Merge pull request #284 from benxu3/async-interpreter

add --debug flag

README.md

@@ -129,7 +129,7 @@ If you want to run local speech-to-text using Whisper, you must install Rust. Fo
 
 To customize the behavior of the system, edit the [system message, model, skills library path,](https://docs.openinterpreter.com/settings/all-settings) etc. in the `profiles` directory under the `server` directory. This file sets up an interpreter, and is powered by Open Interpreter.
 
-To specify the text-to-speech service for the 01 `base_device.py`, set `interpreter.tts` to either "openai" for OpenAI, "elevenlabs" for ElevenLabs, or "coqui" for Coqui (local) in a profile. For the 01 Light, set `SPEAKER_SAMPLE_RATE` to 24000 for Coqui (local) or 22050 for OpenAI TTS. We currently don't support ElevenLabs TTS on the 01 Light.
+To specify the text-to-speech service for the 01 `base_device.py`, set `interpreter.tts` to either "openai" for OpenAI, "elevenlabs" for ElevenLabs, or "coqui" for Coqui (local) in a profile. For the 01 Light, set `SPEAKER_SAMPLE_RATE` in `client.ino` under the `esp32` client directory to 24000 for Coqui (local) or 22050 for OpenAI TTS. We currently don't support ElevenLabs TTS on the 01 Light.
 
 ## Ubuntu Dependencies
 

software/source/clients/base_device.py

@@ -91,6 +91,8 @@ def __init__(self):
         self.server_url = ""
         self.ctrl_pressed = False
         self.tts_service = ""
+        self.debug = False
+        self.playback_latency = None
 
     def fetch_image_from_camera(self, camera_index=CAMERA_DEVICE_INDEX):
         """Captures an image from the specified camera device and saves it to a temporary file. Adds the image to the captured_images list."""
@@ -164,6 +166,10 @@ async def play_audiosegments(self):
         while True:
             try:
                 audio = await self.audiosegments.get()
+                if self.debug and self.playback_latency and isinstance(audio, bytes):
+                    elapsed_time = time.time() - self.playback_latency
+                    print(f"Time from request to playback: {elapsed_time} seconds")
+                    self.playback_latency = None
 
                 if self.tts_service == "elevenlabs":
                     mpv_process.stdin.write(audio)  # type: ignore
@@ -219,6 +225,8 @@ def record_audio(self):
         stream.stop_stream()
         stream.close()
         print("Recording stopped.")
+        if self.debug:
+            self.playback_latency = time.time()
 
         duration = wav_file.getnframes() / RATE
         if duration < 0.3:

software/source/clients/linux/device.py

@@ -3,8 +3,9 @@
 device = Device()
 
 
-def main(server_url):
+def main(server_url, debug):
     device.server_url = server_url
+    device.debug = debug
     device.start()
 
 

software/source/clients/mac/device.py

@@ -3,8 +3,9 @@
 device = Device()
 
 
-def main(server_url):
+def main(server_url, debug):
     device.server_url = server_url
+    device.debug = debug
     device.start()
 
 

software/source/clients/windows/device.py

@@ -3,8 +3,9 @@
 device = Device()
 
 
-def main(server_url):
+def main(server_url, debug):
     device.server_url = server_url
+    device.debug = debug
     device.start()
 
 

software/source/server/async_interpreter.py

@@ -21,7 +21,14 @@
 
 
 class AsyncInterpreter:
-    def __init__(self, interpreter):
+    def __init__(self, interpreter, debug):
+        self.stt_latency = None
+        self.tts_latency = None
+        self.interpreter_latency = None
+        # time from first put to first yield
+        self.tffytfp = None
+        self.debug = debug
+
         self.interpreter = interpreter
         self.audio_chunks = []
 
@@ -126,6 +133,8 @@ def generate(self, message, start_interpreter):
                     # Experimental: The AI voice sounds better with replacements like these, but it should happen at the TTS layer
                     # content = content.replace(". ", ". ... ").replace(", ", ", ... ").replace("!", "! ... ").replace("?", "? ... ")
                     # print("yielding ", content)
+                    if self.tffytfp is None:
+                        self.tffytfp = time.time()
 
                     yield content
 
@@ -157,6 +166,10 @@ def generate(self, message, start_interpreter):
                         )
 
         # Send a completion signal
+        if self.debug:
+            end_interpreter = time.time()
+            self.interpreter_latency = end_interpreter - start_interpreter
+            print("INTERPRETER LATENCY", self.interpreter_latency)
         # self.add_to_output_queue_sync({"role": "server","type": "completion", "content": "DONE"})
 
     async def run(self):
@@ -171,13 +184,20 @@ async def run(self):
         while not self._input_queue.empty():
             input_queue.append(self._input_queue.get())
 
-        message = self.stt.text()
-
-        if self.audio_chunks:
-            audio_bytes = bytearray(b"".join(self.audio_chunks))
-            wav_file_path = bytes_to_wav(audio_bytes, "audio/raw")
-            print("wav_file_path ", wav_file_path)
-            self.audio_chunks = []
+        if self.debug:
+            start_stt = time.time()
+            message = self.stt.text()
+            end_stt = time.time()
+            self.stt_latency = end_stt - start_stt
+            print("STT LATENCY", self.stt_latency)
+
+            if self.audio_chunks:
+                audio_bytes = bytearray(b"".join(self.audio_chunks))
+                wav_file_path = bytes_to_wav(audio_bytes, "audio/raw")
+                print("wav_file_path ", wav_file_path)
+                self.audio_chunks = []
+        else:
+            message = self.stt.text()
 
         print(message)
 
@@ -204,11 +224,22 @@ async def run(self):
                         "end": True,
                     }
                 )
+                if self.debug:
+                    end_tts = time.time()
+                    self.tts_latency = end_tts - self.tts.stream_start_time
+                    print("TTS LATENCY", self.tts_latency)
                 self.tts.stop()
+
                 break
 
     async def _on_tts_chunk_async(self, chunk):
         # print("adding chunk to queue")
+        if self.debug and self.tffytfp is not None and self.tffytfp != 0:
+            print(
+                "time from first yield to first put is ",
+                time.time() - self.tffytfp,
+            )
+            self.tffytfp = 0
         await self._add_to_queue(self._output_queue, chunk)
 
     def on_tts_chunk(self, chunk):

software/source/server/async_server.py

@@ -1,30 +1,18 @@
-# import from the profiles directory the interpreter to be served
-
-# add other profiles to the directory to define other interpreter instances and import them here
-# {.profiles.fast: optimizes for STT/TTS latency with the fastest models }
-# {.profiles.local: uses local models and local STT/TTS }
-# {.profiles.default: uses default interpreter settings with optimized TTS latency }
-
-# from .profiles.fast import interpreter as base_interpreter
-# from .profiles.local import interpreter as base_interpreter
-from .profiles.default import interpreter as base_interpreter
-
 import asyncio
 import traceback
 import json
-from fastapi import FastAPI, WebSocket
+from fastapi import FastAPI, WebSocket, Depends
 from fastapi.responses import PlainTextResponse
 from uvicorn import Config, Server
 from .async_interpreter import AsyncInterpreter
 from fastapi.middleware.cors import CORSMiddleware
 from typing import List, Dict, Any
 import os
+import importlib.util
 
 os.environ["STT_RUNNER"] = "server"
 os.environ["TTS_RUNNER"] = "server"
 
-# interpreter.tts set in the profiles directory!!!!
-interpreter = AsyncInterpreter(base_interpreter)
 
 app = FastAPI()
 
@@ -37,13 +25,19 @@
 )
 
 
+async def get_debug_flag():
+    return app.state.debug
+
+
 @app.get("/ping")
 async def ping():
     return PlainTextResponse("pong")
 
 
 @app.websocket("/")
-async def websocket_endpoint(websocket: WebSocket):
+async def websocket_endpoint(
+    websocket: WebSocket, debug: bool = Depends(get_debug_flag)
+):
     await websocket.accept()
 
     # Send the tts_service value to the client
@@ -91,7 +85,25 @@ async def send_output():
             await websocket.close()
 
 
-async def main(server_host, server_port):
+async def main(server_host, server_port, profile, debug):
+
+    app.state.debug = debug
+
+    # Load the profile module from the provided path
+    spec = importlib.util.spec_from_file_location("profile", profile)
+    profile_module = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(profile_module)
+
+    # Get the interpreter from the profile
+    interpreter = profile_module.interpreter
+
+    if not hasattr(interpreter, 'tts'):
+        print("Setting TTS provider to default: openai")
+        interpreter.tts = "openai"
+
+    # Make it async
+    interpreter = AsyncInterpreter(interpreter, debug)
+
     print(f"Starting server on {server_host}:{server_port}")
     config = Config(app, host=server_host, port=server_port, lifespan="on")
     server = Server(config)

software/source/server/tunnel.py

@@ -6,7 +6,7 @@
 
 
 def create_tunnel(
-    tunnel_method="ngrok", server_host="localhost", server_port=10001, qr=False
+    tunnel_method="ngrok", server_host="localhost", server_port=10001, qr=False, domain=None
 ):
     print_markdown("Exposing server to the internet...")
 
