awacke1's picture
Update app.py
a8499f4 verified
raw
history blame
38.7 kB
import streamlit as st
import asyncio
import websockets
import uuid
import argparse
from datetime import datetime
import os
import random
import time
import hashlib
from PIL import Image
import glob
import base64
import io
import streamlit.components.v1 as components
import edge_tts
from audio_recorder_streamlit import audio_recorder
import nest_asyncio
import re
from streamlit_paste_button import paste_image_button
import pytz
import shutil
import wave
import audioop
# Patch for nested async
nest_asyncio.apply()
# Static config
icons = 'πŸ€–πŸ§ πŸ”¬πŸ“'
START_ROOM = "Sector 🌌"
# Page setup
st.set_page_config(
page_title="πŸ€–πŸ§ MMO Chat BrainπŸ“πŸ”¬",
page_icon=icons,
layout="wide",
initial_sidebar_state="auto"
)
# Funky usernames with voices
FUN_USERNAMES = {
"CosmicJester 🌌": "en-US-AriaNeural",
"PixelPanda 🐼": "en-US-JennyNeural",
"QuantumQuack πŸ¦†": "en-GB-SoniaNeural",
"StellarSquirrel 🐿️": "en-AU-NatashaNeural",
"GizmoGuru βš™οΈ": "en-CA-ClaraNeural",
"NebulaNinja 🌠": "en-US-GuyNeural",
"ByteBuster πŸ’Ύ": "en-GB-RyanNeural",
"GalacticGopher 🌍": "en-AU-WilliamNeural",
"RocketRaccoon πŸš€": "en-CA-LiamNeural",
"EchoElf 🧝": "en-US-AnaNeural",
"PhantomFox 🦊": "en-US-BrandonNeural",
"WittyWizard πŸ§™": "en-GB-ThomasNeural",
"LunarLlama πŸŒ™": "en-AU-FreyaNeural",
"SolarSloth β˜€οΈ": "en-CA-LindaNeural",
"AstroAlpaca πŸ¦™": "en-US-ChristopherNeural",
"CyberCoyote 🐺": "en-GB-ElliotNeural",
"MysticMoose 🦌": "en-AU-JamesNeural",
"GlitchGnome 🧚": "en-CA-EthanNeural",
"VortexViper 🐍": "en-US-AmberNeural",
"ChronoChimp πŸ’": "en-GB-LibbyNeural"
}
# Folders
CHAT_DIR = "chat_logs"
VOTE_DIR = "vote_logs"
STATE_FILE = "user_state.txt"
AUDIO_DIR = "audio_logs"
HISTORY_DIR = "history_logs"
MEDIA_DIR = "media_files"
os.makedirs(CHAT_DIR, exist_ok=True)
os.makedirs(VOTE_DIR, exist_ok=True)
os.makedirs(AUDIO_DIR, exist_ok=True)
os.makedirs(HISTORY_DIR, exist_ok=True)
os.makedirs(MEDIA_DIR, exist_ok=True)
CHAT_FILE = os.path.join(CHAT_DIR, "global_chat.md")
QUOTE_VOTES_FILE = os.path.join(VOTE_DIR, "quote_votes.md")
MEDIA_VOTES_FILE = os.path.join(VOTE_DIR, "media_votes.md")
HISTORY_FILE = os.path.join(HISTORY_DIR, "chat_history.md")
# Unicode digits
UNICODE_DIGITS = {i: f"{i}\uFE0F⃣" for i in range(10)}
# Font collection
UNICODE_FONTS = [
("Normal", lambda x: x),
("Bold", lambda x: "".join(chr(ord(c) + 0x1D400 - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D41A - 0x61) if 'a' <= c <= 'z' else c for c in x)),
("Italic", lambda x: "".join(chr(ord(c) + 0x1D434 - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D44E - 0x61) if 'a' <= c <= 'z' else c for c in x)),
("Bold Italic", lambda x: "".join(chr(ord(c) + 0x1D468 - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D482 - 0x61) if 'a' <= c <= 'z' else c for c in x)),
("Script", lambda x: "".join(chr(ord(c) + 0x1D49C - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D4B6 - 0x61) if 'a' <= c <= 'z' else c for c in x)),
("Bold Script", lambda x: "".join(chr(ord(c) + 0x1D4D0 - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D4EA - 0x61) if 'a' <= c <= 'z' else c for c in x)),
("Fraktur", lambda x: "".join(chr(ord(c) + 0x1D504 - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D51E - 0x61) if 'a' <= c <= 'z' else c for c in x)),
("Bold Fraktur", lambda x: "".join(chr(ord(c) + 0x1D56C - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D586 - 0x61) if 'a' <= c <= 'z' else c for c in x)),
("Double Struck", lambda x: "".join(chr(ord(c) + 0x1D538 - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D552 - 0x61) if 'a' <= c <= 'z' else c for c in x)),
