Some improvements and cleanups in the SmallWebRTCTransport text examples.

examples/foundational/41a-text-only-webrtc.py

@@ -5,30 +5,18 @@
 #
 
 import argparse
-import asyncio
-import io
 import os
-import re
-import shutil
-import sys
 
-import aiohttp
 from dotenv import load_dotenv
 from loguru import logger
-from PIL import Image
 
-from pipecat.audio.vad.silero import SileroVADAnalyzer
 from pipecat.frames.frames import (
-    Frame,
-    FunctionCallResultFrame,
     LLMMessagesAppendFrame,
-    URLImageRawFrame,
 )
 from pipecat.pipeline.pipeline import Pipeline
 from pipecat.pipeline.runner import PipelineRunner
 from pipecat.pipeline.task import PipelineParams, PipelineTask
 from pipecat.processors.aggregators.openai_llm_context import OpenAILLMContext
-from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
 from pipecat.processors.frameworks.rtvi import (
     ActionResult,
     RTVIAction,
@@ -38,6 +26,7 @@ from pipecat.processors.frameworks.rtvi import (
     RTVIProcessor,
     RTVIServerMessageFrame,
 )
+from pipecat.services.openai import OpenAIContextAggregatorPair
 from pipecat.services.openai.llm import OpenAILLMService
 from pipecat.transports.base_transport import TransportParams
 from pipecat.transports.network.small_webrtc import SmallWebRTCTransport
@@ -45,54 +34,28 @@ from pipecat.transports.network.webrtc_connection import SmallWebRTCConnection
 
 load_dotenv(override=True)
 
+
 # This is an example of a text-only chatbot using small webrtc tranport.
 # It uses the small webrtc transport prebuilt web UI.
 # https://github.com/pipecat-ai/small-webrtc-prebuilt
 
 
-async def run_bot(webrtc_connection: SmallWebRTCConnection, _: argparse.Namespace):
+def create_action_llm_append_to_messages(context_aggregator: OpenAIContextAggregatorPair):
-    logger.info(f"Starting bot")
-
-    transport = SmallWebRTCTransport(
-        webrtc_connection=webrtc_connection,
-        params=TransportParams(),
-    )
-
-    # Create an HTTP session for API calls
-    async with aiohttp.ClientSession() as session:
-        llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
-
-        messages = [
-            {
-                "role": "system",
-                "content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Respond to what the user said in a creative and helpful way.",
-            },
-        ]
-
-        context = OpenAILLMContext(messages)
-        context_aggregator = llm.create_context_aggregator(context)
-
     async def action_llm_append_to_messages_handler(
         rtvi: RTVIProcessor, service: str, arguments: dict[str, any]
     ) -> ActionResult:
-            run_immediately = (
+        run_immediately = arguments["run_immediately"] if "run_immediately" in arguments else True
-                arguments["run_immediately"] if "run_immediately" in arguments else True
+        logger.info(f"run_immediately: {run_immediately}")
-            )
-
         if run_immediately:
             await rtvi.interrupt_bot()
-
             # We just interrupted the bot so it should be fine to use the
             # context directly instead of through frame.
             if "messages" in arguments and arguments["messages"]:
-                    mess = arguments["messages"]
                 frame = LLMMessagesAppendFrame(messages=arguments["messages"])
                 await rtvi.push_frame(frame)
 
-            if run_immediately:
             frame = context_aggregator.user().get_context_frame()
             await rtvi.push_frame(frame)
-
         return True
 
     action_llm_append_to_messages = RTVIAction(
@@ -105,7 +68,30 @@ async def run_bot(webrtc_connection: SmallWebRTCConnection, _: argparse.Namespac
         ],
         handler=action_llm_append_to_messages_handler,
     )
+    return action_llm_append_to_messages
 
+
+async def run_bot(webrtc_connection: SmallWebRTCConnection, _: argparse.Namespace):
+    logger.info(f"Starting bot")
+
+    transport = SmallWebRTCTransport(
+        webrtc_connection=webrtc_connection,
+        params=TransportParams(),
+    )
+
+    llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
+
+    messages = [
+        {
+            "role": "system",
+            "content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Respond to what the user said in a creative and helpful way.",
+        },
+    ]
+
+    context = OpenAILLMContext(messages)
+    context_aggregator = llm.create_context_aggregator(context)
+
+    action_llm_append_to_messages = create_action_llm_append_to_messages(context_aggregator)
     rtvi = RTVIProcessor(config=RTVIConfig(config=[]))
     rtvi.register_action(action_llm_append_to_messages)
 

examples/foundational/41b-text-and-audio-webrtc.py

@@ -5,30 +5,19 @@
 #
 
 import argparse
-import asyncio
-import io
 import os
-import re
-import shutil
-import sys
 
-import aiohttp
 from dotenv import load_dotenv
 from loguru import logger
-from PIL import Image
 
 from pipecat.audio.vad.silero import SileroVADAnalyzer
 from pipecat.frames.frames import (
-    Frame,
-    FunctionCallResultFrame,
     LLMMessagesAppendFrame,
-    URLImageRawFrame,
 )
 from pipecat.pipeline.pipeline import Pipeline
 from pipecat.pipeline.runner import PipelineRunner
 from pipecat.pipeline.task import PipelineParams, PipelineTask
 from pipecat.processors.aggregators.openai_llm_context import OpenAILLMContext
-from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
 from pipecat.processors.frameworks.rtvi import (
     ActionResult,
     RTVIAction,
@@ -40,6 +29,7 @@ from pipecat.processors.frameworks.rtvi import (
 )
 from pipecat.services.cartesia.tts import CartesiaTTSService
 from pipecat.services.deepgram.stt import DeepgramSTTService
+from pipecat.services.openai import OpenAIContextAggregatorPair
 from pipecat.services.openai.llm import OpenAILLMService
 from pipecat.transports.base_transport import TransportParams
 from pipecat.transports.network.small_webrtc import SmallWebRTCTransport
@@ -52,44 +42,11 @@ load_dotenv(override=True)
 # https://github.com/pipecat-ai/small-webrtc-prebuilt
 
 
-async def run_bot(webrtc_connection: SmallWebRTCConnection, _: argparse.Namespace):
+def create_action_llm_append_to_messages(context_aggregator: OpenAIContextAggregatorPair):
-    logger.info(f"Starting bot")
-
-    transport = SmallWebRTCTransport(
-        webrtc_connection=webrtc_connection,
-        params=TransportParams(
-            audio_in_enabled=True,
-            audio_out_enabled=True,
-            vad_analyzer=SileroVADAnalyzer(),
-        ),
-    )
-
-    # Create an HTTP session for API calls
-    async with aiohttp.ClientSession() as session:
-        stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
-
-        llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
-
-        tts = CartesiaTTSService(
-            api_key=os.getenv("CARTESIA_API_KEY"), voice_id="71a7ad14-091c-4e8e-a314-022ece01c121"
-        )
-
-        messages = [
-            {
-                "role": "system",
-                "content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Respond to what the user says in a creative and helpful way. Explain to the User they can speak or type text to communicate with you.",
-            },
-        ]
-
-        context = OpenAILLMContext(messages)
-        context_aggregator = llm.create_context_aggregator(context)
-
     async def action_llm_append_to_messages_handler(
         rtvi: RTVIProcessor, service: str, arguments: dict[str, any]
     ) -> ActionResult:
-            run_immediately = (
+        run_immediately = arguments["run_immediately"] if "run_immediately" in arguments else True
-                arguments["run_immediately"] if "run_immediately" in arguments else True
-            )
 
         if run_immediately:
             await rtvi.interrupt_bot()
@@ -117,7 +74,40 @@ async def run_bot(webrtc_connection: SmallWebRTCConnection, _: argparse.Namespac
         ],
         handler=action_llm_append_to_messages_handler,
     )
+    return action_llm_append_to_messages
 
+
+async def run_bot(webrtc_connection: SmallWebRTCConnection, _: argparse.Namespace):
+    logger.info(f"Starting bot")
+
+    transport = SmallWebRTCTransport(
+        webrtc_connection=webrtc_connection,
+        params=TransportParams(
+            audio_in_enabled=True,
+            audio_out_enabled=True,
+            vad_analyzer=SileroVADAnalyzer(),
+        ),
+    )
+
+    stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
+
+    llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
+
+    tts = CartesiaTTSService(
+        api_key=os.getenv("CARTESIA_API_KEY"), voice_id="71a7ad14-091c-4e8e-a314-022ece01c121"
+    )
+
+    messages = [
+        {
+            "role": "system",
+            "content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Respond to what the user says in a creative and helpful way. Explain to the User they can speak or type text to communicate with you.",
+        },
+    ]
+
+    context = OpenAILLMContext(messages)
+    context_aggregator = llm.create_context_aggregator(context)
+
+    action_llm_append_to_messages = create_action_llm_append_to_messages(context_aggregator)
     rtvi = RTVIProcessor(config=RTVIConfig(config=[]))
     rtvi.register_action(action_llm_append_to_messages)