home-bro-client/app/main.py

import os
import time
import asyncio
import requests
import struct
import math
import cv2
import pvporcupine
import pyaudio
from dotenv import load_dotenv, set_key
from fastapi import FastAPI, BackgroundTasks
from fastapi.staticfiles import StaticFiles
from fastapi.responses import FileResponse
from pydantic import BaseModel
import uvicorn

# Load environment variables
load_dotenv()
ENV_PATH = ".env"

from contextlib import asynccontextmanager

@asynccontextmanager
async def lifespan(app: FastAPI):
    # Startup: Start the background loop
    start_interaction_loop()
    yield
    # Shutdown: Cleanup if needed
    if interaction_task:
        interaction_task.cancel()

app = FastAPI(title="Home-Bro Satellite", lifespan=lifespan)
app.mount("/static", StaticFiles(directory="static"), name="static")

class Config(BaseModel):
    bro_name: str
    pico_key: str
    brain_url: str
    wake_word: str
    custom_wake_word_path: str = ""
    audio_index: str = ""

@app.get("/")
async def read_index():
    return FileResponse("static/index.html")

@app.get("/favicon.ico", include_in_schema=False)
async def favicon():
    return FileResponse("static/index.html")

@app.get("/config")
async def get_config():
    return {
        "bro_name": os.getenv("BRO_NAME", "Home-Bro"),
        "pico_key": os.getenv("PICOVOICE_API_KEY", ""),
        "brain_url": os.getenv("BRAIN_URL", "http://localhost:8000"),
        "wake_word": os.getenv("WAKE_WORD", "porcupine"),
        "custom_wake_word_path": os.getenv("CUSTOM_WAKE_WORD_PATH", ""),
        "audio_index": os.getenv("AUDIO_DEVICE_INDEX", "")
    }

@app.get("/check-brain")
async def check_brain(url: str):
    try:
        response = requests.get(f"{url}/api/info", timeout=2)
        return {"online": response.status_code == 200}
    except Exception:
        return {"online": False}

def take_snapshot():
    cam_index = int(os.getenv("CAM_INDEX", 0))
    cap = cv2.VideoCapture(cam_index)
    if not cap.isOpened():
        print("Error: Could not open camera.")
        return None
    ret, frame = cap.read()
    if ret:
        path = "snapshot.jpg"
        cv2.imwrite(path, frame)
        cap.release()
        return path
    cap.release()
    return None

def send_to_brain(image_path):
    brain_url = os.getenv("BRAIN_URL")
    if not brain_url: return None
    try:
        with open(image_path, "rb") as f:
            files = {"file": f}
            response = requests.post(f"{brain_url}/analyze/pi", files=files)
            if response.status_code == 200:
                return response.json()
    except Exception as e:
        print(f"Upload failed: {e}")
    return None

def list_audio_devices():
    pa = pyaudio.PyAudio()
    print("\n--- Verfügbare Audiogeräte ---")
    for i in range(pa.get_device_count()):
        dev = pa.get_device_info_by_index(i)
        print(f"Index {i}: {dev['name']} (Input Channels: {dev['maxInputChannels']})")
    print("------------------------------\n")
    pa.terminate()

def get_wake_word_params():
    keyword = os.getenv("WAKE_WORD", "porcupine")
    custom_path = os.getenv("CUSTOM_WAKE_WORD_PATH", "")
    if keyword == "custom" and custom_path:
        return {"keyword_paths": [custom_path]}
    if keyword in pvporcupine.KEYWORDS:
        return {"keywords": [keyword]}
    return {"keywords": ["porcupine"]}

interaction_task = None

async def brain_interaction_loop():
    list_audio_devices()
    access_key = os.getenv("PICOVOICE_API_KEY")
    if not access_key:
        print("CRITICAL: PICOVOICE_API_KEY Missing.")
        return

    pa = None
    audio_stream = None
    porcupine = None

    try:
        params = get_wake_word_params()
        porcupine = pvporcupine.create(access_key=access_key, **params)
        pa = pyaudio.PyAudio()
        
        device_index = os.getenv("AUDIO_DEVICE_INDEX")
        input_device_index = int(device_index) if device_index and device_index.strip() else None
        
        audio_stream = pa.open(
            rate=porcupine.sample_rate,
            channels=1,
            format=pyaudio.paInt16,
            input=True,
            frames_per_buffer=porcupine.frame_length,
            input_device_index=input_device_index
        )

        print(f"Listening for '{os.getenv('WAKE_WORD', 'porcupine')}' on device {input_device_index}...")

        frame_count = 0
        while True:
            await asyncio.sleep(0.01)
            pcm = audio_stream.read(porcupine.frame_length, exception_on_overflow=False)
            pcm_unpacked = struct.unpack_from("h" * porcupine.frame_length, pcm)

            frame_count += 1
            if frame_count % 100 == 0:
                rms = math.sqrt(sum(x*x for x in pcm_unpacked) / len(pcm_unpacked))
                if rms < 10:
                    print(f"DEBUG: Pegel sehr niedrig ({rms:.2f}). Check Mic!")

            if porcupine.process(pcm_unpacked) >= 0:
                print(f"[{os.getenv('BRO_NAME', 'Home-Bro')}] Wake-word detected!")
                img_path = take_snapshot()
                if img_path:
                    result = send_to_brain(img_path)
                    if result:
                        print(f"Brain says: {result.get('comment')}")

    except asyncio.CancelledError:
        print("Restarting Loop...")
    except Exception as e:
        print(f"Error: {e}")
    finally:
        if audio_stream: audio_stream.close()
        if pa: pa.terminate()
        if porcupine: porcupine.delete()

def start_interaction_loop():
    global interaction_task
    interaction_task = asyncio.create_task(brain_interaction_loop())

async def restart_interaction_loop():
    global interaction_task
    if interaction_task:
        interaction_task.cancel()
        try: await interaction_task
        except asyncio.CancelledError: pass
    start_interaction_loop()

@app.post("/config")
async def save_config(config: Config, background_tasks: BackgroundTasks):
    settings = {
        "BRO_NAME": config.bro_name,
        "PICOVOICE_API_KEY": config.pico_key,
        "BRAIN_URL": config.brain_url,
        "WAKE_WORD": config.wake_word,
        "CUSTOM_WAKE_WORD_PATH": config.custom_wake_word_path,
        "AUDIO_DEVICE_INDEX": config.audio_index
    }
    for key, value in settings.items():
        os.environ[key] = value
        set_key(ENV_PATH, key, value)
    background_tasks.add_task(restart_interaction_loop)
    return {"status": "ok"}


if __name__ == "__main__":
    uvicorn.run(app, host="0.0.0.0", port=8080)