Merge pull request #6 from openworm/development

Added initial Cook reader (herm only, no muscles)

.gitignore

@@ -95,3 +95,13 @@ arm64
 /examples/test/images/c302_C1_Oscillator.net.nml
 /examples/test/images/LEMS_c302_C1_Oscillator.xml
 /c302/VarshneyDataReader2.py
+/
+/examples/test/images/LEMS_c302_C1_Full.xml
+/examples/test/images/LEMS_c302_C1_Muscles.xml
+/examples/test/images/LEMS_c302_C1_Social.xml
+/examples/test/images/LEMS_c302_C1_Syns.xml
+/examples/test/images/c302_C1_Full.net.nml
+/examples/test/images/c302_C1_Muscles.net.nml
+/examples/test/images/c302_C1_Pharyngeal.net.nml
+/examples/test/images/c302_C1_Social.net.nml
+/examples/test/images/c302_C1_Syns.net.nml

c302/ConnectomeReader.py

@@ -28,11 +28,11 @@ def convert_to_preferred_muscle_name(muscle):
         return muscle + "???"
 
 def get_all_muscle_prefixes():
-    return ["pm", "vm", "um", "BWM-D", "BWM-V", "LegacyBodyWallMuscles"]
+    return ["pm", "vm", "um", "BWM-D", "BWM-V", "LegacyBodyWallMuscles", "vBWM", "dBWM"]
 
 
 def get_body_wall_muscle_prefixes():
-    return ["BWM-D", "BWM-V", "LegacyBodyWallMuscles"]
+    return ["BWM-D", "BWM-V", "LegacyBodyWallMuscles", "vBWM", "dBWM"]
 
 
 def is_muscle(cell):
@@ -48,6 +48,15 @@ def is_body_wall_muscle(cell):
 def is_neuron(cell):
     return not is_body_wall_muscle(cell)
 
+
+def remove_leading_index_zero(cell):
+    """
+    Returns neuron name with an index without leading zero. E.g. VB01 -> VB1.
+    """
+    if is_neuron(cell) and cell[-2:].startswith("0"):
+        return "%s%s" % (cell[:-2], cell[-1:])
+    return cell
+
 class ConnectionInfo:
 