@@ -99,8 +99,13 @@ def create_tunnel(
 
         # If ngrok is installed, start it on the specified port
         # process = subprocess.Popen(f'ngrok http {server_port} --log=stdout', shell=True, stdout=subprocess.PIPE)
+
+        if domain:
+            domain = f"--domain={domain}"
+        else:
+            domain = ""
         process = subprocess.Popen(
-            f"ngrok http {server_port} --scheme http,https  --log=stdout",
+            f"ngrok http {server_port} --scheme http,https {domain} --log=stdout",
             shell=True,
             stdout=subprocess.PIPE,
         )

software/start.py

@@ -6,6 +6,7 @@
 import importlib
 from source.server.tunnel import create_tunnel
 from source.server.async_server import main
+import subprocess
 
 import signal
 
@@ -41,6 +42,25 @@ def run(
     qr: bool = typer.Option(
         False, "--qr", help="Display QR code to scan to connect to the server"
     ),
+    domain: str = typer.Option(
+        None, "--domain", help="Connect ngrok to a custom domain"
+    ),
+    profiles: bool = typer.Option(
+        False,
+        "--profiles",
+        help="Opens the folder where this script is contained",
+    ),
+    profile: str = typer.Option(
+        "default.py", # default
+        "--profile",
+        help="Specify the path to the profile, or the name of the file if it's in the `profiles` directory (run `--profiles` to open the profiles directory)",
+    ),
+    debug: bool = typer.Option(
+        False,
+        "--debug",
+        help="Print latency measurements and save microphone recordings locally for manual playback.",
+    ),
+
 ):
     _run(
         server=server,
@@ -52,6 +72,10 @@ def run(
         server_url=server_url,
         client_type=client_type,
         qr=qr,
+        debug=debug,
+        domain=domain,
+        profiles=profiles,
+        profile=profile,
     )
 
 
@@ -65,8 +89,34 @@ def _run(
     server_url: str = None,
     client_type: str = "auto",
     qr: bool = False,
+    debug: bool = False,
+    domain = None,
+    profiles = None,
+    profile = None,
 ):
 
+    profiles_dir = os.path.join(os.path.dirname(os.path.realpath(__file__)), "source", "server", "profiles")
+
+    if profiles:
+        if platform.system() == "Windows":
+            subprocess.Popen(['explorer', profiles_dir])
+        elif platform.system() == "Darwin":
+            subprocess.Popen(['open', profiles_dir])
+        elif platform.system() == "Linux":
+            subprocess.Popen(['xdg-open', profiles_dir])
+        else:
+            subprocess.Popen(['open', profiles_dir])
+        exit(0)
+
+    if profile:
+        if not os.path.isfile(profile):
+            profile = os.path.join(profiles_dir, profile)
+            if not os.path.isfile(profile):
+                profile += ".py"
+                if not os.path.isfile(profile):
+                    print(f"Invalid profile path: {profile}")
+                    exit(1)
+
     system_type = platform.system()
     if system_type == "Windows":
         server_host = "localhost"
@@ -84,7 +134,6 @@ def handle_exit(signum, frame):
     signal.signal(signal.SIGINT, handle_exit)
 
     if server:
-        # print(f"Starting server with mobile = {mobile}")
         loop = asyncio.new_event_loop()
         asyncio.set_event_loop(loop)
         server_thread = threading.Thread(
@@ -93,14 +142,16 @@ def handle_exit(signum, frame):
                 main(
                     server_host,
                     server_port,
+                    profile,
+                    debug,
                 ),
             ),
         )
         server_thread.start()
 
     if expose:
         tunnel_thread = threading.Thread(
-            target=create_tunnel, args=[tunnel_service, server_host, server_port, qr]
+            target=create_tunnel, args=[tunnel_service, server_host, server_port, qr, domain]
         )
         tunnel_thread.start()
 
@@ -125,7 +176,7 @@ def handle_exit(signum, frame):
             f".clients.{client_type}.device", package="source"
         )
 
-        client_thread = threading.Thread(target=module.main, args=[server_url])
+        client_thread = threading.Thread(target=module.main, args=[server_url, debug])
         client_thread.start()
 
     try: