[feat] update some feature

sync code of  fastrtc,
add text support through datachannel,
fix safari connect problem
support chat without camera or mic
This commit is contained in:
huangbinchao.hbc
2025-03-25 18:05:10 +08:00
parent e1fb40a8a8
commit aefb08150f
222 changed files with 28698 additions and 5889 deletions

View File

@@ -0,0 +1,15 @@
---
title: Talk to Gemini
emoji: ♊️
colorFrom: purple
colorTo: red
sdk: gradio
sdk_version: 5.16.0
app_file: app.py
pinned: false
license: mit
short_description: Talk to Gemini using Google's multimodal API
tags: [webrtc, websocket, gradio, secret|TWILIO_ACCOUNT_SID, secret|TWILIO_AUTH_TOKEN, secret|GEMINI_API_KEY]
---
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

View File

@@ -0,0 +1,15 @@
---
title: Talk to Gemini (Gradio UI)
emoji: ♊️
colorFrom: purple
colorTo: red
sdk: gradio
sdk_version: 5.16.0
app_file: app.py
pinned: false
license: mit
short_description: Talk to Gemini (Gradio UI)
tags: [webrtc, websocket, gradio, secret|TWILIO_ACCOUNT_SID, secret|TWILIO_AUTH_TOKEN, secret|GEMINI_API_KEY]
---
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

181
demo/talk_to_gemini/app.py Normal file
View File

@@ -0,0 +1,181 @@
import asyncio
import base64
import json
import os
import pathlib
from typing import AsyncGenerator, Literal
import gradio as gr
import numpy as np
from dotenv import load_dotenv
from fastapi import FastAPI
from fastapi.responses import HTMLResponse
from fastrtc import (
AsyncStreamHandler,
Stream,
get_twilio_turn_credentials,
wait_for_item,
)
from google import genai
from google.genai.types import (
LiveConnectConfig,
PrebuiltVoiceConfig,
SpeechConfig,
VoiceConfig,
)
from gradio.utils import get_space
from pydantic import BaseModel
current_dir = pathlib.Path(__file__).parent
load_dotenv()
def encode_audio(data: np.ndarray) -> str:
"""Encode Audio data to send to the server"""
return base64.b64encode(data.tobytes()).decode("UTF-8")
class GeminiHandler(AsyncStreamHandler):
"""Handler for the Gemini API"""
def __init__(
self,
expected_layout: Literal["mono"] = "mono",
output_sample_rate: int = 24000,
output_frame_size: int = 480,
) -> None:
super().__init__(
expected_layout,
output_sample_rate,
output_frame_size,
input_sample_rate=16000,
)
self.input_queue: asyncio.Queue = asyncio.Queue()
self.output_queue: asyncio.Queue = asyncio.Queue()
self.quit: asyncio.Event = asyncio.Event()
def copy(self) -> "GeminiHandler":
return GeminiHandler(
expected_layout="mono",
output_sample_rate=self.output_sample_rate,
output_frame_size=self.output_frame_size,
)
async def start_up(self):
if not self.phone_mode:
await self.wait_for_args()
api_key, voice_name = self.latest_args[1:]
else:
api_key, voice_name = None, "Puck"
client = genai.Client(
api_key=api_key or os.getenv("GEMINI_API_KEY"),
http_options={"api_version": "v1alpha"},
)
config = LiveConnectConfig(
response_modalities=["AUDIO"], # type: ignore
speech_config=SpeechConfig(
voice_config=VoiceConfig(
prebuilt_voice_config=PrebuiltVoiceConfig(
voice_name=voice_name,
)
)
),
)
async with client.aio.live.connect(
model="gemini-2.0-flash-exp", config=config
) as session:
async for audio in session.start_stream(
stream=self.stream(), mime_type="audio/pcm"
):
if audio.data:
array = np.frombuffer(audio.data, dtype=np.int16)
self.output_queue.put_nowait((self.output_sample_rate, array))
async def stream(self) -> AsyncGenerator[bytes, None]:
while not self.quit.is_set():
try:
audio = await asyncio.wait_for(self.input_queue.get(), 0.1)
yield audio
except (asyncio.TimeoutError, TimeoutError):
pass
async def receive(self, frame: tuple[int, np.ndarray]) -> None:
_, array = frame
array = array.squeeze()
audio_message = encode_audio(array)
self.input_queue.put_nowait(audio_message)
async def emit(self) -> tuple[int, np.ndarray] | None:
return await wait_for_item(self.output_queue)
def shutdown(self) -> None:
self.quit.set()
stream = Stream(
modality="audio",
mode="send-receive",
handler=GeminiHandler(),
rtc_configuration=get_twilio_turn_credentials() if get_space() else None,
concurrency_limit=5 if get_space() else None,
time_limit=90 if get_space() else None,
additional_inputs=[
gr.Textbox(
label="API Key",
type="password",
value=os.getenv("GEMINI_API_KEY") if not get_space() else "",
),
gr.Dropdown(
label="Voice",
choices=[
"Puck",
"Charon",
"Kore",
"Fenrir",
"Aoede",
],
value="Puck",
),
],
)
class InputData(BaseModel):
webrtc_id: str
voice_name: str
api_key: str
app = FastAPI()
stream.mount(app)
@app.post("/input_hook")
async def _(body: InputData):
stream.set_input(body.webrtc_id, body.api_key, body.voice_name)
return {"status": "ok"}
@app.get("/")
async def index():
rtc_config = get_twilio_turn_credentials() if get_space() else None
html_content = (current_dir / "index.html").read_text()
html_content = html_content.replace("__RTC_CONFIGURATION__", json.dumps(rtc_config))
return HTMLResponse(content=html_content)
if __name__ == "__main__":
import os
if (mode := os.getenv("MODE")) == "UI":
stream.ui.launch(server_port=7860)
elif mode == "PHONE":
stream.fastphone(host="0.0.0.0", port=7860)
else:
import uvicorn
uvicorn.run(app, host="0.0.0.0", port=7860)

