COMP: Fix runtime warning from ipython

IPython/parallel.py:13: ShimWarning: The `IPython.parallel` package has
been deprecated. You should import from ipyparallel instead.

https://github.com/ipython/ipyparallel

https://ipyparallel.readthedocs.org/en/latest/

Author: hjmjohnson

nipype/pipeline/plugins/ipython.py

@@ -10,7 +10,7 @@
 IPython_not_loaded = False
 try:
     from IPython import __version__ as IPyversion
-    from IPython.parallel.error import TimeoutError
+    from ipyparallel.error import TimeoutError
 except:
     IPython_not_loaded = True
 
@@ -42,7 +42,7 @@ class IPythonPlugin(DistributedPluginBase):
 
     def __init__(self, plugin_args=None):
         if IPython_not_loaded:
-            raise ImportError('IPython parallel could not be imported')
+            raise ImportError('ipyparallel could not be imported')
         super(IPythonPlugin, self).__init__(plugin_args=plugin_args)
         self.iparallel = None
         self.taskclient = None
@@ -51,11 +51,11 @@ def __init__(self, plugin_args=None):
 
     def run(self, graph, config, updatehash=False):
         """Executes a pre-defined pipeline is distributed approaches
-        based on IPython's parallel processing interface
+        based on IPython's ipyparallel processing interface
         """
         # retrieve clients again
         try:
-            name = 'IPython.parallel'
+            name = 'ipyparallel'
             __import__(name)
             self.iparallel = sys.modules[name]
         except ImportError:

nipype/pipeline/plugins/ipythonx.py

@@ -21,14 +21,14 @@ class IPythonXPlugin(DistributedPluginBase):
 
     def __init__(self, plugin_args=None):
         if IPython_not_loaded:
-            raise ImportError('IPython parallel could not be imported')
+            raise ImportError('ipyparallel could not be imported')
         super(IPythonXPlugin, self).__init__(plugin_args=plugin_args)
         self.ipyclient = None
         self.taskclient = None
 
     def run(self, graph, config, updatehash=False):
         """Executes a pre-defined pipeline is distributed approaches
-        based on IPython's parallel processing interface
+        based on IPython's ipyparallel processing interface
         """
         # retrieve clients again
         try: