Implement websocket client class (#14)

# Description
Implements a WS communication-based client
Refactor to minimize duplicate code


# Issues
<!-- If this is related to or closes an issue/other PR, please note them
here -->

# Other Notes
- Goes with NeonGeckoCom/neon-hana#20

---------

Co-authored-by: Daniel McKnight <[email protected]>

.github/workflows/publish_test_build.yml

@@ -15,6 +15,7 @@ jobs:
     with:
       version_file: "neon_nodes/version.py"
       publish_prerelease: true
+      update_changelog: true
   trigger_os_build:
     runs-on: ubuntu-latest
     needs: publish_alpha_release

README.md

@@ -7,3 +7,19 @@ processing, and presenting a response to the user.
 The voice client will start a service that listens for a wake word on the local
 system, sends recorded speech to a HANA endpoint for processing, and plays back
 the response.
+
+## Websocket Client
+The websocket client starts a local listener service and establishes a websocket
+connection to a remove HANA server. Compared to the Voice Client, this has 
+lower latency and allows for asynchronous messages from the HANA server.
+
+## Configuration
+This service is configured via `~/.config/neon/neon.yaml`.
+
+```yaml
+neon_node:
+  description: Neon Node  # Friendly description of the node
+  hana_address: https://hana.neonaiservices.com  # Hana server HTTP address
+  hana_username: node_user  # Hana node user username
+  hana_password: node_password  # Hana node user password
+```

neon_nodes/__init__.py

@@ -23,3 +23,30 @@
 # LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
 # NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
 # SOFTWARE,  EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+from mock import Mock
+from ovos_utils.log import LOG
+
+class MockTransformers(Mock):
+    def transform(self, chunk):
+        return chunk, dict()
+
+
+def on_ready():
+    LOG.info("ready")
+
+
+def on_stopping():
+    LOG.info("stopping")
+
+
+def on_error(e="unknown"):
+    LOG.error(e)
+
+
+def on_alive():
+    LOG.debug("alive")
+
+
+def on_started():
+    LOG.debug("started")

neon_nodes/configuration/system.yaml

@@ -1,8 +1,13 @@
 neon_node:
   description: Neon Node
+  # TODO: Stable release address
+  hana_address: https://hana.neonaibeta.com
+  hana_username: neon
+  hana_password: neon
 microphone:
   module: ovos-microphone-plugin-alsa
 listener:
+  sample_rate: 16000
   wake_word: hey_mycroft
   stand_up_word: ""
   VAD:

neon_nodes/voice_client.py

@@ -47,30 +47,7 @@
 from pydub import AudioSegment
 from pydub.playback import play
 
-
-class MockTransformers(Mock):
-    def transform(self, chunk):
-        return chunk, dict()
-
-
-def on_ready():
-    LOG.info("ready")
-
-
-def on_stopping():
-    LOG.info("stopping")
-
-
-def on_error(e="unknown"):
-    LOG.error(e)
-
-
-def on_alive():
-    LOG.debug("alive")
-
-
-def on_started():
-    LOG.debug("started")
+from neon_nodes import on_alive, on_error, on_ready, on_started, on_stopping, MockTransformers
 
 
 class NeonVoiceClient:
@@ -81,7 +58,9 @@ def __init__(self, bus=None, ready_hook=on_ready, error_hook=on_error,
         self.stopping_hook = stopping_hook
         alive_hook()
         self.config = Configuration()
-        self._device_data = self.config.get('neon_node', {})
+        self._node_config = self.config.get('neon_node', {})
+        self._hana_address = self._node_config.get('hana_address')
+
         LOG.init(self.config.get("logging"))
         self.bus = bus or FakeBus()
         self.lang = self.config.get('lang') or "en-us"
@@ -108,7 +87,7 @@ def __init__(self, bus=None, ready_hook=on_ready, error_hook=on_error,
         self._error_sound = None
 
         self._network_info = dict()
-        self._node_data = dict()
+        self._node_data = {"description": self._node_config.get("description")}
 
         started_hook()
         self.run()
@@ -240,18 +219,21 @@ def get_audio_response(self, audio: bytes):
         audio_data = b64encode(audio).decode("utf-8")
         transcript = request_backend("neon/get_stt",
                                      {"encoded_audio": audio_data,
-                                      "lang_code": self.lang})
+                                      "lang_code": self.lang},
+                                     server_url=self._hana_address)
         transcribed = transcript['transcripts'][0]
         LOG.info(transcribed)
         response = request_backend("neon/get_response",
                                    {"lang_code": self.lang,
                                     "user_profile": self.user_profile,
                                     "node_data": self.node_data,
-                                    "utterance": transcribed})
+                                    "utterance": transcribed},
+                                   server_url=self._hana_address)
         answer = response['answer']
         LOG.info(answer)
         audio = request_backend("neon/get_tts", {"lang_code": self.lang,
-                                                 "to_speak": answer})
+                                                 "to_speak": answer},
+                                server_url=self._hana_address)
         audio_bytes = b64decode(audio['encoded_audio'])
         play(AudioSegment.from_file(io.BytesIO(audio_bytes), format="wav"))
         LOG.info(f"Playback completed")
@@ -272,6 +254,4 @@ def main(*args, **kwargs):
 
 
 if __name__ == "__main__":
-    # environ.setdefault("OVOS_CONFIG_BASE_FOLDER", "neon")
-    # environ.setdefault("OVOS_CONFIG_FILENAME", "diana.yaml")
     main()