Merge pull request #86 from daily-co/transport-refactor

Starting refactor of transports into their own directory

src/dailyai/pipeline/merge_pipeline.py

@@ -12,9 +12,10 @@ class SequentialMergePipeline(Pipeline):
         self.pipelines = pipelines
 
     async def run_pipeline(self):
-        for pipeline in self.pipelines:
+        for idx, pipeline in enumerate(self.pipelines):
             while True:
                 frame = await pipeline.sink.get()
+                print(idx, frame)
                 if isinstance(
                         frame, EndFrame) or isinstance(
                         frame, EndPipeFrame):

src/dailyai/transports/abstract_transport.py

@@ -0,0 +1,41 @@
+from abc import abstractmethod
+import asyncio
+import logging
+import time
+
+from dailyai.pipeline.frame_processor import FrameProcessor
+from dailyai.pipeline.pipeline import Pipeline
+
+
+class AbstractTransport:
+    def __init__(self, **kwargs):
+        self.send_queue = asyncio.Queue()
+        self.receive_queue = asyncio.Queue()
+        self.completed_queue = asyncio.Queue()
+
+        duration_minutes = kwargs.get("duration_minutes") or 10
+        self._expiration = time.time() + duration_minutes * 60
+
+        self._mic_enabled = kwargs.get("mic_enabled") or False
+        self._mic_sample_rate = kwargs.get("mic_sample_rate") or 16000
+        self._camera_enabled = kwargs.get("camera_enabled") or False
+        self._camera_width = kwargs.get("camera_width") or 1024
+        self._camera_height = kwargs.get("camera_height") or 768
+        self._speaker_enabled = kwargs.get("speaker_enabled") or False
+        self._speaker_sample_rate = kwargs.get("speaker_sample_rate") or 16000
+        self._fps = kwargs.get("fps") or 8
+
+        self._logger: logging.Logger = logging.getLogger("dailyai.transport")
+
+    @abstractmethod
+    async def run(self, pipeline: Pipeline, override_pipeline_source_queue=True):
+        pass
+
+    @abstractmethod
+    async def run_interruptible_pipeline(
+        self,
+        pipeline: Pipeline,
+        pre_processor: FrameProcessor | None = None,
+        post_processor: FrameProcessor | None = None,
+    ):
+        pass

src/dailyai/transports/daily_transport.py

@@ -24,10 +24,10 @@ from daily import (
     VirtualSpeakerDevice,
 )
 
-from dailyai.services.base_transport_service import BaseTransportService
+from dailyai.transports.threaded_transport import ThreadedTransport
 
 
-class DailyTransportService(BaseTransportService, EventHandler):
+class DailyTransport(ThreadedTransport, EventHandler):
     _daily_initialized = False
     _lock = threading.Lock()
 
@@ -48,7 +48,7 @@ class DailyTransportService(BaseTransportService, EventHandler):
         start_transcription: bool = False,
         **kwargs,
     ):
-        # This will call BaseTransportService.__init__ method, not EventHandler
+        # This will call ThreadedTransport.__init__ method, not EventHandler
         super().__init__(**kwargs)
 
         self._room_url: str = room_url
@@ -140,10 +140,10 @@ class DailyTransportService(BaseTransportService, EventHandler):
 
     def _prerun(self):
         # Only initialize Daily once
-        if not DailyTransportService._daily_initialized:
-            with DailyTransportService._lock:
+        if not DailyTransport._daily_initialized:
+            with DailyTransport._lock:
                 Daily.init()
-                DailyTransportService._daily_initialized = True
+                DailyTransport._daily_initialized = True
         self.client = CallClient(event_handler=self)
 
         if self._mic_enabled:

src/dailyai/transports/local_transport.py

@@ -3,10 +3,10 @@ import numpy as np
 import tkinter as tk
 import pyaudio
 
-from dailyai.services.base_transport_service import BaseTransportService
+from dailyai.transports.threaded_transport import ThreadedTransport
 
