Merge branch 'main' into Raspberry-pi-docs

OS/01/.env.example

@@ -14,9 +14,15 @@ WHISPER_MODEL_NAME="ggml-tiny.en.bin"
 PIPER_VOICE_URL="https://huggingface.co/rhasspy/piper-voices/resolve/main/en/en_US/lessac/medium/"
 PIPER_VOICE_NAME="en_US-lessac-medium.onnx"
 
+# Expose through Ngrok
+# Uncomment following line with your Ngrok auth token (https://dashboard.ngrok.com/get-started/your-authtoken)
+#NGROK_AUTHTOKEN="AUTH TOKEN"
+
 # If SERVER_START, this is where we'll serve the server.
 # If DEVICE_START, this is where the device expects the server to be.
 SERVER_URL=ws://localhost:8000/
+# If you are setting up Ngrok then either change the below to Ngrok URL if running device separately, else comment it
+SERVER_CONNECTION_URL=ws://localhost:8000/
 SERVER_START=True
 DEVICE_START=True
 

OS/01/device.py

@@ -38,11 +38,6 @@ RATE = 44100  # Sample rate
 RECORDING = False  # Flag to control recording state
 SPACEBAR_PRESSED = False  # Flag to track spacebar press state
 
-# Configuration for WebSocket
-WS_URL = os.getenv('SERVER_URL')
-if not WS_URL:
-    raise ValueError("The environment variable SERVER_URL is not set. Please set it to proceed.")
-
 # Specify OS
 current_platform = get_system_info()
 
@@ -203,6 +198,11 @@ async def websocket_communication(WS_URL):
 
 if __name__ == "__main__":
     async def main():
+        # Configuration for WebSocket
+        WS_URL = os.getenv('SERVER_CONNECTION_URL')
+        if not WS_URL:
+            raise ValueError("The environment variable SERVER_CONNECTION_URL is not set. Please set it to proceed.")
+
         # Start the WebSocket communication
         asyncio.create_task(websocket_communication(WS_URL))
 

OS/01/i.py

@@ -9,7 +9,6 @@ from interpreter import OpenInterpreter
 
 
 def configure_interpreter(interpreter: OpenInterpreter):
-def configure_interpreter(interpreter):
     ### SYSTEM MESSAGE
 
     # The system message is where most of the 01's behavior is configured.

OS/01/requirements.txt

@@ -11,3 +11,4 @@ ffmpeg-python
 textual
 pydub
 python-dotenv
+ngrok

OS/01/server.py

@@ -14,6 +14,7 @@ import threading
 import uvicorn
 import re
 from fastapi import FastAPI
+from fastapi.responses import PlainTextResponse
 from threading import Thread
 from starlette.websockets import WebSocket
 from stt import stt_bytes
@@ -24,6 +25,7 @@ import urllib.parse
 from utils.kernel import put_kernel_messages_into_queue
 from i import configure_interpreter
 from interpreter import interpreter
+import ngrok
 
 from utils.logs import setup_logging
 from utils.logs import logger
@@ -93,6 +95,10 @@ if os.getenv('CODE_RUNNER') == "device":
 # Configure interpreter
 interpreter = configure_interpreter(interpreter)
 
+@app.get("/ping")
+async def ping():
+    return PlainTextResponse("pong")
+
 @app.websocket("/")
 async def websocket_endpoint(websocket: WebSocket):
     await websocket.accept()
@@ -229,6 +235,17 @@ async def stream_or_play_tts(sentence):
         await to_device.put({"role": "assistant", "type": "audio", "format": "audio/mp3", "content": str(audio_bytes)})
         await to_device.put({"role": "assistant", "type": "audio", "format": "audio/mp3", "end": True})
 
+async def setup_ngrok(ngrok_auth_token, parsed_url):
+    # Set up Ngrok
+    logger.info("Setting up Ngrok")
+    ngrok_listener = await ngrok.forward(f"{parsed_url.hostname}:{parsed_url.port}", authtoken=ngrok_auth_token)
+    ngrok_parsed_url = urllib.parse.urlparse(ngrok_listener.url())
+
+    # Setup SERVER_URL environment variable for device to use
+    connection_url = f"wss://{ngrok_parsed_url.hostname}/"
+    logger.info(f"Ngrok established at {ngrok_parsed_url.geturl()}")
+    logger.info(f"\033[1mSERVER_CONNECTION_URL should be set to \"{connection_url}\"\033[0m")
+
 
 from uvicorn import Config, Server
 
@@ -247,6 +264,12 @@ if __name__ == "__main__":
         if not server_url:
             raise ValueError("The environment variable SERVER_URL is not set. Please set it to proceed.")
         parsed_url = urllib.parse.urlparse(server_url)
+
+        # Set up Ngrok
+        ngrok_auth_token = os.getenv('NGROK_AUTHTOKEN')
+        if ngrok_auth_token is not None:
+            await setup_ngrok(ngrok_auth_token, parsed_url)
+            
         logger.info("Starting `server.py`...")
 
         config = Config(app, host=parsed_url.hostname, port=parsed_url.port, lifespan='on')

OS/01/start.sh

@@ -63,6 +63,13 @@ fi
 
 start_device() {
     echo "Starting device..."
+    if [[ -n $NGROK_AUTHTOKEN ]]; then
+        echo "Waiting for Ngrok to setup"
+        sleep 7
+        read -p "Enter the Ngrok URL: " ngrok_url
+        export SERVER_CONNECTION_URL=$ngrok_url
+        echo "SERVER_CONNECTION_URL set to $SERVER_CONNECTION_URL"
+    fi
     python device.py &
     DEVICE_PID=$!
     echo "Device started as process $DEVICE_PID"
@@ -90,18 +97,18 @@ stop_processes() {
 # Trap SIGINT and SIGTERM to stop processes when the script is terminated
 trap stop_processes SIGINT SIGTERM
 
-# DEVICE
-# Start device if DEVICE_START is True
-if [[ "$DEVICE_START" == "True" ]]; then
-    start_device
-fi
-
 # SERVER
 # Start server if SERVER_START is True
 if [[ "$SERVER_START" == "True" ]]; then
     start_server
 fi
 
+# DEVICE
+# Start device if DEVICE_START is True
+if [[ "$DEVICE_START" == "True" ]]; then
+    start_device
+fi
+
 # Wait for device and server processes to exit
 wait $DEVICE_PID
 wait $SERVER_PID