View File

@@ -0,0 +1,452 @@
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Gemini Voice Chat</title>
<style>
:root {
--color-accent: #6366f1;
--color-background: #0f172a;
--color-surface: #1e293b;
--color-text: #e2e8f0;
--boxSize: 8px;
--gutter: 4px;
}
body {
margin: 0;
padding: 0;
background-color: var(--color-background);
color: var(--color-text);
font-family: system-ui, -apple-system, sans-serif;
min-height: 100vh;
display: flex;
flex-direction: column;
align-items: center;
justify-content: center;
}
.container {
width: 90%;
max-width: 800px;
background-color: var(--color-surface);
padding: 2rem;
border-radius: 1rem;
box-shadow: 0 25px 50px -12px rgba(0, 0, 0, 0.25);
}
.wave-container {
position: relative;
display: flex;
min-height: 100px;
max-height: 128px;
justify-content: center;
align-items: center;
margin: 2rem 0;
}
.box-container {
display: flex;
justify-content: space-between;
height: 64px;
width: 100%;
}
.box {
height: 100%;
width: var(--boxSize);
background: var(--color-accent);
border-radius: 8px;
transition: transform 0.05s ease;
}
.controls {
display: grid;
gap: 1rem;
margin-bottom: 2rem;
}
.input-group {
display: flex;
flex-direction: column;
gap: 0.5rem;
}
label {
font-size: 0.875rem;
font-weight: 500;
}
input,
select {
padding: 0.75rem;
border-radius: 0.5rem;
border: 1px solid rgba(255, 255, 255, 0.1);
background-color: var(--color-background);
color: var(--color-text);
font-size: 1rem;
}
button {
padding: 1rem 2rem;
border-radius: 0.5rem;
border: none;
background-color: var(--color-accent);
color: white;
font-weight: 600;
cursor: pointer;
transition: all 0.2s ease;
}
button:hover {
opacity: 0.9;
transform: translateY(-1px);
}
.icon-with-spinner {
display: flex;
align-items: center;
justify-content: center;
gap: 12px;
min-width: 180px;
}
.spinner {
width: 20px;
height: 20px;
border: 2px solid white;
border-top-color: transparent;
border-radius: 50%;
animation: spin 1s linear infinite;
flex-shrink: 0;
}
@keyframes spin {
to {
transform: rotate(360deg);
}
}
.pulse-container {
display: flex;
align-items: center;
justify-content: center;
gap: 12px;
min-width: 180px;
}
.pulse-circle {
width: 20px;
height: 20px;
border-radius: 50%;
background-color: white;
opacity: 0.2;
flex-shrink: 0;
transform: translateX(-0%) scale(var(--audio-level, 1));
transition: transform 0.1s ease;
}
/* Add styles for toast notifications */
.toast {
position: fixed;
top: 20px;
left: 50%;
transform: translateX(-50%);
padding: 16px 24px;
border-radius: 4px;
font-size: 14px;
z-index: 1000;
display: none;
box-shadow: 0 2px 5px rgba(0, 0, 0, 0.2);
}
.toast.error {
background-color: #f44336;
color: white;
}
.toast.warning {
background-color: #ffd700;
color: black;
}
</style>
</head>
<body>
<!-- Add toast element after body opening tag -->
<div id="error-toast" class="toast"></div>
<div style="text-align: center">
<h1>Gemini Voice Chat</h1>
<p>Speak with Gemini using real-time audio streaming</p>
<p>
Get a Gemini API key
<a href="https://ai.google.dev/gemini-api/docs/api-key">here</a>
</p>
</div>
<div class="container">
<div class="controls">
<div class="input-group">
<label for="api-key">API Key</label>
<input type="password" id="api-key" placeholder="Enter your API key">
</div>
<div class="input-group">
<label for="voice">Voice</label>
<select id="voice">
<option value="Puck">Puck</option>
<option value="Charon">Charon</option>
<option value="Kore">Kore</option>
<option value="Fenrir">Fenrir</option>
<option value="Aoede">Aoede</option>
</select>
</div>
</div>
<div class="wave-container">
<div class="box-container">
<!-- Boxes will be dynamically added here -->
</div>
</div>
<button id="start-button">Start Recording</button>
</div>
<audio id="audio-output"></audio>
<script>
let peerConnection;
let audioContext;
let dataChannel;
let isRecording = false;
let webrtc_id;
const startButton = document.getElementById('start-button');
const apiKeyInput = document.getElementById('api-key');
const voiceSelect = document.getElementById('voice');
const audioOutput = document.getElementById('audio-output');
const boxContainer = document.querySelector('.box-container');
const numBars = 32;
for (let i = 0; i < numBars; i++) {
const box = document.createElement('div');
box.className = 'box';
boxContainer.appendChild(box);
}
function updateButtonState() {
if (peerConnection && (peerConnection.connectionState === 'connecting' || peerConnection.connectionState === 'new')) {
startButton.innerHTML = `
<div class="icon-with-spinner">
<div class="spinner"></div>
<span>Connecting...</span>
</div>
`;
} else if (peerConnection && peerConnection.connectionState === 'connected') {
startButton.innerHTML = `
<div class="pulse-container">
<div class="pulse-circle"></div>
<span>Stop Recording</span>
</div>
`;
} else {
startButton.innerHTML = 'Start Recording';
}
}
function showError(message) {
const toast = document.getElementById('error-toast');
toast.textContent = message;
toast.className = 'toast error';
toast.style.display = 'block';
// Hide toast after 5 seconds
setTimeout(() => {
toast.style.display = 'none';
}, 5000);
}
async function setupWebRTC() {
const config = __RTC_CONFIGURATION__;
peerConnection = new RTCPeerConnection(config);
webrtc_id = Math.random().toString(36).substring(7);
const timeoutId = setTimeout(() => {
const toast = document.getElementById('error-toast');
toast.textContent = "Connection is taking longer than usual. Are you on a VPN?";
toast.className = 'toast warning';
toast.style.display = 'block';
// Hide warning after 5 seconds
setTimeout(() => {
toast.style.display = 'none';
}, 5000);
}, 5000);
try {
const stream = await navigator.mediaDevices.getUserMedia({ audio: true });
stream.getTracks().forEach(track => peerConnection.addTrack(track, stream));
// Update audio visualization setup
audioContext = new AudioContext();
analyser_input = audioContext.createAnalyser();
const source = audioContext.createMediaStreamSource(stream);
source.connect(analyser_input);
analyser_input.fftSize = 64;
dataArray_input = new Uint8Array(analyser_input.frequencyBinCount);
function updateAudioLevel() {
analyser_input.getByteFrequencyData(dataArray_input);
const average = Array.from(dataArray_input).reduce((a, b) => a + b, 0) / dataArray_input.length;
const audioLevel = average / 255;
const pulseCircle = document.querySelector('.pulse-circle');
if (pulseCircle) {
console.log("audioLevel", audioLevel);
pulseCircle.style.setProperty('--audio-level', 1 + audioLevel);
}
animationId = requestAnimationFrame(updateAudioLevel);
}
updateAudioLevel();
// Add connection state change listener
peerConnection.addEventListener('connectionstatechange', () => {
console.log('connectionstatechange', peerConnection.connectionState);
if (peerConnection.connectionState === 'connected') {
clearTimeout(timeoutId);
const toast = document.getElementById('error-toast');
toast.style.display = 'none';
}
updateButtonState();
});
// Handle incoming audio
peerConnection.addEventListener('track', (evt) => {
if (audioOutput && audioOutput.srcObject !== evt.streams[0]) {
audioOutput.srcObject = evt.streams[0];
audioOutput.play();
// Set up audio visualization on the output stream
audioContext = new AudioContext();
analyser = audioContext.createAnalyser();
const source = audioContext.createMediaStreamSource(evt.streams[0]);
source.connect(analyser);
analyser.fftSize = 2048;
dataArray = new Uint8Array(analyser.frequencyBinCount);
updateVisualization();
}
});
// Create data channel for messages
dataChannel = peerConnection.createDataChannel('text');
dataChannel.onmessage = (event) => {
const eventJson = JSON.parse(event.data);
if (eventJson.type === "error") {
showError(eventJson.message);
} else if (eventJson.type === "send_input") {
fetch('/input_hook', {
method: 'POST',
headers: {
'Content-Type': 'application/json',
},
body: JSON.stringify({
webrtc_id: webrtc_id,
api_key: apiKeyInput.value,
voice_name: voiceSelect.value
})
});
}
};
// Create and send offer
const offer = await peerConnection.createOffer();
await peerConnection.setLocalDescription(offer);
await new Promise((resolve) => {
if (peerConnection.iceGatheringState === "complete") {
resolve();
} else {
const checkState = () => {
if (peerConnection.iceGatheringState === "complete") {
peerConnection.removeEventListener("icegatheringstatechange", checkState);
resolve();
}
};
peerConnection.addEventListener("icegatheringstatechange", checkState);
}
});
const response = await fetch('/webrtc/offer', {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({
sdp: peerConnection.localDescription.sdp,
type: peerConnection.localDescription.type,
webrtc_id: webrtc_id,
})
});
const serverResponse = await response.json();
if (serverResponse.status === 'failed') {
showError(serverResponse.meta.error === 'concurrency_limit_reached'
? `Too many connections. Maximum limit is ${serverResponse.meta.limit}`
: serverResponse.meta.error);
stop();
startButton.textContent = 'Start Recording';
return;
}
await peerConnection.setRemoteDescription(serverResponse);
} catch (err) {
clearTimeout(timeoutId);
console.error('Error setting up WebRTC:', err);
showError('Failed to establish connection. Please try again.');
stop();
startButton.textContent = 'Start Recording';
}
}
function updateVisualization() {
if (!analyser) return;
analyser.getByteFrequencyData(dataArray);
const bars = document.querySelectorAll('.box');
for (let i = 0; i < bars.length; i++) {
const barHeight = (dataArray[i] / 255) * 2;
bars[i].style.transform = `scaleY(${Math.max(0.1, barHeight)})`;
}
animationId = requestAnimationFrame(updateVisualization);
}
function stopWebRTC() {
if (peerConnection) {
peerConnection.close();
}
if (animationId) {
cancelAnimationFrame(animationId);
}
if (audioContext) {
audioContext.close();
}
updateButtonState();
}
startButton.addEventListener('click', () => {
if (!isRecording) {
setupWebRTC();
startButton.classList.add('recording');
} else {
stopWebRTC();
startButton.classList.remove('recording');
}
isRecording = !isRecording;
});
</script>
</body>
</html>

View File

@@ -0,0 +1,4 @@
fastrtc
python-dotenv
google-genai
twilio