION), (0.3, pnl.GatingSignal),
(0.3, pnl.GatingSignal()), (0.3, pnl.GatingProjection),
(0.3, "GP_OBJECT"), (0.3, pnl.GatingMechanism),
(0.3, pnl.GatingMechanism())
]
for i, gating_tuple in enumerate(
[j for j in zip(gating_spec_list, reversed(gating_spec_list))]):
G1, G2 = gating_tuple
# This shenanigans is to avoid assigning the same instantiated ControlProjection more than once
if G1 is 'GP_OBJECT':
G1 = pnl.GatingProjection()
elif isinstance(G1, tuple) and G1[1] is 'GP_OBJECT':
G1 = (G1[0], pnl.GatingProjection())
if G2 is 'GP_OBJECT':
G2 = pnl.GatingProjection()
elif isinstance(G2, tuple) and G2[1] is 'GP_OBJECT':
G2 = (G2[0], pnl.GatingProjection())
T = pnl.TransferMechanism(name='T-GATING-{}'.format(i),
input_states=[G1],
output_states=[G2])
assert T.input_states[0].mod_afferents[0].name in \
'GatingProjection for T-GATING-{}[InputState-0]'.format(i)
assert T.output_states[0].mod_afferents[0].name in \
'GatingProjection for T-GATING-{}[OutputState-0]'.format(i)
def test_2_item_tuple_from_gating_signal_to_output_states(self):
D4 = pnl.DDM(name='D4')
# Single name
G = pnl.GatingMechanism(gating_signals=[(pnl.DECISION_VARIABLE, D4)])
assert G.gating_signals[0].name == 'D4[DECISION_VARIABLE] GatingSignal'
assert G.gating_signals[0].efferents[0].receiver.name == 'DECISION_VARIABLE'
# List of names
G = pnl.GatingMechanism(gating_signals=[([pnl.DECISION_VARIABLE, pnl.RESPONSE_TIME], D4)])
assert G.gating_signals[0].name == 'D4[DECISION_VARIABLE, RESPONSE_TIME] GatingSignal'
assert G.gating_signals[0].efferents[0].receiver.name == 'DECISION_VARIABLE'
assert G.gating_signals[0].efferents[1].receiver.name == 'RESPONSE_TIME'
def test_multiple_modulatory_projections_with_state_name(self):
M = pnl.DDM(name='MY DDM')
C = pnl.ControlMechanism(control_signals=[{
'DECISION_CONTROL': [
M.parameter_states[pnl.DRIFT_RATE], M.parameter_states[
pnl.THRESHOLD]
]
}])
G = pnl.GatingMechanism(gating_signals=[{
'DDM_OUTPUT_GATE': [
M.output_states[pnl.DECISION_VARIABLE], M.output_states[
pnl.RESPONSE_TIME]
]
}])
assert len(C.control_signals) == 1
assert C.control_signals[0].name == 'DECISION_CONTROL'
assert len(C.control_signals[0].efferents) == 2
assert M.parameter_states[pnl.DRIFT_RATE].mod_afferents[
0] == C.control_signals[0].efferents[0]
assert M.parameter_states[pnl.THRESHOLD].mod_afferents[
0] == C.control_signals[0].efferents[1]
assert len(G.gating_signals) == 1
assert G.gating_signals[0].name == 'DDM_OUTPUT_GATE'
assert len(G.gating_signals[0].efferents) == 2
assert M.output_states[pnl.DECISION_VARIABLE].mod_afferents[
0] == G.gating_signals[0].efferents[0]
assert M.output_states[pnl.RESPONSE_TIME].mod_afferents[
0] == G.gating_signals[0].efferents[1]
def test_multiple_modulatory_projections_with_port_Name(self):
M = pnl.DDM(name='MY DDM')
C = pnl.ControlMechanism(control_signals=[{
'DECISION_CONTROL': [
M.parameter_ports[psyneulink.core.components.functions.
distributionfunctions.DRIFT_RATE],
M.parameter_ports[psyneulink.core.globals.keywords.THRESHOLD]
]
}])
G = pnl.GatingMechanism(gating_signals=[{
'DDM_OUTPUT_GATE': [
M.output_ports[pnl.DECISION_VARIABLE], M.output_ports[
pnl.RESPONSE_TIME]
]
}])
assert len(C.control_signals) == 1
assert C.control_signals[0].name == 'DECISION_CONTROL'
assert len(C.control_signals[0].efferents) == 2
assert M.parameter_ports[
psyneulink.core.components.functions.distributionfunctions.
DRIFT_RATE].mod_afferents[0] == C.control_signals[0].efferents[0]
assert M.parameter_ports[
psyneulink.core.globals.keywords.
THRESHOLD].mod_afferents[0] == C.control_signals[0].efferents[1]
assert len(G.gating_signals) == 1
assert G.gating_signals[0].name == 'DDM_OUTPUT_GATE'
assert len(G.gating_signals[0].efferents) == 2
assert M.output_ports[pnl.DECISION_VARIABLE].mod_afferents[
0] == G.gating_signals[0].efferents[0]
assert M.output_ports[pnl.RESPONSE_TIME].mod_afferents[
0] == G.gating_signals[0].efferents[1]
def test_multiple_modulatory_projections_with_mech_and_port_Name_specs(
self):
M = pnl.DDM(name='MY DDM')
C = pnl.ControlMechanism(control_signals=[{
pnl.MECHANISM:
M,
pnl.PARAMETER_PORTS: [
psyneulink.core.components.functions.distributionfunctions.
DRIFT_RATE, psyneulink.core.globals.keywords.THRESHOLD
]
}])
G = pnl.GatingMechanism(gating_signals=[{
pnl.MECHANISM:
M,
pnl.OUTPUT_PORTS: [pnl.DECISION_VARIABLE, pnl.RESPONSE_TIME]
}])
assert len(C.control_signals) == 1
assert len(C.control_signals[0].efferents) == 2
assert M.parameter_ports[
psyneulink.core.components.functions.distributionfunctions.