("Sans Serif", lambda x: "".join(chr(ord(c) + 0x1D5A0 - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D5BA - 0x61) if 'a' <= c <= 'z' else c for c in x)),
("Sans Serif Bold", lambda x: "".join(chr(ord(c) + 0x1D5D4 - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D5EE - 0x61) if 'a' <= c <= 'z' else c for c in x)),
("Sans Serif Italic", lambda x: "".join(chr(ord(c) + 0x1D608 - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D622 - 0x61) if 'a' <= c <= 'z' else c for c in x)),
("Sans Serif Bold Italic", lambda x: "".join(chr(ord(c) + 0x1D63C - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D656 - 0x61) if 'a' <= c <= 'z' else c for c in x)),
("Monospace", lambda x: "".join(chr(ord(c) + 0x1D670 - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D68A - 0x61) if 'a' <= c <= 'z' else c for c in x)),
("Circled", lambda x: "".join(chr(ord(c) - 0x41 + 0x24B6) if 'A' <= c <= 'Z' else chr(ord(c) - 0x61 + 0x24D0) if 'a' <= c <= 'z' else c for c in x)),
("Squared", lambda x: "".join(chr(ord(c) - 0x41 + 0x1F130) if 'A' <= c <= 'Z' else c for c in x)),
("Negative Circled", lambda x: "".join(chr(ord(c) - 0x41 + 0x1F150) if 'A' <= c <= 'Z' else c for c in x)),
("Negative Squared", lambda x: "".join(chr(ord(c) - 0x41 + 0x1F170) if 'A' <= c <= 'Z' else c for c in x)),
("Regional Indicator", lambda x: "".join(chr(ord(c) - 0x41 + 0x1F1E6) if 'A' <= c <= 'Z' else c for c in x)),
]
# Global state
if 'server_running' not in st.session_state:
st.session_state.server_running = False
if 'server_task' not in st.session_state:
st.session_state.server_task = None
if 'active_connections' not in st.session_state:
st.session_state.active_connections = {}
if 'media_notifications' not in st.session_state:
st.session_state.media_notifications = []
if 'last_chat_update' not in st.session_state:
st.session_state.last_chat_update = 0
if 'displayed_chat_lines' not in st.session_state:
st.session_state.displayed_chat_lines = []
if 'old_val' not in st.session_state:
st.session_state.old_val = ""
if 'last_query' not in st.session_state:
st.session_state.last_query = ""
if 'message_text' not in st.session_state:
st.session_state.message_text = ""
if 'audio_cache' not in st.session_state:
st.session_state.audio_cache = {}
if 'pasted_image_data' not in st.session_state:
st.session_state.pasted_image_data = None
if 'quote_line' not in st.session_state:
st.session_state.quote_line = None
if 'refresh_rate' not in st.session_state:
st.session_state.refresh_rate = 5
if 'base64_cache' not in st.session_state:
st.session_state.base64_cache = {}
if 'transcript_history' not in st.session_state:
st.session_state.transcript_history = []
if 'last_transcript' not in st.session_state:
st.session_state.last_transcript = ""
if 'image_hashes' not in st.session_state:
st.session_state.image_hashes = set()
# Utility functions
def format_timestamp_prefix(username):
central = pytz.timezone('US/Central')
now = datetime.now(central)
return f"{now.strftime('%I-%M-%p-ct-%m-%d-%Y')}-by-{username}"
def compute_image_hash(image_data):
if isinstance(image_data, Image.Image):
img_byte_arr = io.BytesIO()
image_data.save(img_byte_arr, format='PNG')
img_bytes = img_byte_arr.getvalue()
else:
img_bytes = image_data
return hashlib.md5(img_bytes).hexdigest()[:8]
def get_node_name():
parser = argparse.ArgumentParser(description='Start a chat node with a specific name')
parser.add_argument('--node-name', type=str, default=None)
parser.add_argument('--port', type=int, default=8501)
args = parser.parse_args()
username = st.session_state.get('username', 'System 🌟')
log_action(username, "🌐🍼 - Node naming - christening the beast!")
