Merge pull request #69 from NeuroML/experimental

Add slight delay in generate network when extra_lems_components present

examples/LEMS_SimExample3.xml

@@ -2,7 +2,7 @@
 
     <!-- 
 
-        This LEMS file has been automatically generated using PyNeuroML v1.2.3 (libNeuroML v0.5.8)
+        This LEMS file has been automatically generated using PyNeuroML v1.2.13 (libNeuroML v0.5.9)
 
      -->
 

examples/LEMS_SimExample4.xml

@@ -2,7 +2,7 @@
 
     <!-- 
 
-        This LEMS file has been automatically generated using PyNeuroML v1.2.3 (libNeuroML v0.5.8)
+        This LEMS file has been automatically generated using PyNeuroML v1.2.13 (libNeuroML v0.5.9)
 
      -->
 

examples/test_files/.test.ex4.jnmleden.omt

@@ -14,7 +14,7 @@ experiments:
         spike detection:
           method: threshold
           threshold: -64
-        tolerance: 0.0002527679474217274
+        tolerance: 0.00025278058645101804
 
   v1:
     observables:
@@ -26,4 +26,4 @@ experiments:
         spike detection:
           method: threshold
           threshold: -69
-        tolerance: 0.0002327287423609406
+        tolerance: 0.00023269200151868833

neuromllite/MatrixHandler.py

@@ -11,16 +11,20 @@
 from neuromllite.NetworkGenerator import _get_rng_for_network
 
 import numpy as np
+import os
 
 
 class MatrixHandler(ConnectivityHandler):
     colormaps_used = []
 
     weight_arrays_to_show = {}
+    weight_array_figures = {}
 
-    def __init__(self, level=10, nl_network=None):
+    def __init__(self, level=10, nl_network=None, show_already=True, save_figs_to_dir=None):
         self.nl_network = nl_network
         self.level = level
+        self.show_already = show_already
+        self.save_figs_to_dir=save_figs_to_dir
 
         self.rng, seed = _get_rng_for_network(self.nl_network)
 
@@ -34,11 +38,6 @@ def print_settings(self):
         print_v("*    level:                   %s" % self.level)
         print_v("*    is_cell_level:           %s" % self.is_cell_level())
         print_v("*    CUTOFF_INH_SYN_MV:       %s" % self.CUTOFF_INH_SYN_MV)
-        # print_v('*    include_inputs:          %s'%self.include_inputs)
-        # print_v('*    scale_by_post_pop_size:  %s'%self.scale_by_post_pop_size)
-        # print_v('*    scale_by_post_pop_cond:  %s'%self.scale_by_post_pop_cond)
-        # print_v('*    min_weight_to_show:      %s'%self.min_weight_to_show)
-        # print_v('*    min_weight_to_show:      %s'%self.min_weight_to_show)
         print_v("*")
         print_v("* Used values: ")
         print_v("*    colormaps_used:          %s" % self.colormaps_used)
@@ -68,6 +67,9 @@ def finalise_document(self):
         for v in self.proj_post_pops.values():
             all_pops.append(v)
 
+        self.weight_matrices_generated = {}
+        self.weight_matrices_saved = {}
+
         for pop in all_pops:
             if self.is_cell_level():
                 for i in range(self.pop_sizes[pop]):
@@ -376,10 +378,22 @@ def finalise_document(self):
                 if proj_type in cbar_labels:
                     cbar.set_label(cbar_labels[proj_type])
 
+                if self.save_figs_to_dir:
+                    safe = proj_type.replace(' ','_').replace('(','-').replace(')','-').replace('*','_').replace('/','_')
+                    save_figure_to = os.path.join(self.save_figs_to_dir, '%s_%s.png'%(self.network_id, safe))
+                    plt.savefig(save_figure_to, bbox_inches="tight")
+                    print_v("Saved image to %s of plot: %s" % (save_figure_to, title))
+                    self.weight_array_figures[title] = save_figure_to
+
+
+
         print_v("Generating matrix for: %s" % self.network_id)
+
+
         self.print_settings()
 
-        plt.show()
+        if self.show_already:
+            plt.show()
 
     def handle_population(
         self,
@@ -535,3 +549,31 @@ def finalise_projection(
 
     def finalise_input_source(self, inputListId):
         pass
+
+if __name__ == "__main__":
+
+    from neuromllite.utils import load_network_json
+
+    tests = ['/Users/padraig/neuroConstruct/osb/cerebral_cortex/networks/del-Molino2017/NeuroML/delMolinoEtAl_low_baseline.json',
+    '/Users/padraig/NeuroMLlite/examples/Example12_MultiComp.json']
+
+    for test in tests:
+
+        network = load_network_json(test)
+
+        from neuromllite.NetworkGenerator import generate_network
+
+        level = 1
+        handler = MatrixHandler(level, 
+                                nl_network=network, 
+                                show_already=True,
+                                save_figs_to_dir='.')
+
+        generate_network(
+            network, handler, always_include_props=True, base_dir=os.path.dirname(test)
+        )
+
+        print_v("Done with MatrixHandler...")
+
+        for w in handler.weight_arrays_to_show:
+            print('%s:\n%s'%(w, handler.weight_arrays_to_show[w]))

neuromllite/NetworkGenerator.py

@@ -11,6 +11,7 @@
 import numpy as np
 import os
 import random
+import time
 
 import lems.api as lems  # from pylems
 
@@ -913,6 +914,7 @@ def generate_neuroml2_from_network(
 
     if len(extra_lems_components.components) > 0:
         extra_lems_components.export_to_file(extra_lems_file)
+        time.sleep(0.5)
 
     if validate and format == "xml":
         from pyneuroml import pynml