DRIFT_RATE].mod_afferents[0] == C.control_signals[0].efferents[0]
assert M.parameter_ports[
psyneulink.core.globals.keywords.
THRESHOLD].mod_afferents[0] == C.control_signals[0].efferents[1]
assert len(G.gating_signals) == 1
assert len(G.gating_signals[0].efferents) == 2
assert M.output_ports[pnl.DECISION_VARIABLE].mod_afferents[
0] == G.gating_signals[0].efferents[0]
assert M.output_ports[pnl.RESPONSE_TIME].mod_afferents[
0] == G.gating_signals[0].efferents[1]
def test_multiple_modulatory_projection_specs(self):
M = pnl.DDM(name='MY DDM')
C = pnl.ControlMechanism(control_signals=[{
pnl.PROJECTIONS: [
M.parameter_ports[
psyneulink.core.components.functions.nonstateful.
distributionfunctions.DRIFT_RATE], M.parameter_ports[
psyneulink.core.globals.keywords.THRESHOLD]
]
}])
G = pnl.GatingMechanism(gating_signals=[{
pnl.PROJECTIONS: [
M.output_ports[pnl.DECISION_VARIABLE], M.output_ports[
pnl.RESPONSE_TIME]
]
}])
assert len(C.control_signals) == 1
assert len(C.control_signals[0].efferents) == 2
assert M.parameter_ports[
psyneulink.core.components.functions.nonstateful.
distributionfunctions.
DRIFT_RATE].mod_afferents[0] == C.control_signals[0].efferents[0]
assert M.parameter_ports[
psyneulink.core.globals.keywords.
THRESHOLD].mod_afferents[0] == C.control_signals[0].efferents[1]
assert len(G.gating_signals) == 1
assert len(G.gating_signals[0].efferents) == 2
assert M.output_ports[pnl.DECISION_VARIABLE].mod_afferents[
0] == G.gating_signals[0].efferents[0]
assert M.output_ports[pnl.RESPONSE_TIME].mod_afferents[
0] == G.gating_signals[0].efferents[1]
def test_multiple_modulatory_projection_specs(self):
M = pnl.DDM(name='MY DDM')
C = pnl.ControlMechanism(control_signals=[{
pnl.PROJECTIONS: [
M.parameter_states[pnl.DRIFT_RATE], M.parameter_states[
pnl.THRESHOLD]
]
}])
G = pnl.GatingMechanism(gating_signals=[{
pnl.PROJECTIONS: [
M.output_states[pnl.DECISION_VARIABLE], M.output_states[
pnl.RESPONSE_TIME]
]
}])
assert len(C.control_signals) == 1
assert len(C.control_signals[0].efferents) == 2
assert M.parameter_states[pnl.DRIFT_RATE].mod_afferents[
0] == C.control_signals[0].efferents[0]
assert M.parameter_states[pnl.THRESHOLD].mod_afferents[
0] == C.control_signals[0].efferents[1]
assert len(G.gating_signals) == 1
assert len(G.gating_signals[0].efferents) == 2
assert M.output_states[pnl.DECISION_VARIABLE].mod_afferents[
0] == G.gating_signals[0].efferents[0]
assert M.output_states[pnl.RESPONSE_TIME].mod_afferents[
0] == G.gating_signals[0].efferents[1]
def test_multiple_modulatory_projections_with_mech_and_state_name_specs(
self):
M = pnl.DDM(name='MY DDM')
C = pnl.ControlMechanism(control_signals=[{
pnl.MECHANISM:
M,
pnl.PARAMETER_STATES: [pnl.DRIFT_RATE, pnl.THRESHOLD]
}])
G = pnl.GatingMechanism(gating_signals=[{
pnl.MECHANISM:
M,
pnl.OUTPUT_STATES: [pnl.DECISION_VARIABLE, pnl.RESPONSE_TIME]
}])
assert len(C.control_signals) == 1
assert len(C.control_signals[0].efferents) == 2
assert M.parameter_states[pnl.DRIFT_RATE].mod_afferents[
0] == C.control_signals[0].efferents[0]
assert M.parameter_states[pnl.THRESHOLD].mod_afferents[
0] == C.control_signals[0].efferents[1]
assert len(G.gating_signals) == 1
assert len(G.gating_signals[0].efferents) == 2
assert M.output_states[pnl.DECISION_VARIABLE].mod_afferents[
0] == G.gating_signals[0].efferents[0]
assert M.output_states[pnl.RESPONSE_TIME].mod_afferents[
0] == G.gating_signals[0].efferents[1]
def test_formats_for_gating_specification_of_input_and_output_states(self):
gating_spec_list = [
pnl.GATING, pnl.GATING_SIGNAL, pnl.GATING_PROJECTION,
pnl.GatingSignal, pnl.GatingSignal,
pnl.GatingSignal(), pnl.GatingProjection, "GP_OBJECT",
pnl.GatingMechanism, pnl.ModulatoryMechanism,
pnl.GatingMechanism(), (0.3, pnl.GATING), (0.3, pnl.GATING_SIGNAL),
(0.3, pnl.GATING_PROJECTION), (0.3, pnl.GatingSignal),
(0.3, pnl.GatingSignal()), (0.3, pnl.GatingProjection),
(0.3, "GP_OBJECT"), (0.3, pnl.GatingMechanism),
(0.3, pnl.ModulatoryMechanism), (0.3, pnl.GatingMechanism())
]
for i, gating_tuple in enumerate(
[j for j in zip(gating_spec_list, reversed(gating_spec_list))]):
G1, G2 = gating_tuple
# This shenanigans is to avoid assigning the same instantiated ControlProjection more than once
if G1 is 'GP_OBJECT':
G1 = pnl.GatingProjection()
elif isinstance(G1, tuple) and G1[1] is 'GP_OBJECT':
G1 = (G1[0], pnl.GatingProjection())
if G2 is 'GP_OBJECT':
G2 = pnl.GatingProjection()
elif isinstance(G2, tuple) and G2[1] is 'GP_OBJECT':
G2 = (G2[0], pnl.GatingProjection())
T = pnl.TransferMechanism(name='T-GATING-{}'.format(i),
input_states=[G1],
output_states=[G2])
assert T.input_states[0].mod_afferents[0].name in \
'GatingProjection for T-GATING-{}[InputState-0]'.format(i)
assert T.output_states[0].mod_afferents[0].name in \
'GatingProjection for T-GATING-{}[OutputState-0]'.format(i)
