deviceserver: get the module logger at global level.

Instead of having the individual functions getting the module logger
themselves, get the module logger at global level as recommended by
the logging documentation.

Author: carandraug

microscope/deviceserver.py

@@ -38,6 +38,10 @@
 
 import microscope.devices
 
+
+_logger = logging.getLogger(__name__)
+
+
 # Pyro configuration. Use pickle because it can serialize numpy ndarrays.
 Pyro4.config.SERIALIZERS_ACCEPTED.add('pickle')
 Pyro4.config.SERIALIZER = 'pickle'
@@ -136,7 +140,6 @@ def clone(self):
 
     def run(self):
         cls_name = self._device_def['cls'].__name__
-        logger = logging.getLogger(__name__)
 
         # If the multiprocessing start method is fork, the child
         # process gets a copy of the root logger.  The copy is
@@ -172,8 +175,8 @@ def run(self):
             try:
                 self._device.initialize()
             except Exception as e:
-                logger.info("Failed to start device. Retrying in 5s.",
-                            exc_info=e)
+                _logger.info("Failed to start device. Retrying in 5s.",
+                             exc_info=e)
                 time.sleep(5)
             else:
                 break
@@ -194,7 +197,7 @@ def run(self):
         log_handler.setFormatter(_create_log_formatter(cls_name))
         root_logger.addHandler(log_handler)
 
-        logger.info('Device initialized; starting daemon.')
+        _logger.info('Device initialized; starting daemon.')
 
         pyro_daemon.register(self._device, cls_name)
         if isinstance(self._device, microscope.devices.ControllerDevice):
@@ -216,9 +219,10 @@ def run(self):
         pyro_thread = Thread(target = pyro_daemon.requestLoop)
         pyro_thread.daemon = True
         pyro_thread.start()
-        logger.info('Serving %s' % pyro_daemon.uriFor(self._device))
+        _logger.info('Serving %s' % pyro_daemon.uriFor(self._device))
         if isinstance(self._device, microscope.devices.FloatingDeviceMixin):
-            logger.info('Device UID on port %s is %s' % (port, self._device.get_id()))
+            _logger.info('Device UID on port %s is %s'
+                         % (port, self._device.get_id()))
         # Wait for termination event. We should just be able to call
         # wait() on the exit_event, but this causes issues with locks
         # in multiprocessing - see http://bugs.python.org/issue30975 .
@@ -234,7 +238,6 @@ def run(self):
 
 
 def serve_devices(devices, exit_event=None):
-    logger = logging.getLogger(__name__)
     root_logger = logging.getLogger()
 
     log_handler = RotatingFileHandler("__MAIN__.log")
@@ -260,7 +263,7 @@ def serve_devices(devices, exit_event=None):
     def term_func(sig, frame):
         """Terminate subprocesses cleanly."""
         if parent == multiprocessing.current_process ():
-            logger.debug("Shutting down all servers.")
+            _logger.debug("Shutting down all servers.")
             exit_event.set()
             # Join keep_alive_thread so that it can't modify the list
             # of servers.
@@ -280,7 +283,7 @@ def term_func(sig, frame):
 
     # Group devices by class.
     if not by_class:
-        logger.critical("No valid devices specified. Exiting")
+        _logger.critical("No valid devices specified. Exiting")
         sys.exit()
 
     for cls, devs in by_class.items():
@@ -316,27 +319,26 @@ def keep_alive():
                 if s.is_alive():
                     continue
                 else:
-                    logger.info(("DeviceServer Failure. Process %s is dead with"
-                                 " exitcode %s. Restarting...")
-                                % (s.pid, s.exitcode))
+                    _logger.info("DeviceServer Failure. Process %s is dead with"
+                                 " exitcode %s. Restarting..."
+                                 % (s.pid, s.exitcode))
                     servers.remove(s)
                     servers.append(s.clone())
 
                     try:
                         s.join(30)
                     except:
-                        logger.error("... could not join PID %s." % (old_pid))
+                        _logger.error("... could not join PID %s." % (old_pid))
                     else:
                         old_pid = s.pid
                         del (s)
                         servers[-1].start()
-                        logger.info(("... DeviceServer with PID %s restarted"
-                                     " as PID %s.")
-                                    % (old_pid, servers[-1].pid))
+                        _logger.info("... DeviceServer with PID %s restarted"
+                                     " as PID %s." % (old_pid, servers[-1].pid))
             if len(servers) == 0:
                 # Log and exit if no servers running. May want to change this
                 # if we add some interface to interactively restart servers.
-                logger.info("No servers running. Exiting.")
+                _logger.info("No servers running. Exiting.")
                 exit_event.set()
             try:
                 time.sleep(5)
@@ -350,20 +352,20 @@ def keep_alive():
         try:
             time.sleep(5)
         except (KeyboardInterrupt, IOError):
-            logger.debug("KeyboardInterrupt or IOError")
+            _logger.debug("KeyboardInterrupt or IOError")
             exit_event.set()
 
-    logger.debug("Shutting down servers ...")
+    _logger.debug("Shutting down servers ...")
     while len(servers) > 0:
         for s in servers:
             if not s.is_alive():
                 servers.remove(s)
                 del(s)
         time.sleep(1)
-    logger.info(" ... No more servers running.")
-    logger.debug("Joining threads ...")
+    _logger.info(" ... No more servers running.")
+    _logger.debug("Joining threads ...")
     keep_alive_thread.join()
-    logger.debug("... Threads joined. Exiting.")
+    _logger.debug("... Threads joined. Exiting.")
     return
 
 
@@ -413,7 +415,6 @@ def validate_devices(configfile):
 
 def __console__():
     """Serve devices from a console process."""
-    logger = logging.getLogger(__name__)
     root_logger = logging.getLogger()
     if __debug__:
         root_logger.setLevel(logging.DEBUG)
@@ -427,13 +428,13 @@ def __console__():
     root_logger.addFilter(Filter())
 
     if len(sys.argv) < 2:
-        logger.critical("No config file specified. Exiting.")
+        _logger.critical("No config file specified. Exiting.")
         devices = []
     else:
         try:
             devices = validate_devices(sys.argv[1])
         except Exception as e:
-            logger.critical(e)
+            _logger.critical(e)
             devices = []
 
     if not devices: