Merge branch 'master' into pyup-pin-nbsphinx-0.3.3

docs/conf.py

@@ -91,9 +91,9 @@
 # built documents.
 #
 # The short X.Y version.
-version = '0.82'
+version = '0.84'
 # The full version, including alpha/beta/rc tags.
-release = '0.82'
+release = '0.84'
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.

docs/requirements.txt

@@ -1,4 +1,4 @@
 ###### Requirements without Version Specifiers ######
-numpydoc
 nbsphinx==0.3.3
+numpydoc==0.8.0
 ipykernel

pymaid/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.82"
+__version__ = "0.84"
 
 from pymaid import config
 

pymaid/core.py

@@ -280,6 +280,20 @@ def __init__(self, x, remote_instance=None, meta_data=None):
                 except BaseException:
                     pass
 
+    def __dir__(self):
+        """ Custom __dir__ to add some parameters that we want to make
+        searchable.
+        """
+        add_attributes = ['n_open_ends', 'n_branch_nodes', 'n_end_nodes',
+                          'cable_length', 'root', 'neuron_name',
+                          'nodes', 'annotations', 'partners', 'review_status',
+                          'connectors', 'presynapses', 'postsynapses',
+                          'gap_junctions', 'soma', 'root', 'tags',
+                          'n_presynapses', 'n_postsynapses', 'n_connectors',
+                          'bbox']
+
+        return list(set(super().__dir__() + add_attributes))
+
     def __getattr__(self, key):
         # This is to catch empty neurons (e.g. after pruning)
         if 'nodes' in self.__dict__ and \
@@ -335,6 +349,8 @@ def __getattr__(self, key):
         elif key == 'tags':
             self.get_skeleton()
             return self.tags
+        elif key == 'sampling_resolution':
+            return self.n_nodes / self.cable_length
         elif key == 'n_open_ends':
             if 'nodes' in self.__dict__:
                 closed = self.tags.get('ends', []) \
@@ -344,36 +360,43 @@ def __getattr__(self, key):
                     + self.tags.get('soma', [])
                 return len([n for n in self.nodes[self.nodes.type == 'end'].treenode_id.tolist() if n not in closed])
             else:
+                logger.info('No skeleton data available. Use .get_skeleton() to fetch.')
                 return 'NA'
         elif key == 'n_branch_nodes':
             if 'nodes' in self.__dict__:
                 return self.nodes[self.nodes.type == 'branch'].shape[0]
             else:
+                logger.info('No skeleton data available. Use .get_skeleton() to fetch.')
                 return 'NA'
         elif key == 'n_end_nodes':
             if 'nodes' in self.__dict__:
                 return self.nodes[self.nodes.type == 'end'].shape[0]
             else:
+                logger.info('No skeleton data available. Use .get_skeleton() to fetch.')
                 return 'NA'
         elif key == 'n_nodes':
             if 'nodes' in self.__dict__:
                 return self.nodes.shape[0]
             else:
+                logger.info('No skeleton data available. Use .get_skeleton() to fetch.')
                 return 'NA'
         elif key == 'n_connectors':
             if 'connectors' in self.__dict__:
                 return self.connectors.shape[0]
             else:
+                logger.info('No skeleton data available. Use .get_skeleton() to fetch.')
                 return 'NA'
         elif key == 'n_presynapses':
             if 'connectors' in self.__dict__:
                 return self.connectors[self.connectors.relation == 0].shape[0]
             else:
+                logger.info('No skeleton data available. Use .get_skeleton() to fetch.')
                 return 'NA'
         elif key == 'n_postsynapses':
             if 'connectors' in self.__dict__:
                 return self.connectors[self.connectors.relation == 1].shape[0]
             else:
+                logger.info('No skeleton data available. Use .get_skeleton() to fetch.')
                 return 'NA'
         elif key == 'cable_length':
             if 'nodes' in self.__dict__:
@@ -384,6 +407,13 @@ def __getattr__(self, key):
                     w = nx.get_edge_attributes(self.graph, 'weight').values()
                 return sum(w) / 1000
             else:
+                logger.info('No skeleton data available. Use .get_skeleton() to fetch.')
+                return 'NA'
+        elif key == 'bbox':
+            if 'nodes' in self.__dict__:
+                return self.nodes.describe().loc[['min','max'],['x','y','z']].values.T
+            else:
+                logger.info('No skeleton data available. Use .get_skeleton() to fetch.')
                 return 'NA'
         else:
             raise AttributeError('Attribute "%s" not found' % key)
@@ -470,7 +500,9 @@ def _clear_temp_attr(self, exclude=[]):
         for a in [at for at in temp_att if at not in exclude]:
             try:
                 delattr(self, a)
+                logger.debug('Neuron {}: {} cleared'.format(self.skeleton_id, a))
             except BaseException:
+                logger.debug('Neuron {}: Unable to clear temporary attribute "{}"'.format(self.skeleton_id, a))
                 pass
 
         temp_node_cols = ['flow_centrality', 'strahler_index']
@@ -844,10 +876,11 @@ def prune_distal_to(self, node, inplace=True):
 
         for n in node:
             prox = graph_utils.cut_neuron(x, n, ret='proximal')
+            # Reinitialise with proximal data
             x.__init__(prox, x._remote_instance, x.meta_data)
-
-        # Clear temporary attributes is done by cut_neuron
-        # x._clear_temp_attr()
+            # Remove potential "left over" attributes (happens if we use a copy)
+            x._clear_temp_attr(exclude=['graph', 'igraph', 'type',
+                                        'classify_nodes'])
 
         if not inplace:
             return x
@@ -879,7 +912,11 @@ def prune_proximal_to(self, node, inplace=True):
 
         for n in node:
             dist = graph_utils.cut_neuron(x, n, ret='distal')
+            # Reinitialise with distal data
             x.__init__(dist, x._remote_instance, x.meta_data)
+            # Remove potential "left over" attributes (happens if we use a copy)
+            x._clear_temp_attr(exclude=['graph', 'igraph', 'type',
+                                        'classify_nodes'])
 
         # Clear temporary attributes is done by cut_neuron
         # x._clear_temp_attr()
@@ -1431,13 +1468,28 @@ def __len__(self):
         """Use skeleton ID here, otherwise this is terribly slow."""
         return len(self.skeleton_id)
 
+    def __dir__(self):
+        """ Custom __dir__ to add some parameters that we want to make
+        searchable.
+        """
+        add_attributes = ['n_open_ends', 'n_branch_nodes', 'n_end_nodes',
+                          'cable_length', 'root', 'neuron_name',
+                          'nodes', 'annotations', 'partners', 'review_status',
+                          'connectors', 'presynapses', 'postsynapses',
+                          'gap_junctions', 'soma', 'root', 'tags',
+                          'n_presynapses', 'n_postsynapses', 'n_connectors',
+                          'skeleton_id', 'empty', 'shape', 'bbox']
+
+        return list(set(super().__dir__() + add_attributes))
+
     def __getattr__(self, key):
         if key == 'shape':
             return (self.__len__(),)
         elif key in ['n_nodes', 'n_connectors', 'n_presynapses',
                      'n_postsynapses', 'n_open_ends', 'n_end_nodes',
                      'cable_length', 'tags', 'igraph', 'soma', 'root',
-                     'segments', 'graph', 'n_branch_nodes', 'dps']:
+                     'segments', 'graph', 'n_branch_nodes', 'dps',
+                     'sampling_resolution']:
             self.get_skeletons(skip_existing=True)
             return np.array([getattr(n, key) for n in self.neurons])
         elif key == 'neuron_name':
@@ -1455,6 +1507,8 @@ def __getattr__(self, key):
                 this_n['skeleton_id'] = n.skeleton_id
                 data.append(this_n)
             return pd.concat(data, axis=0, ignore_index=True)
+        elif key == 'bbox':
+            return self.nodes.describe().loc[['min','max'],['x','y','z']].values.T
         elif key == '_remote_instance':
             all_instances = [
                 n._remote_instance for n in self.neurons if n._remote_instance != None]

pymaid/fetch.py

@@ -1578,9 +1578,8 @@ def get_connector_links(x, with_tags=False, chunk_size=50, remote_instance=None)
 
     Parameters
     ----------
-    x :                 list of connector IDs | CatmaidNeuron | CatmaidNeuronList
-                        Connector ID(s) to retrieve details for. If
-                        CatmaidNeuron/List, will use their connectors.
+    x :                 int | CatmaidNeuron | CatmaidNeuronList
+                        Neurons/Skeleton IDs to retrieve link details for.
     with_tags :         bool, optional
                         If True will also return dictionary of connector tags.
     chunk_size :        int, optional

pymaid/graph_utils.py

@@ -887,7 +887,12 @@ def reroot_neuron(x, new_root, inplace=False):
 
 
 def cut_neuron(x, cut_node, ret='both'):
-    """ Split neuron at given point. Returns two new neurons.
+    """ Split neuron at given point and returns two new neurons.
+
+    Note
+    ----
+    Split is performed between cut node and its parent node. However, cut node
+    will still be present in both resulting neurons.
 
     Parameters
     ----------
@@ -970,23 +975,22 @@ def _cut_igraph(x, cut_node, ret):
     g = x.igraph.copy()
 
     # Get vertex index
-    cut_node = g.vs.find(node_id=cut_node).index
+    cut_ix = g.vs.find(node_id=cut_node).index
 
     # Get edge to parent
-    e = g.es.find(_source=cut_node)
+    e = g.es.find(_source=cut_ix)
 
     # Remove edge
     g.delete_edges(e)
 
     # Make graph undirected -> otherwise .decompose() throws an error
     # This issue is fixed in the up-to-date branch of igraph-python
-    # (which is not on PYPI...)
+    # (which is not on PyPI O_o )
     g.to_undirected(combine_edges='first')
 
-    # Get subgraph
+    # Get subgraph -> fastest way to get sets of nodes for subsetting
     a, b = g.decompose(mode='WEAK')
-
-    # Important: a,b are now UNDIRECTED graphs -> we must not keep using them.
+    # IMPORTANT: a,b are now UNDIRECTED graphs -> we must not keep using them!
 
     if x.root[0] in a.vs['node_id']:
         dist_graph, prox_graph = b, a
@@ -1004,7 +1008,7 @@ def _cut_igraph(x, cut_node, ret):
         dist._clear_temp_attr(exclude=['igraph', 'type', 'classify_nodes'])
 
     if ret == 'proximal' or ret == 'both':
-        prox = subset_neuron(x, prox_graph.vs['node_id'], clear_temp=False)
+        prox = subset_neuron(x, prox_graph.vs['node_id'] + [cut_node], clear_temp=False)
 
         # Change new root for dist
         prox.nodes.loc[prox.nodes.treenode_id == cut_node, 'type'] = 'end'
@@ -1104,6 +1108,8 @@ def subset_neuron(x, subset, clear_temp=True, remove_disconnected=True,
             Cut neuron at specific point.
 
     """
+    if isinstance(x, core.CatmaidNeuronList) and len(x) == 1:
+        x = x[0]
 
     if not isinstance(x, core.CatmaidNeuron):
         raise TypeError('Can only process data of type "CatmaidNeuron", not\