with pytest.raises(pnl.ProjectionError) as error_text:
T1 = pnl.ProcessingMechanism(
name='T1', input_states=[pnl.ModulatoryMechanism()])
assert 'Primary OutputState of ModulatoryMechanism-0 (ControlSignal-0) ' \
'cannot be used as a sender of a Projection to InputState of T1' in error_text.value.args[0]
with pytest.raises(pnl.ProjectionError) as error_text:
T2 = pnl.ProcessingMechanism(
name='T2', output_states=[pnl.ModulatoryMechanism()])
assert 'Primary OutputState of ModulatoryMechanism-1 (ControlSignal-0) ' \
'cannot be used as a sender of a Projection to OutputState of T2' in error_text.value.args[0]
def test_2_item_tuple_from_input_and_output_ports_to_gating_signals(self):
G = pnl.GatingMechanism(gating_signals=['a', 'b'])
T = pnl.TransferMechanism(name='T',
input_ports=[(3, G)],
output_ports=[(2, G.gating_signals['b'])])
assert T.input_ports[0].mod_afferents[0].sender == G.gating_signals[0]
assert T.output_ports[0].mod_afferents[0].sender == G.gating_signals[1]
def test_gating(benchmark, comp_mode):
Input_Layer = pnl.TransferMechanism(
name='Input_Layer',
default_variable=np.zeros((2,)),
function=pnl.Logistic()
)
Output_Layer = pnl.TransferMechanism(
name='Output_Layer',
default_variable=[0, 0, 0],
function=pnl.Linear(),
output_ports={
pnl.NAME: 'RESULTS USING UDF',
pnl.FUNCTION: pnl.Linear(slope=pnl.GATING)
}
)
Gating_Mechanism = pnl.GatingMechanism(
size=[1],
gating_signals=[Output_Layer.output_port]
)
p_pathway = [Input_Layer, Output_Layer]
stim_list = {
Input_Layer: [[-1, 30], [-1, 30], [-1, 30], [-1, 30]],
Gating_Mechanism: [[0.0], [0.5], [1.0], [2.0]]
}
comp = pnl.Composition(name="comp")
comp.add_linear_processing_pathway(p_pathway)
comp.add_node(Gating_Mechanism)
comp.run(num_trials=4, inputs=stim_list, execution_mode=comp_mode)
expected_results = [
[np.array([0., 0., 0.])],
[np.array([0.63447071, 0.63447071, 0.63447071])],
[np.array([1.26894142, 1.26894142, 1.26894142])],
[np.array([2.53788284, 2.53788284, 2.53788284])]
]
np.testing.assert_allclose(comp.results, expected_results)
if benchmark.enabled:
benchmark(comp.run, num_trials=4, inputs=stim_list, execution_mode=comp_mode)
class TestProjectionSpecificationFormats:
def test_projection_specification_formats(self):
"""Test various matrix and Projection specifications
Also tests assignment of Projections to pathay of Composition using add_linear_processing_pathway:
- Projection explicitly specified in sequence (M1_M2_proj)
- Projection pre-constructed and assigned to Mechanisms, but not specified in pathway(M2_M3_proj)
- Projection specified in pathway that is duplicate one preconstructed and assigned to Mechanisms (M3_M4_proj)
(currently it should be ignored; in the future, if/when Projections between the same sender and receiver
in different Compositions are allowed, then it should be used)
"""
M1 = pnl.ProcessingMechanism(size=2)
M2 = pnl.ProcessingMechanism(size=5)
M3 = pnl.ProcessingMechanism(size=4)
M4 = pnl.ProcessingMechanism(size=3)
M1_M2_matrix = (np.arange(2 * 5).reshape((2, 5)) + 1) / (2 * 5)
M2_M3_matrix = (np.arange(5 * 4).reshape((5, 4)) + 1) / (5 * 4)
M3_M4_matrix_A = (np.arange(4 * 3).reshape((4, 3)) + 1) / (4 * 5)
M3_M4_matrix_B = (np.arange(4 * 3).reshape((4, 3)) + 1) / (4 * 3)
M1_M2_proj = pnl.MappingProjection(matrix=M1_M2_matrix)
M2_M3_proj = pnl.MappingProjection(sender=M2,
receiver=M3,
matrix={
pnl.VALUE: M2_M3_matrix,
pnl.FUNCTION:
pnl.AccumulatorIntegrator,
pnl.FUNCTION_PARAMS: {
pnl.DEFAULT_VARIABLE:
M2_M3_matrix,
pnl.INITIALIZER:
M2_M3_matrix
}
})
M3_M4_proj_A = pnl.MappingProjection(sender=M3,
receiver=M4,
matrix=M3_M4_matrix_A)
c = pnl.Composition()
c.add_linear_processing_pathway(
pathway=[M1, M1_M2_proj, M2, M3, M3_M4_matrix_B, M4])
assert np.allclose(M2_M3_proj.matrix.base, M2_M3_matrix)
assert M2.efferents[0] is M2_M3_proj
assert np.allclose(M3.efferents[0].matrix.base, M3_M4_matrix_A)
# This is if different Projections are allowed between the same sender and receiver in different Compositions:
# assert np.allclose(M3.efferents[1].matrix, M3_M4_matrix_B)
c.run(inputs={M1: [2, -30]})
# assert np.allclose(c.results, [[-130.19166667, -152.53333333, -174.875]])
assert np.allclose(c.results, [[-78.115, -91.52, -104.925]])
def test_multiple_modulatory_projection_specs(self):
M = pnl.DDM(name='MY DDM')
C = pnl.ControlMechanism(control_signals=[{
pnl.PROJECTIONS: [
M.parameter_ports[psyneulink.core.components.functions.
distributionfunctions.DRIFT_RATE],
M.parameter_ports[psyneulink.core.globals.keywords.THRESHOLD]
]
}])
G = pnl.GatingMechanism(gating_signals=[{
pnl.PROJECTIONS: [
M.output_ports[pnl.DECISION_VARIABLE], M.output_ports[
pnl.RESPONSE_TIME]
]
}])
assert len(C.control_signals) == 1
assert len(C.control_signals[0].efferents) == 2
assert M.parameter_ports[
psyneulink.core.components.functions.distributionfunctions.