return args.node_name or f"node-{uuid.uuid4().hex[:8]}", args.port
def log_action(username, action):
if 'action_log' not in st.session_state:
st.session_state.action_log = {}
user_log = st.session_state.action_log.setdefault(username, {})
current_time = time.time()
user_log = {k: v for k, v in user_log.items() if current_time - v < 10}
st.session_state.action_log[username] = user_log
if action not in user_log:
central = pytz.timezone('US/Central')
with open(HISTORY_FILE, 'a') as f:
f.write(f"[{datetime.now(central).strftime('%Y-%m-%d %H:%M:%S')}] {username}: {action}\n")
user_log[action] = current_time
def clean_text_for_tts(text):
cleaned = re.sub(r'[#*!\[\]]+', '', text)
cleaned = ' '.join(cleaned.split())
return cleaned[:200] if cleaned else "No text to speak"
async def save_chat_entry(username, message, is_markdown=False):
await asyncio.to_thread(log_action, username, "πŸ’¬πŸ”’ - Chat saver - words locked tight!")
central = pytz.timezone('US/Central')
timestamp = datetime.now(central).strftime("%Y-%m-%d %H:%M:%S")
if is_markdown:
entry = f"[{timestamp}] {username}:\n```markdown\n{message}\n```"
else:
entry = f"[{timestamp}] {username}: {message}"
await asyncio.to_thread(lambda: open(CHAT_FILE, 'a').write(f"{entry}\n"))
voice = FUN_USERNAMES.get(username, "en-US-AriaNeural")
cleaned_message = clean_text_for_tts(message)
audio_file = await async_edge_tts_generate(cleaned_message, voice)
if audio_file:
with open(HISTORY_FILE, 'a') as f:
f.write(f"[{timestamp}] {username}: Audio generated - {audio_file}\n")
await broadcast_message(f"{username}|{message}", "chat")
st.session_state.last_chat_update = time.time()
return audio_file
async def load_chat():
username = st.session_state.get('username', 'System 🌟')
await asyncio.to_thread(log_action, username, "πŸ“œπŸš€ - Chat loader - history unleashed!")
if not os.path.exists(CHAT_FILE):
await asyncio.to_thread(lambda: open(CHAT_FILE, 'a').write(f"# {START_ROOM} Chat\n\nWelcome to the cosmic hub - start chatting! 🎀\n"))
with open(CHAT_FILE, 'r') as f:
content = await asyncio.to_thread(f.read)
return content
async def get_user_list(chat_content):
username = st.session_state.get('username', 'System 🌟')
await asyncio.to_thread(log_action, username, "πŸ‘₯πŸŽ‰ - User lister - who’s in the gang!")
users = set()
for line in chat_content.split('\n'):
if line.strip() and ': ' in line:
user = line.split(': ')[1].split(' ')[0]
users.add(user)
return sorted(list(users))
async def has_joined_before(client_id, chat_content):
username = st.session_state.get('username', 'System 🌟')
await asyncio.to_thread(log_action, username, "πŸšͺπŸ” - Join checker - been here before?")
return any(f"Client-{client_id}" in line for line in chat_content.split('\n'))
async def get_message_suggestions(chat_content, prefix):
username = st.session_state.get('username', 'System 🌟')
await asyncio.to_thread(log_action, username, "πŸ’‘πŸ“ - Suggestion maker - old quips resurface!")
lines = chat_content.split('\n')
messages = [line.split(': ', 1)[1] for line in lines if ': ' in line and line.strip()]
return [msg for msg in messages if msg.lower().startswith(prefix.lower())][:5]
async def save_vote(file, item, user_hash, username, comment=""):
await asyncio.to_thread(log_action, username, "πŸ‘πŸ“Š - Vote saver - cheers recorded!")
central = pytz.timezone('US/Central')
timestamp = datetime.now(central).strftime("%Y-%m-%d %H:%M:%S")
entry = f"[{timestamp}] {user_hash} voted for {item}"
await asyncio.to_thread(lambda: open(file, 'a').write(f"{entry}\n"))
await asyncio.to_thread(lambda: open(HISTORY_FILE, "a").write(f"- {timestamp} - User {user_hash} voted for {item}\n"))
chat_message = f"{username} upvoted: \"{item}\""
if comment:
chat_message += f" - {comment}"
await save_chat_entry(username, chat_message)
async def load_votes(file):
username = st.session_state.get('username', 'System 🌟')
await asyncio.to_thread(log_action, username, "πŸ†πŸ“ˆ - Vote counter - tallying the love!")