     def __init__(self,

c302/Cook2019DataReader.py

@@ -0,0 +1,188 @@
+# -*- coding: utf-8 -*-
+
+############################################################
+
+#    A simple script to read the values in herm_full_edgelist.csv.
+
+#    This is on of a number of interchangeable "Readers" which can 
+#    be used to get connection data for c302
+
+############################################################
+
+import csv
+
+from c302.ConnectomeReader import ConnectionInfo
+from c302.ConnectomeReader import analyse_connections
+from c302.ConnectomeReader import convert_to_preferred_muscle_name
+from c302.ConnectomeReader import is_neuron
+from c302.ConnectomeReader import is_body_wall_muscle
+from c302.ConnectomeReader import remove_leading_index_zero
+
+from openpyxl import load_workbook
+
+import os
+import numpy as np
+
+from c302 import print_
+
+spreadsheet_location = os.path.dirname(os.path.abspath(__file__))+"/data/"
+filename = "%sSI 5 Connectome adjacency matrices.xlsx" % spreadsheet_location
+
+HERM_CHEM = 'hermaphrodite chemical'
+HERM_GAP_SYMM = 'herm gap jn symmetric'
+
+pre_range = {HERM_CHEM:range(4,304),
+             HERM_GAP_SYMM:range(4,472)}
+
+post_range = {HERM_CHEM:range(4,457),
+             HERM_GAP_SYMM:range(4,472)}
+
+
+def get_synclass(cell, syntype):
+    #TODO: fix this dirty hack
+    if syntype == "GapJunction":
+        return "Generic_GJ"
+    else:
+        if cell.startswith("DD") or cell.startswith("VD"):
+            return "GABA"
+        return "Acetylcholine"
+
+class Cook2019DataReader():
+
+    verbose = False
+
+    def __init__(self):
+
+        wb = load_workbook(filename)
+        print_("Opened the Excel file: " + filename)   
+
+        self.pre_cells = {}
+        self.post_cells = {}
+        self.conn_nums = {}
+
+        for conn_type in [HERM_CHEM, HERM_GAP_SYMM]:
+
+            sheet = wb.get_sheet_by_name(conn_type)
+            print('Looking at sheet: %s'%conn_type)
+
+            self.pre_cells[conn_type] = []
+            self.post_cells[conn_type] = []
+
+            for i in pre_range[conn_type]:
+                self.pre_cells[conn_type].append(sheet['C%i'%i].value)
+
+            if self.verbose: print(' - Pre cells for %s (%i):\n%s'%(conn_type, len(self.pre_cells[conn_type]), self.pre_cells[conn_type]))
+
+            for i in post_range[conn_type]:
+                self.post_cells[conn_type].append(sheet.cell(row=3, column=i).value)
+
+            if self.verbose: print(' - Post cells for %s (%i):\n%s'%(conn_type, len(self.post_cells[conn_type]), self.post_cells[conn_type]))
+
+            self.conn_nums[conn_type] = np.zeros([len(self.pre_cells[conn_type]),len(self.post_cells[conn_type])], dtype=int)
+
+            for i in range(len(self.pre_cells[conn_type])):
+                for j in range(len(self.post_cells[conn_type])):
+                    row =4+i
+                    col = 4+j
+                    val = sheet.cell(row=row, column=col).value
+                    #print('Cell (%i,%i) [row %i, col %i] = %s'%(i,j,row, col, val))
+                    if val is not None:
+                        self.conn_nums[conn_type][i,j] = int(val)
+
+            if self.verbose: print(' - Conns for %s (%s):\n%s'%(conn_type, self.conn_nums[conn_type].shape, self.conn_nums[conn_type]))
+
+
+
+    def read_data(self, include_nonconnected_cells=False):
+        """
+        Args:
+            include_nonconnected_cells (bool): Also append neurons without known connections to other neurons to the 'cells' list. True if they should get appended, False otherwise.
+        Returns:
+            cells (:obj:`list` of :obj:`str`): List of neurons
+            conns (:obj:`list` of :obj:`ConnectionInfo`): List of connections from neuron to neuron
+        """
+
+        if not include_nonconnected_cells:
+            raise Exception('Option include_nonconnected_cells=False not supported')
+
+        conns = []
+        cells = []
+        for conn_type in [HERM_CHEM, HERM_GAP_SYMM]:
+
+            for pre_index in range(len(self.pre_cells[conn_type])):
+                for post_index in range(len(self.post_cells[conn_type])):
+
+                    pre = remove_leading_index_zero(self.pre_cells[conn_type][pre_index])
+                    post = remove_leading_index_zero(self.post_cells[conn_type][post_index])
+
+                    if is_body_wall_muscle(post):
+                        continue  # post is a BWM so ignore
+
+                    num = self.conn_nums[conn_type][pre_index,post_index]
+                    if num>0:
+                        syntype = 'Send' if conn_type == HERM_CHEM else "GapJunction"
+                        synclass = get_synclass(pre, syntype)
+
+                        conns.append(ConnectionInfo(pre, post, num, syntype, synclass))
+                        #print ConnectionInfo(pre, post, num, syntype, synclass)
+                        if pre not in cells:
+                            cells.append(pre)
+                        if post not in cells:
+                            cells.append(post)
+
+        return cells, conns
+
+
+    def read_muscle_data(self):
+        """
+        Returns:
+            neurons (:obj:`list` of :obj:`str`): List of motor neurons. Each neuron has at least one connection with a post-synaptic muscle cell.
+            muscles (:obj:`list` of :obj:`str`): List of muscle cells.
+            conns (:obj:`list` of :obj:`ConnectionInfo`): List of neuron-muscle connections.
+        """
+
+        neurons = []
+        muscles = []
+        conns = []
+
+        '''
+        with open(filename, 'r') as f:
+            reader = csv.DictReader(f)
+            print_("Opened file: " + filename)
+
+            for row in reader:
+                pre, post, num, syntype, synclass = parse_row(row)
+
+                if not (is_neuron(pre) or is_body_wall_muscle(pre)) or not is_body_wall_muscle(post):
+                    continue
+
+                if is_neuron(pre):
+                    pre = remove_leading_index_zero(pre)
+                else:
+                    pre = get_old_muscle_name(pre)
+                post = get_old_muscle_name(post)
+
+                conns.append(ConnectionInfo(pre, post, num, syntype, synclass))
+                if is_neuron(pre) and pre not in neurons:
+                    neurons.append(pre)
+                elif is_body_wall_muscle(pre) and pre not in muscles:
+                    muscles.append(pre)
+                if post not in muscles:
+                    muscles.append(post)''' 
+
+        return neurons, muscles, conns
+
+
+def main():
+
+    cdr = Cook2019DataReader()
+    read_data = cdr.read_data
+    read_muscle_data = cdr.read_muscle_data
+
+    cells, neuron_conns = read_data(include_nonconnected_cells=True)
+    neurons2muscles, muscles, muscle_conns = read_muscle_data()
+
+    analyse_connections(cells, neuron_conns, neurons2muscles, muscles, muscle_conns)
+
+if __name__ == '__main__':
+    main()

c302/Cook2019HermReader.py

@@ -0,0 +1,19 @@
+# Temporary class to allow this to be used in comparison notebook. 
+# Should be tidied up.
+
+from c302.Cook2019DataReader import Cook2019DataReader
+
+from c302.ConnectomeReader import analyse_connections
+
+cdr = Cook2019DataReader()
+read_data = cdr.read_data
+read_muscle_data = cdr.read_muscle_data
+
+def main1():
+    cells, neuron_conns = read_data(include_nonconnected_cells=True)
+    neurons2muscles, muscles, muscle_conns = read_muscle_data()
+    analyse_connections(cells, neuron_conns, neurons2muscles, muscles, muscle_conns)
+
+
+if __name__ == '__main__':
+    main1()

c302/__init__.py

@@ -465,7 +465,7 @@ def _get_cell_info(bnd, cells):
     for name in fixed_up_names:
         cell = next(ctx(Cell).query(name=name).load(), None)
         if cell is None:
-            print_("No matching cell for %s" % name)
+            #print_("No matching cell for %s" % name)
             continue
 
         normalized_name = cell.name()

c302/c302_utils.py

@@ -577,6 +577,8 @@ def generate_conn_matrix(nml_doc, save_fig_dir=None, verbose=False, figsize=defa
     configs = ['c302_C0_Muscles.net.nml']
     configs = ['c302_C0_Syns.net.nml', 'c302_C0_Social.net.nml','c302_C0_Muscles.net.nml','c302_C0_Pharyngeal.net.nml','c302_C0_Oscillator.net.nml','c302_C0_Full.net.nml']
 
+    figsize=(6.4,4.8)
+
     if '-phar' in sys.argv:
 
         configs = ['c302_C0_Pharyngeal.net.nml']
@@ -585,11 +587,21 @@ def generate_conn_matrix(nml_doc, save_fig_dir=None, verbose=False, figsize=defa
 
         configs = ['c302_C1_Oscillator.net.nml']
 
+    elif '-musc' in sys.argv:
+
+        configs = ['c302_C1_Muscles.net.nml']
+        figsize=(10,10)
+
+    elif '-full' in sys.argv:
+
+        configs = ['c302_C1_Full.net.nml']
+        figsize=(10,10)
+
     for c in configs:
 
         nml_doc = read_neuroml2_file('examples/%s'%c)
 
-        generate_conn_matrix(nml_doc, save_fig_dir='./examples/summary/images', figsize=(6.4,4.8))
+        generate_conn_matrix(nml_doc, save_fig_dir='./examples/summary/images', figsize=figsize)
 
     if not '-nogui' in sys.argv:
         plt.show()