DRIFT_RATE].mod_afferents[0] == C.control_signals[0].efferents[0]
assert M.parameter_ports[
psyneulink.core.globals.keywords.
THRESHOLD].mod_afferents[0] == C.control_signals[0].efferents[1]
assert len(G.gating_signals) == 1
assert len(G.gating_signals[0].efferents) == 2
assert M.output_ports[pnl.DECISION_VARIABLE].mod_afferents[
0] == G.gating_signals[0].efferents[0]
assert M.output_ports[pnl.RESPONSE_TIME].mod_afferents[
0] == G.gating_signals[0].efferents[1]
def test_multiple_modulatory_projections_with_port_Name(self):
M = pnl.DDM(name='MY DDM')
C = pnl.ControlMechanism(control_signals=[{
'DECISION_CONTROL': [
M.parameter_ports[psyneulink.core.components.functions.
distributionfunctions.DRIFT_RATE],
M.parameter_ports[psyneulink.core.globals.keywords.THRESHOLD]
]
}])
G = pnl.GatingMechanism(gating_signals=[{
'DDM_OUTPUT_GATE': [
M.output_ports[pnl.DECISION_VARIABLE], M.output_ports[
pnl.RESPONSE_TIME]
]
}])
assert len(C.control_signals) == 1
assert C.control_signals[0].name == 'DECISION_CONTROL'
assert len(C.control_signals[0].efferents) == 2
assert M.parameter_ports[
psyneulink.core.components.functions.distributionfunctions.
DRIFT_RATE].mod_afferents[0] == C.control_signals[0].efferents[0]
assert M.parameter_ports[
psyneulink.core.globals.keywords.
THRESHOLD].mod_afferents[0] == C.control_signals[0].efferents[1]
assert len(G.gating_signals) == 1
assert G.gating_signals[0].name == 'DDM_OUTPUT_GATE'
assert len(G.gating_signals[0].efferents) == 2
assert M.output_ports[pnl.DECISION_VARIABLE].mod_afferents[
0] == G.gating_signals[0].efferents[0]
assert M.output_ports[pnl.RESPONSE_TIME].mod_afferents[
0] == G.gating_signals[0].efferents[1]
def test_multiple_modulatory_projections_with_mech_and_port_Name_specs(
self):
M = pnl.DDM(name='MY DDM')
C = pnl.ControlMechanism(control_signals=[{
pnl.MECHANISM:
M,
pnl.PARAMETER_PORTS: [
psyneulink.core.components.functions.distributionfunctions.
DRIFT_RATE, psyneulink.core.globals.keywords.THRESHOLD
]
}])
G = pnl.GatingMechanism(gating_signals=[{
pnl.MECHANISM:
M,
pnl.OUTPUT_PORTS: [pnl.DECISION_VARIABLE, pnl.RESPONSE_TIME]
}])
assert len(C.control_signals) == 1
assert len(C.control_signals[0].efferents) == 2
assert M.parameter_ports[
psyneulink.core.components.functions.distributionfunctions.
DRIFT_RATE].mod_afferents[0] == C.control_signals[0].efferents[0]
assert M.parameter_ports[
psyneulink.core.globals.keywords.
THRESHOLD].mod_afferents[0] == C.control_signals[0].efferents[1]
assert len(G.gating_signals) == 1
assert len(G.gating_signals[0].efferents) == 2
assert M.output_ports[pnl.DECISION_VARIABLE].mod_afferents[
0] == G.gating_signals[0].efferents[0]
assert M.output_ports[pnl.RESPONSE_TIME].mod_afferents[
0] == G.gating_signals[0].efferents[1]
def test_mapping_projection_with_mech_and_port_Name_specs(self):
R1 = pnl.TransferMechanism(output_ports=['OUTPUT_1', 'OUTPUT_2'])
R2 = pnl.TransferMechanism(default_variable=[[0], [0]],
input_ports=['INPUT_1', 'INPUT_2'])
T = pnl.TransferMechanism(input_ports=[{
pnl.MECHANISM:
R1,
pnl.OUTPUT_PORTS: ['OUTPUT_1', 'OUTPUT_2']
}],
output_ports=[{
pnl.MECHANISM:
R2,
pnl.INPUT_PORTS: ['INPUT_1', 'INPUT_2']
}])
assert len(R1.output_ports) == 2
assert len(R2.input_ports) == 2
assert len(T.input_ports) == 1
for input_port in T.input_ports:
for projection in input_port.path_afferents:
assert projection.sender.owner is R1
assert len(T.output_ports) == 1
for output_port in T.output_ports:
for projection in output_port.efferents:
assert projection.receiver.owner is R2
def test_mapping_projection_using_2_item_tuple_with_list_of_port_Names(
self):
T1 = pnl.TransferMechanism(name='T1', input_ports=[[0, 0], [0, 0, 0]])
T2 = pnl.TransferMechanism(name='T2',
output_ports=[
(['InputPort-0', 'InputPort-1'], T1)
])
assert len(T2.output_ports) == 1
assert T2.output_ports[0].efferents[0].receiver.name == 'InputPort-0'
assert T2.output_ports[0].efferents[0].matrix.base.shape == (1, 2)
assert T2.output_ports[0].efferents[1].receiver.name == 'InputPort-1'
assert T2.output_ports[0].efferents[1].matrix.base.shape == (1, 3)
def test_mapping_projection_using_2_item_tuple_and_3_item_tuples_with_index_specs(
self):
T1 = pnl.TransferMechanism(name='T1', input_ports=[[0, 0], [0, 0, 0]])
T2 = pnl.TransferMechanism(name='T2',
input_ports=['a', 'b', 'c'],
output_ports=[
(['InputPort-0', 'InputPort-1'], T1),
('InputPort-0', (pnl.OWNER_VALUE, 2),
T1),
(['InputPort-0', 'InputPort-1'], 1, T1)
])
assert len(T2.output_ports) == 3
assert T2.output_ports[0].efferents[0].receiver.name == 'InputPort-0'
assert T2.output_ports[0].efferents[0].matrix.base.shape == (1, 2)
assert T2.output_ports[0].efferents[1].receiver.name == 'InputPort-1'
assert T2.output_ports[0].efferents[1].matrix.base.shape == (1, 3)
assert T2.output_ports[1].owner_value_index == 2
assert T2.output_ports[2].owner_value_index == 1
def test_2_item_tuple_from_control_signal_to_parameter_port(self):
D = pnl.DDM(name='D')
# Single name
C = pnl.ControlMechanism(
control_signals=[(psyneulink.core.components.functions.