if not os.path.exists(file):
await asyncio.to_thread(lambda: open(file, 'w').write("# Vote Tally\n\nNo votes yet - get clicking! πŸ–±οΈ\n"))
with open(file, 'r') as f:
content = await asyncio.to_thread(f.read)
lines = content.strip().split('\n')[2:]
votes = {}
user_votes = set()
for line in lines:
if line.strip() and 'voted for' in line:
user_hash = line.split('] ')[1].split(' voted for ')[0]
item = line.split('voted for ')[1]
vote_key = f"{user_hash}-{item}"
if vote_key not in user_votes:
votes[item] = votes.get(item, 0) + 1
user_votes.add(vote_key)
return votes
async def generate_user_hash():
username = st.session_state.get('username', 'System 🌟')
await asyncio.to_thread(log_action, username, "πŸ”‘πŸ•΅οΈ - Hash generator - secret codes ahoy!")
if 'user_hash' not in st.session_state:
st.session_state.user_hash = hashlib.md5(str(random.getrandbits(128)).encode()).hexdigest()[:8]
return st.session_state.user_hash
async def async_edge_tts_generate(text, voice, rate=0, pitch=0, file_format="mp3"):
username = st.session_state.get('username', 'System 🌟')
await asyncio.to_thread(log_action, username, "🎢🌟 - Audio maker - voices come alive!")
timestamp = format_timestamp_prefix(username)
filename = f"{timestamp}.{file_format}"
filepath = os.path.join(AUDIO_DIR, filename)
communicate = edge_tts.Communicate(text, voice, rate=f"{rate:+d}%", pitch=f"{pitch:+d}Hz")
try:
await communicate.save(filepath)
return filepath if os.path.exists(filepath) else None
except edge_tts.exceptions.NoAudioReceived:
with open(HISTORY_FILE, 'a') as f:
central = pytz.timezone('US/Central')
f.write(f"[{datetime.now(central).strftime('%Y-%m-%d %H:%M:%S')}] {username}: Audio failed - No audio received for '{text}'\n")
return None
def play_and_download_audio(file_path):
if file_path and os.path.exists(file_path):
st.audio(file_path)
if file_path not in st.session_state.base64_cache:
with open(file_path, "rb") as f:
b64 = base64.b64encode(f.read()).decode()
st.session_state.base64_cache[file_path] = b64
b64 = st.session_state.base64_cache[file_path]
dl_link = f'<a href="data:audio/mpeg;base64,{b64}" download="{os.path.basename(file_path)}">🎡 Download {os.path.basename(file_path)}</a>'
st.markdown(dl_link, unsafe_allow_html=True)
async def save_pasted_image(image, username):
await asyncio.to_thread(log_action, username, "πŸ“ΈπŸ’Ύ - Image saver - pics preserved!")
img_hash = compute_image_hash(image)
if img_hash in st.session_state.image_hashes:
return None
timestamp = format_timestamp_prefix(username)
filename = f"{timestamp}-{img_hash}.png"
filepath = os.path.join(MEDIA_DIR, filename)
await asyncio.to_thread(image.save, filepath, "PNG")
st.session_state.image_hashes.add(img_hash)
return filepath
async def get_video_html(video_path, width="100%"):
username = st.session_state.get('username', 'System 🌟')
await asyncio.to_thread(log_action, username, "πŸŽ₯🎬 - Video renderer - movies roll!")
with open(video_path, 'rb') as f:
video_data = await asyncio.to_thread(f.read)
video_url = f"data:video/mp4;base64,{base64.b64encode(video_data).decode()}"
return f'<video width="{width}" controls autoplay><source src="{video_url}" type="video/mp4">Your browser does not support the video tag.</video>'
async def get_audio_html(audio_path, width="100%"):
username = st.session_state.get('username', 'System 🌟')
await asyncio.to_thread(log_action, username, "🎢✈️ - Audio renderer - sounds soar!")
