Merge pull request #2752 from clinssen/sli_tests_to_py_aeif_cond_beta_multisynapse

Port `aeif_cond_beta_multisynapse_test` from SLI-2-Py and add test for `..._alpha_...`

Author: terhorstd

testsuite/pytests/conftest.py

@@ -32,6 +32,7 @@ def test_gsl():
 """
 
 import dataclasses
+import os
 import pathlib
 import sys
 
@@ -79,6 +80,11 @@ def have_threads():
     return nest.ll_api.sli_func("is_threaded")
 
 
+@pytest.fixture(scope="session")
+def report_dir() -> pathlib.Path:
+    return pathlib.Path(os.environ.get("REPORTDIR", ""))
+
+
 @pytest.fixture(autouse=True)
 def skipif_missing_threads(request, have_threads):
     """

testsuite/pytests/test_aeif_cond_alpha_multisynapse.py

@@ -19,38 +19,38 @@
 # You should have received a copy of the GNU General Public License
 # along with NEST.  If not, see <http://www.gnu.org/licenses/>.
 
-"""
-This test creates a multisynapse neuron and first checks if time constants
-can be set correctly.
-
-Afterwards, it simulates the mutisynapse neuron with n (n=4) different time
-constants and records the neuron's synaptic current. At the same time, it simulates
-n (n=4) single synapse neurons with according parameters.
-At the end, it compares the multisynapse neuron currents with each according single
-synapse current.
-"""
-
 import nest
 import numpy as np
 import pytest
 
 
 @pytest.mark.skipif_missing_gsl
 class TestAeifCondAlphaMultisynapse:
-    def test_single_multi_synapse_equivalence(self, have_plotting):
-        simulation_t = 2500.0  # ms
+    r"""
+    This test creates a multisynapse neuron and first checks if time constants
+    can be set correctly.
+
+    Afterwards, it simulates the mutisynapse neuron with n (n=4) different time
+    constants and records the neuron's synaptic current. At the same time, it simulates
+    n (n=4) single synapse neurons with according parameters.
+    At the end, it compares the multisynapse neuron currents with each according single
+    synapse current.
+    """
 
-        dt = 0.1
+    def test_single_multi_synapse_equivalence(self, have_plotting, report_dir):
+        simulation_t = 2500.0  # total simulation time [ms]
 
-        E_ex = 0.0  # mV
-        E_in = -85.0  # mV
-        V_peak = 0.0
+        dt = 0.1  # time step [ms]
+
+        E_ex = 0.0  # excitatory reversal potential [mV]
+        E_in = -85.0  # inhibitory reversal potential [mV]
+        V_peak = 0.0  # spike detection threshold [mV]
         a = 4.0
         b = 80.5
-        tau_syn = [0.2, 0.5, 1.0, 10.0]
-        weight = [1.0, 5.0, 1.0, -1.0]
-        E_rev = [E_ex, E_ex, E_ex, E_in]
-        spike_time = 1.0
+        tau_syn = [0.2, 0.5, 1.0, 10.0]  # synaptic times [ms]
+        weight = [1.0, 5.0, 1.0, -1.0]  # synaptic weights
+        E_rev = [E_ex, E_ex, E_ex, E_in]  # synaptic reversal potentials [mV]
+        spike_time = 1.0  # time at which the single spike occurs [ms]
 
         # The delays have to be ordered and needs enough space between them to avoid one PSC from affecting the next
         delays = [1.0, 500.0, 1500.0, 2250.0]  # ms
@@ -147,12 +147,14 @@ def test_single_multi_synapse_equivalence(self, have_plotting):
                 _ax.legend()
 
             ax[-1].semilogy(multisynapse_neuron_vm.get("events")["times"], error, label="errror")
-            fig.savefig("test_aeif_cond_alpha_multisynapse.png")
+
+            fig.savefig(report_dir / "test_aeif_cond_alpha_multisynapse.png")
 
         # compare with a large tolerance because previous PSPs affect subsequent PSPs in the multisynapse neuron
         np.testing.assert_allclose(error, 0, atol=1e-6)
 
     def test_recordables(self):
+        r"""Test that the right number of recordables are created when setting ``record_from``."""
         nest.ResetKernel()
 
         nrn = nest.Create("aeif_cond_alpha_multisynapse")
@@ -166,7 +168,7 @@ def test_recordables(self):
         assert len(nrn.recordables) == 3
 
     def test_resize_recordables(self):
-        """Test that the recordable g's change when changing the number of receptor ports"""
+        r"""Test that the recordable g's change when changing the number of receptor ports"""
         nest.ResetKernel()
 
         E_rev1 = [0.0, 0.0, -85.0]
@@ -185,8 +187,8 @@ def test_resize_recordables(self):
         nrn.set({"E_rev": E_rev3, "tau_syn": tau_syn3})
         assert len(nrn.recordables) == 6
 
-    def test_g_alpha_dynamics(self, have_plotting):
-        """Test that g has alpha function dynamics"""
+    def test_g_alpha_dynamics(self, have_plotting, report_dir):
+        r"""Test that g has alpha function dynamics"""
 
         dt = 0.1  # time step
 
@@ -196,7 +198,7 @@ def test_g_alpha_dynamics(self, have_plotting):
         E_rev = [0.0, 0.0, -85.0, 20.0]  # synaptic reversal potentials
         tau_syn = [40.0, 20.0, 30.0, 25.0]  # synaptic time constants
         weight = [1.0, 0.5, 2.0, 1.0]  # synaptic weights
-        delays = [1.0, 3.0, 10.0, 10.0]  # ms - synaptic delays
+        delays = [1.0, 3.0, 10.0, 10.0]  # synaptic delays [ms]
         spike_time = 10.0  # time at which the single spike occurs
         total_t = 500.0  # total simulation time
 
@@ -258,6 +260,6 @@ def alpha_function(t, W=1.0, tau=1.0, t0=0.0):
                 for _ax in ax:
                     _ax.legend()
 
-                fig.savefig("test_aeif_cond_alpha_multisynapse_psc_shape_ " + str(i) + ".png")
+                fig.savefig(report_dir / f"test_aeif_cond_alpha_multisynapse_psc_shape_{i}.png")
 
             np.testing.assert_allclose(sim_g, theo_g)