distributionfunctions.DRIFT_RATE, D)])
assert C.control_signals[0].name == 'D[drift_rate] ControlSignal'
assert C.control_signals[0].efferents[0].receiver.name == 'drift_rate'
# List of names
C = pnl.ControlMechanism(control_signals=[([
psyneulink.core.components.functions.distributionfunctions.
DRIFT_RATE, psyneulink.core.globals.keywords.THRESHOLD
], D)])
assert C.control_signals[
0].name == 'D[drift_rate, threshold] ControlSignal'
assert C.control_signals[0].efferents[0].receiver.name == 'drift_rate'
assert C.control_signals[0].efferents[1].receiver.name == 'threshold'
def test_2_item_tuple_from_parameter_port_to_control_signals(self):
C = pnl.ControlMechanism(control_signals=['a', 'b'])
D = pnl.DDM(name='D3',
function=psyneulink.core.components.functions.
distributionfunctions.DriftDiffusionAnalytical(
drift_rate=(3, C),
threshold=(2, C.control_signals['b'])))
assert D.parameter_ports[
psyneulink.core.components.functions.distributionfunctions.
DRIFT_RATE].mod_afferents[0].sender == C.control_signals[0]
assert D.parameter_ports[
psyneulink.core.globals.keywords.
THRESHOLD].mod_afferents[0].sender == C.control_signals[1]
def test_2_item_tuple_from_gating_signal_to_output_ports(self):
D4 = pnl.DDM(name='D4')
# Single name
G = pnl.GatingMechanism(gating_signals=[(pnl.DECISION_VARIABLE, D4)])
assert G.gating_signals[0].name == 'D4[DECISION_VARIABLE] GatingSignal'
assert G.gating_signals[0].efferents[
0].receiver.name == 'DECISION_VARIABLE'
# List of names
G = pnl.GatingMechanism(
gating_signals=[([pnl.DECISION_VARIABLE, pnl.RESPONSE_TIME], D4)])
assert G.gating_signals[
0].name == 'D4[DECISION_VARIABLE, RESPONSE_TIME] GatingSignal'
assert G.gating_signals[0].efferents[
0].receiver.name == 'DECISION_VARIABLE'
assert G.gating_signals[0].efferents[
1].receiver.name == 'RESPONSE_TIME'
def test_2_item_tuple_from_input_and_output_ports_to_gating_signals(self):
G = pnl.GatingMechanism(gating_signals=['a', 'b'])
T = pnl.TransferMechanism(name='T',
input_ports=[(3, G)],
output_ports=[(2, G.gating_signals['b'])])
assert T.input_ports[0].mod_afferents[0].sender == G.gating_signals[0]
assert T.output_ports[0].mod_afferents[0].sender == G.gating_signals[1]
control_spec_list = [
pnl.CONTROL, pnl.CONTROL_SIGNAL, pnl.CONTROL_PROJECTION,
pnl.ControlSignal,
pnl.ControlSignal(), pnl.ControlProjection, "CP_OBJECT",
pnl.ControlMechanism,
pnl.ControlMechanism(), pnl.ControlMechanism, (0.3, pnl.CONTROL),
(0.3, pnl.CONTROL_SIGNAL), (0.3, pnl.CONTROL_PROJECTION),
(0.3, pnl.ControlSignal), (0.3, pnl.ControlSignal()),
(0.3, pnl.ControlProjection), (0.3, "CP_OBJECT"),
(0.3, pnl.ControlMechanism), (0.3, pnl.ControlMechanism()),
(0.3, pnl.ControlMechanism)
]
@pytest.mark.parametrize(
'noise, gain',
[(noise, gain) for noise, gain in
[j for j in zip(control_spec_list, reversed(control_spec_list))]])
def test_formats_for_control_specification_for_mechanism_and_function_params(
self, noise, gain):
# This shenanigans is to avoid assigning the same instantiated ControlProjection more than once
if noise == 'CP_OBJECT':
noise = pnl.ControlProjection()
elif isinstance(noise, tuple) and noise[1] == 'CP_OBJECT':
noise = (noise[0], pnl.ControlProjection())
if gain == 'CP_OBJECT':
gain = pnl.ControlProjection()
elif isinstance(gain, tuple) and gain[1] == 'CP_OBJECT':
gain = (gain[0], pnl.ControlProjection())
R = pnl.RecurrentTransferMechanism(
# NOTE: fixed name prevents failures due to registry naming
# for parallel test runs
name='R-CONTROL',
noise=noise,
function=psyneulink.core.components.functions.transferfunctions.
Logistic(gain=gain))
assert R.parameter_ports[pnl.NOISE].mod_afferents[0].name in \
'ControlProjection for R-CONTROL[noise]'
assert R.parameter_ports[pnl.GAIN].mod_afferents[0].name in \
'ControlProjection for R-CONTROL[gain]'
gating_spec_list = [
pnl.GATING, pnl.CONTROL, pnl.GATING_SIGNAL, pnl.CONTROL_SIGNAL,
pnl.GATING_PROJECTION, pnl.CONTROL_PROJECTION, pnl.GatingSignal,
pnl.ControlSignal,
pnl.GatingSignal(),
pnl.ControlSignal(), pnl.GatingProjection, "GP_OBJECT",
pnl.GatingMechanism, pnl.ControlMechanism,
pnl.GatingMechanism(),
pnl.ControlMechanism(), (0.3, pnl.GATING), (0.3, pnl.CONTROL),
(0.3, pnl.GATING_SIGNAL), (0.3, pnl.CONTROL_SIGNAL),
(0.3, pnl.GATING_PROJECTION), (0.3, pnl.CONTROL_PROJECTION),
(0.3, pnl.GatingSignal), (0.3, pnl.ControlSignal),
(0.3, pnl.GatingSignal()), (0.3, pnl.ControlSignal()),
(0.3, pnl.GatingProjection), (0.3, pnl.ControlProjection),
(0.3, "GP_OBJECT"), (0.3, pnl.GatingMechanism),
(0.3, pnl.ControlMechanism), (0.3, pnl.GatingMechanism()),
(0.3, pnl.ControlMechanism())
]
@pytest.mark.parametrize(
'input_port, output_port',
[(inp, outp) for inp, outp in
[j for j in zip(gating_spec_list, reversed(gating_spec_list))]])
def test_formats_for_gating_specification_of_input_and_output_ports(
self, input_port, output_port):
G_IN, G_OUT = input_port, output_port
# This shenanigans is to avoid assigning the same instantiated ControlProjection more than once
if G_IN == 'GP_OBJECT':
G_IN = pnl.GatingProjection()
elif isinstance(G_IN, tuple) and G_IN[1] == 'GP_OBJECT':
G_IN = (G_IN[0], pnl.GatingProjection())
if G_OUT == 'GP_OBJECT':
G_OUT = pnl.GatingProjection()
elif isinstance(G_OUT, tuple) and G_OUT[1] == 'GP_OBJECT':
G_OUT = (G_OUT[0], pnl.GatingProjection())
if isinstance(G_IN, tuple):
IN_NAME = G_IN[1]
else:
IN_NAME = G_IN
IN_CONTROL = pnl.CONTROL in repr(IN_NAME).split(".")[-1].upper()
if isinstance(G_OUT, tuple):
OUT_NAME = G_OUT[1]
else:
OUT_NAME = G_OUT
OUT_CONTROL = pnl.CONTROL in repr(OUT_NAME).split(".")[-1].upper()
T = pnl.TransferMechanism(name='T-GATING',
input_ports=[G_IN],
output_ports=[G_OUT])
if IN_CONTROL:
assert T.input_ports[0].mod_afferents[0].name in \
'ControlProjection for T-GATING[InputPort-0]'
else:
assert T.input_ports[0].mod_afferents[0].name in \
'GatingProjection for T-GATING[InputPort-0]'
if OUT_CONTROL:
assert T.output_ports[0].mod_afferents[0].name in \
'ControlProjection for T-GATING[OutputPort-0]'
else:
assert T.output_ports[0].mod_afferents[0].name in \
'GatingProjection for T-GATING[OutputPort-0]'
# with pytest.raises(pnl.ProjectionError) as error_text:
# T1 = pnl.ProcessingMechanism(name='T1', input_ports=[pnl.ControlMechanism()])
# assert 'Primary OutputPort of ControlMechanism-0 (ControlSignal-0) ' \
# 'cannot be used as a sender of a Projection to InputPort of T1' in error_text.value.args[0]
#
# with pytest.raises(pnl.ProjectionError) as error_text:
# T2 = pnl.ProcessingMechanism(name='T2', output_ports=[pnl.ControlMechanism()])
# assert 'Primary OutputPort of ControlMechanism-1 (ControlSignal-0) ' \
# 'cannot be used as a sender of a Projection to OutputPort of T2' in error_text.value.args[0]
def test_no_warning_when_matrix_specified(self):
with pytest.warns(None) as w:
c = pnl.Composition()
m0 = pnl.ProcessingMechanism(default_variable=[0, 0, 0, 0])
p0 = pnl.MappingProjection(matrix=[[0, 0, 0, 0], [0, 0, 0, 0],
[0, 0, 0, 0], [0, 0, 0, 0]])
m1 = pnl.TransferMechanism(default_variable=[0, 0, 0, 0])
c.add_linear_processing_pathway([m0, p0, m1])
for warn in w:
if r'elementwise comparison failed; returning scalar instead' in warn.message.args[
0]:
raise
# KDM: this is a good candidate for pytest.parametrize
def test_masked_mapping_projection(self):
t1 = pnl.TransferMechanism(size=2)
t2 = pnl.TransferMechanism(size=2)
proj = pnl.MaskedMappingProjection(sender=t1,
receiver=t2,
matrix=[[1, 2], [3, 4]],
mask=[[1, 0], [0, 1]],
mask_operation=pnl.ADD)
c = pnl.Composition(pathways=[[t1, proj, t2]])
val = c.execute(inputs={t1: [1, 2]})
assert np.allclose(val, [[8, 12]])
t1 = pnl.TransferMechanism(size=2)
t2 = pnl.TransferMechanism(size=2)
proj = pnl.MaskedMappingProjection(sender=t1,
receiver=t2,
matrix=[[1, 2], [3, 4]],
mask=[[1, 0], [0, 1]],
mask_operation=pnl.MULTIPLY)
c = pnl.Composition(pathways=[[t1, proj, t2]])
val = c.execute(inputs={t1: [1, 2]})
assert np.allclose(val, [[1, 8]])
t1 = pnl.TransferMechanism(size=2)
t2 = pnl.TransferMechanism(size=2)
proj = pnl.MaskedMappingProjection(sender=t1,
receiver=t2,
mask=[[1, 2], [3, 4]],
mask_operation=pnl.MULTIPLY)
c = pnl.Composition(pathways=[[t1, proj, t2]])
val = c.execute(inputs={t1: [1, 2]})
assert np.allclose(val, [[1, 8]])
def test_masked_mapping_projection_mask_conficts_with_matrix(self):
with pytest.raises(pnl.MaskedMappingProjectionError) as error_text:
t1 = pnl.TransferMechanism(size=2)
t2 = pnl.TransferMechanism(size=2)
pnl.MaskedMappingProjection(sender=t1,
receiver=t2,
mask=[[1, 2, 3], [4, 5, 6]],
mask_operation=pnl.MULTIPLY)
assert "Shape of the 'mask'" in str(error_text.value)
assert "((2, 3)) must be the same as its 'matrix' ((2, 2))" in str(
error_text.value)
# FIX 7/22/15 [JDC] - REPLACE WITH MORE ELABORATE TESTS OF DUPLICATE PROJECTIONS:
# SAME FROM OutputPort; SAME TO InputPort
# TEST ERROR MESSAGES GENERATED BY VARIOUS _check_for_duplicates METHODS
# def test_duplicate_projection_detection_and_warning(self):
#
# with pytest.warns(UserWarning) as record:
# T1 = pnl.TransferMechanism(name='T1')
# T2 = pnl.TransferMechanism(name='T2')
# T3 = pnl.TransferMechanism(name='T3')
# T4 = pnl.TransferMechanism(name='T4')
#
# MP1 = pnl.MappingProjection(sender=T1,receiver=T2,name='MP1')
# MP2 = pnl.MappingProjection(sender=T1,receiver=T2,name='MP2')
# pnl.proc(T1,MP1,T2,T3)
# pnl.proc(T1,MP2,T2,T4)
#
# # hack to find a specific warning (other warnings may be generated by the Process construction)
# correct_message_found = False
# for warning in record:
# if "that already has an identical Projection" in str(warning.message):
# correct_message_found = True
# break
#
# assert len(T2.afferents)==1
# assert correct_message_found
def test_duplicate_projection_creation_error(self):
from psyneulink.core.components.projections.projection import DuplicateProjectionError
with pytest.raises(DuplicateProjectionError) as record:
T1 = pnl.TransferMechanism(name='T1')
T2 = pnl.TransferMechanism(name='T2')
pnl.MappingProjection(sender=T1, receiver=T2, name='MP1')
pnl.MappingProjection(sender=T1, receiver=T2, name='MP2')
assert 'Attempt to assign Projection to InputPort-0 of T2 that already has an identical Projection.' \
in record.value.args[0]
name='Hidden Layer_2',
default_variable=[0, 0, 0, 0],
function=psyneulink.core.components.functions.nonstateful.transferfunctions.Logistic()
)
Output_Layer = pnl.TransferMechanism(
name='Output Layer',
default_variable=[0, 0, 0],
function=psyneulink.core.components.functions.nonstateful.transferfunctions.Logistic
)
Gating_Mechanism = pnl.GatingMechanism(
# default_gating_allocation=0.0,
size=[1],
gating_signals=[
Hidden_Layer_1,
Hidden_Layer_2,
Output_Layer,
]
)
Input_Weights_matrix = (np.arange(2 * 5).reshape((2, 5)) + 1) / (2 * 5)
Middle_Weights_matrix = (np.arange(5 * 4).reshape((5, 4)) + 1) / (5 * 4)
Output_Weights_matrix = (np.arange(4 * 3).reshape((4, 3)) + 1) / (4 * 3)
# TEST PROCESS.LEARNING WITH:
# CREATION OF FREE STANDING PROJECTIONS THAT HAVE NO LEARNING (Input_Weights, Middle_Weights and Output_Weights)
# INLINE CREATION OF PROJECTIONS (Input_Weights, Middle_Weights and Output_Weights)
# NO EXPLICIT CREATION OF PROJECTIONS (Input_Weights, Middle_Weights and
# Output_Weights)
def test_gating_with_UDF():
def my_linear_fct(x,
m=2.0,
b=0.0,
params={
pnl.ADDITIVE_PARAM: 'b',
pnl.MULTIPLICATIVE_PARAM: 'm'
}):
return m * x + b
def my_simple_linear_fct(x, m=1.0, b=0.0):
return m * x + b
def my_exp_fct(
x,
r=1.0,
# b=pnl.CONTROL,
b=0.0,
params={
pnl.ADDITIVE_PARAM: 'b',
pnl.MULTIPLICATIVE_PARAM: 'r'
}):
return x**r + b
def my_sinusoidal_fct(input,
phase=0,
amplitude=1,
params={
pnl.ADDITIVE_PARAM: 'phase',
pnl.MULTIPLICATIVE_PARAM: 'amplitude'
}):
frequency = input[0]
t = input[1]
return amplitude * np.sin(2 * np.pi * frequency * t + phase)
Input_Layer = pnl.TransferMechanism(name='Input_Layer',
default_variable=np.zeros((2, )),
function=psyneulink.core.components.
functions.transferfunctions.Logistic)
Output_Layer = pnl.TransferMechanism(
name='Output_Layer',
default_variable=[0, 0, 0],
function=psyneulink.core.components.functions.transferfunctions.Linear,
# function=pnl.Logistic,
# output_states={pnl.NAME: 'RESULTS USING UDF',
# pnl.VARIABLE: [(pnl.OWNER_VALUE,0), pnl.TIME_STEP],
# pnl.FUNCTION: my_sinusoidal_fct}
output_states={
pnl.NAME:
'RESULTS USING UDF',
# pnl.VARIABLE: (pnl.OWNER_VALUE, 0),
pnl.FUNCTION:
psyneulink.core.components.functions.transferfunctions.Linear(
slope=pnl.GATING)
# pnl.FUNCTION: pnl.Logistic(gain=pnl.GATING)
# pnl.FUNCTION: my_linear_fct
# pnl.FUNCTION: my_exp_fct
# pnl.FUNCTION:pnl.UserDefinedFunction(custom_function=my_simple_linear_fct,
# params={pnl.ADDITIVE_PARAM:'b',
# pnl.MULTIPLICATIVE_PARAM:'m',
# },
# m=pnl.GATING,
# b=2.0
# )
})
Gating_Mechanism = pnl.GatingMechanism(
# default_gating_allocation=0.0,
size=[1],
gating_signals=[
# Output_Layer
Output_Layer.output_state,
])
p = pnl.Process(size=2,
pathway=[Input_Layer, Output_Layer],
prefs={
pnl.VERBOSE_PREF: False,
pnl.REPORT_OUTPUT_PREF: False
})
g = pnl.Process(default_variable=[1.0], pathway=[Gating_Mechanism])
stim_list = {
Input_Layer: [[-1, 30], [-1, 30], [-1, 30], [-1, 30]],
Gating_Mechanism: [[0.0], [0.5], [1.0], [2.0]]
}
mySystem = pnl.System(processes=[p, g])
mySystem.reportOutputPref = False
results = mySystem.run(
num_trials=4,
inputs=stim_list,
)
expected_results = [[np.array([0., 0., 0.])],
[np.array([0.63447071, 0.63447071, 0.63447071])],
[np.array([1.26894142, 1.26894142, 1.26894142])],
[np.array([2.53788284, 2.53788284, 2.53788284])]]
np.testing.assert_allclose(results, expected_results)
# pnl.FUNCTION: pnl.Logistic(gain=pnl.GATING)
# pnl.FUNCTION: my_linear_fct
# pnl.FUNCTION: my_exp_fct
# pnl.FUNCTION:pnl.UserDefinedFunction(custom_function=my_simple_linear_fct,
# params={pnl.ADDITIVE_PARAM:'b',
# pnl.MULTIPLICATIVE_PARAM:'m',
# },
# m=pnl.GATING,
# b=2.0
# )
})
Gating_Mechanism = pnl.GatingMechanism(
# default_gating_allocation=0.0,
size=[1],
gating_signals=[
# Output_Layer
Output_Layer.output_port,
])
def print_header(system):
print("\n\n**** Time: ", system.scheduler.get_clock(system).simple_time)
def show_target(context=None):
print('Gated: ',
Gating_Mechanism.gating_signals[0].efferents[0].receiver.owner.name,
Gating_Mechanism.gating_signals[0].efferents[0].receiver.name)
print('- Input_Layer.value: ',
Input_Layer.parameters.value.get(context))
def test_gating_with_UDF_with_composition():
def my_linear_fct(x,
m=2.0,
b=0.0,
params={
pnl.ADDITIVE_PARAM: 'b',
pnl.MULTIPLICATIVE_PARAM: 'm'
}):
return m * x + b
def my_simple_linear_fct(x, m=1.0, b=0.0):
return m * x + b
def my_exp_fct(
x,
r=1.0,
# b=pnl.CONTROL,
b=0.0,
params={
pnl.ADDITIVE_PARAM: 'b',
pnl.MULTIPLICATIVE_PARAM: 'r'
}):
return x**r + b
def my_sinusoidal_fct(input,
phase=0,
amplitude=1,
params={
pnl.ADDITIVE_PARAM: 'phase',
pnl.MULTIPLICATIVE_PARAM: 'amplitude'
}):
frequency = input[0]
t = input[1]
return amplitude * np.sin(2 * np.pi * frequency * t + phase)
Input_Layer = pnl.TransferMechanism(
name='Input_Layer',
default_variable=np.zeros((2, )),
function=psyneulink.core.components.functions.nonstateful.
transferfunctions.Logistic)
Output_Layer = pnl.TransferMechanism(
name='Output_Layer',
default_variable=[0, 0, 0],
function=psyneulink.core.components.functions.nonstateful.
transferfunctions.Linear,
# function=pnl.Logistic,
# output_ports={pnl.NAME: 'RESULTS USING UDF',
# pnl.VARIABLE: [(pnl.OWNER_VALUE,0), pnl.TIME_STEP],
# pnl.FUNCTION: my_sinusoidal_fct}
output_ports={
pnl.NAME:
'RESULTS USING UDF',
# pnl.VARIABLE: (pnl.OWNER_VALUE, 0),
pnl.FUNCTION:
psyneulink.core.components.functions.nonstateful.transferfunctions.
Linear(slope=pnl.GATING)
})
Gating_Mechanism = pnl.GatingMechanism(
size=[1],
gating_signals=[
# Output_Layer
Output_Layer.output_port,
])
comp = Composition()
comp.add_linear_processing_pathway(pathway=[Input_Layer, Output_Layer])
comp.add_node(Gating_Mechanism)
stim_list = {
Input_Layer: [[-1, 30], [-1, 30], [-1, 30], [-1, 30]],
Gating_Mechanism: [[0.0], [0.5], [1.0], [2.0]]
}
comp.run(num_trials=4, inputs=stim_list)
expected_results = [[np.array([0., 0., 0.])],
[np.array([0.63447071, 0.63447071, 0.63447071])],
[np.array([1.26894142, 1.26894142, 1.26894142])],
[np.array([2.53788284, 2.53788284, 2.53788284])]]
np.testing.assert_allclose(comp.results, expected_results)
# pnl.FUNCTION: pnl.Logistic(gain=pnl.GATING)
# pnl.FUNCTION: my_linear_fct
# pnl.FUNCTION: my_exp_fct
# pnl.FUNCTION:pnl.UserDefinedFunction(custom_function=my_simple_linear_fct,
# params={pnl.ADDITIVE_PARAM:'b',
# pnl.MULTIPLICATIVE_PARAM:'m',
# },
# m=pnl.GATING,
# b=2.0
# )
})
Gating_Mechanism = pnl.GatingMechanism(
# default_gating_policy=0.0,
size=[1],
gating_signals=[
# Output_Layer
Output_Layer.output_state,
])
p = pnl.Process(size=2,
pathway=[Input_Layer, Output_Layer],
prefs={
pnl.VERBOSE_PREF: False,
pnl.REPORT_OUTPUT_PREF: False
})
g = pnl.Process(default_variable=[1.0], pathway=[Gating_Mechanism])
stim_list = {
Input_Layer: [[-1, 30], [-1, 30], [-1, 30], [-1, 30]],