[VOL-2241] Python 3 refactor of pyvoltha

Majority of work was manual fixes to bytes and strings types, which are
different in py3, but same in py2. As the OMCI library does a lot of
these comparisons and scapy then renders packets, this was frequently
nontrival to debug.

Also:

- Removed grpc dep which wasn't being used, not py3 compatible
- s/Alarms/Events/ to work with protobuf changes per VOL-2224
- Automatic fixes via modernize tooling
- Removed unused OrderedWeakValueDict code
- Removed frameio send_frame specific to Darwin (MacOS), which had no
  corresponding linux code
- Use library functions for hex and unicode conversions
- Various other cleanups and fixes (EOL whitespace, etc.)

Also more (Matt):

 - handle stringify better, check if already string
 - use binary string for binary work
 - import new thread paths
 - update requirements.txt for newer libraries needed with newer python
 - return proper tuple for unpacking
 - bytes string formatting fixed
 - fix mock task unit test

Even more (Zack):

- Python 2/3 compat for _thread by using 'future'
- Bump version to 2.3.0

Change-Id: I53b596d374a944bfb80d0b112f21bcc1f8bcee6e
diff --git a/test/unit/extensions/omci/mock/mock_olt_handler.py b/test/unit/extensions/omci/mock/mock_olt_handler.py
index 142dbd8..9a51082 100644
--- a/test/unit/extensions/omci/mock/mock_olt_handler.py
+++ b/test/unit/extensions/omci/mock/mock_olt_handler.py
@@ -14,9 +14,11 @@
 # limitations under the License.
 #
 
+from __future__ import absolute_import
 import sys
-from mock_adapter_agent import MockDevice
+from .mock_adapter_agent import MockDevice
 from nose.twistedtools import reactor
+from six.moves import range
 
 
 class MockOltHandler(MockDevice):
@@ -45,7 +47,7 @@
 
         self.enabled = True                # OLT is enabled/active
         self.activated_onus = set()        # Activated ONU serial numbers
-        self.enabled_pons = range(0, 16)   # Enabled PONs
+        self.enabled_pons = list(range(0, 16))   # Enabled PONs
         self.max_tx = sys.maxint           # Fail after this many tx requests
         self.latency = 0.0                 # OMCI response latency (keep small)
 
@@ -101,7 +103,7 @@
     def _deliver_proxy_message(self, proxy_address, response):
         from common.frameio.frameio import hexify
         self._adapter_agent.receive_proxied_message(proxy_address,
-                                                    hexify(str(response)))
+                                                    hexify(response))
 
     def receive_proxied_message(self, _, __):
         assert False, 'This is never called on the OLT side of proxy messaging'