 
-class LocalTransportService(BaseTransportService):
+class LocalTransport(ThreadedTransport):
     def __init__(self, **kwargs):
         super().__init__(**kwargs)
         self._sample_width = kwargs.get("sample_width") or 2

src/dailyai/transports/threaded_transport.py

@@ -27,6 +27,7 @@ from dailyai.pipeline.frames import (
 )
 from dailyai.pipeline.pipeline import Pipeline
 from dailyai.services.ai_services import TTSService
+from dailyai.transports.abstract_transport import AbstractTransport
 
 torch.set_num_threads(1)
 
@@ -72,20 +73,13 @@ class VADState(Enum):
     STOPPING = 4
 
 
-class BaseTransportService:
+class ThreadedTransport(AbstractTransport):
 
     def __init__(
         self,
         **kwargs,
     ) -> None:
-        self._mic_enabled = kwargs.get("mic_enabled") or False
-        self._mic_sample_rate = kwargs.get("mic_sample_rate") or 16000
-        self._camera_enabled = kwargs.get("camera_enabled") or False
-        self._camera_width = kwargs.get("camera_width") or 1024
-        self._camera_height = kwargs.get("camera_height") or 768
-        self._speaker_enabled = kwargs.get("speaker_enabled") or False
-        self._speaker_sample_rate = kwargs.get("speaker_sample_rate") or 16000
-        self._fps = kwargs.get("fps") or 8
+        super().__init__(**kwargs)
         self._vad_start_s = kwargs.get("vad_start_s") or 0.2
         self._vad_stop_s = kwargs.get("vad_stop_s") or 0.8
         self._context = kwargs.get("context") or []
@@ -105,14 +99,6 @@ class BaseTransportService:
         self._vad_state = VADState.QUIET
         self._user_is_speaking = False
 
-        duration_minutes = kwargs.get("duration_minutes") or 10
-        self._expiration = time.time() + duration_minutes * 60
-
-        self.send_queue = asyncio.Queue()
-        self.receive_queue = asyncio.Queue()
-
-        self.completed_queue = asyncio.Queue()
-
         self._threadsafe_send_queue = queue.Queue()
 
         self._images = None
@@ -125,8 +111,6 @@ class BaseTransportService:
         self._stop_threads = threading.Event()
         self._is_interrupted = threading.Event()
 
-        self._logger: logging.Logger = logging.getLogger()
-
     async def run(self, pipeline: Pipeline | None = None, override_pipeline_source_queue=True):
         self._prerun()
 
@@ -193,8 +177,7 @@ class BaseTransportService:
     async def run_interruptible_pipeline(
         self,
         pipeline: Pipeline,
-        allow_interruptions=True,
-        pre_processor=None,
+        pre_processor: FrameProcessor | None = None,
         post_processor: FrameProcessor | None = None,
     ):
         pipeline.set_sink(self.send_queue)

src/dailyai/transports/websocket_transport.py

@@ -7,7 +7,8 @@ from dailyai.pipeline.frame_processor import FrameProcessor
 from dailyai.pipeline.frames import AudioFrame, ControlFrame, EndFrame, Frame, TTSEndFrame, TTSStartFrame, TextFrame
 from dailyai.pipeline.pipeline import Pipeline
 from dailyai.serializers.protobuf_serializer import ProtobufFrameSerializer
-from dailyai.services.base_transport_service import BaseTransportService
+from dailyai.transports.abstract_transport import AbstractTransport
+from dailyai.transports.threaded_transport import ThreadedTransport
 
 
 class WebSocketFrameProcessor(FrameProcessor):
@@ -45,7 +46,7 @@ class WebSocketFrameProcessor(FrameProcessor):
             yield frame
 
 
-class WebsocketTransport(BaseTransportService):
+class WebsocketTransport(AbstractTransport):
     def __init__(self, **kwargs):
         super().__init__(**kwargs)
         self._sample_width = kwargs.get("sample_width", 2)