audio_url = f"data:audio/mpeg;base64,{base64.b64encode(await asyncio.to_thread(open, audio_path, 'rb').read()).decode()}"
return f'<audio controls style="width: {width};"><source src="{audio_url}" type="audio/mpeg">Your browser does not support the audio element.</audio>'
async def websocket_handler(websocket, path):
username = st.session_state.get('username', 'System 🌟')
await asyncio.to_thread(log_action, username, "πŸŒπŸ”— - Websocket handler - chat links up!")
try:
client_id = str(uuid.uuid4())
room_id = "chat"
st.session_state.active_connections.setdefault(room_id, {})[client_id] = websocket
chat_content = await load_chat()
username = st.session_state.get('username', random.choice(list(FUN_USERNAMES.keys())))
if not await has_joined_before(client_id, chat_content):
await save_chat_entry(f"Client-{client_id}", f"{username} has joined {START_ROOM}!")
async for message in websocket:
parts = message.split('|', 1)
if len(parts) == 2:
username, content = parts
await save_chat_entry(username, content)
except websockets.ConnectionClosed:
pass
finally:
if room_id in st.session_state.active_connections and client_id in st.session_state.active_connections[room_id]:
del st.session_state.active_connections[room_id][client_id]
async def broadcast_message(message, room_id):
username = st.session_state.get('username', 'System 🌟')
await asyncio.to_thread(log_action, username, "πŸ“’βœˆοΈ - Message broadcaster - words fly far!")
if room_id in st.session_state.active_connections:
disconnected = []
for client_id, ws in st.session_state.active_connections[room_id].items():
try:
await ws.send(message)
except websockets.ConnectionClosed:
disconnected.append(client_id)
for client_id in disconnected:
del st.session_state.active_connections[room_id][client_id]
async def run_websocket_server():
username = st.session_state.get('username', 'System 🌟')
await asyncio.to_thread(log_action, username, "πŸ–₯οΈπŸŒ€ - Server starter - web spins up!")
if not st.session_state.server_running:
server = await websockets.serve(websocket_handler, '0.0.0.0', 8765)
st.session_state.server_running = True
await server.wait_closed()
async def save_audio_recording(audio_bytes, username, voice):
timestamp = format_timestamp_prefix(username)
voice_id = voice.split('-')[-1].lower()
filename = f"rec_{username}_{voice_id}_{timestamp}.mp3"
filepath = os.path.join(AUDIO_DIR, filename)
with open(filepath, 'wb') as f:
f.write(audio_bytes)
cleaned_text = "Audio recording"
audio_file = await async_edge_tts_generate(cleaned_text, voice, file_format="mp3")
if audio_file:
os.rename(audio_file, filepath)
return filepath
async def process_voice_input(audio_bytes, username, voice):
await asyncio.to_thread(log_action, username, "πŸŽ€πŸ“ - Voice processor - speech to text!")
if audio_bytes:
audio_file = await save_audio_recording(audio_bytes, username, voice)
text = f"Voice recording saved: {os.path.basename(audio_file)}"
await save_chat_entry(username, text)
return audio_file
return None
async def perform_ai_lookup(query, vocal_summary=True, extended_refs=False, titles_summary=True, full_audio=False, useArxiv=True, useArxivAudio=False):
username = st.session_state.get('username', 'System 🌟')
result = f"AI Lookup Result for '{query}' (Arxiv: {useArxiv}, Audio: {useArxivAudio})"
await save_chat_entry(username, result)
if useArxivAudio:
audio_file = await async_edge_tts_generate(result, FUN_USERNAMES.get(username, "en-US-AriaNeural"))
if audio_file:
st.audio(audio_file)
def delete_user_files():
protected_files = {'app.py', 'requirements.txt', 'README.md'}
deleted_files = []
directories = [MEDIA_DIR, AUDIO_DIR, CHAT_DIR, VOTE_DIR, HISTORY_DIR]
for directory in directories:
if os.path.exists(directory):
for root, _, files in os.walk(directory):
for file in files:
file_path = os.path.join(root, file)
if os.path.basename(file_path) not in protected_files:
try:
os.remove(file_path)
deleted_files.append(file_path)
except Exception as e:
st.error(f"Failed to delete {file_path}: {e}")
try:
shutil.rmtree(directory, ignore_errors=True)
os.makedirs(directory, exist_ok=True)
except Exception as e:
st.error(f"Failed to remove directory {directory}: {e}")
st.session_state.image_hashes.clear()
st.session_state.audio_cache.clear()
st.session_state.base64_cache.clear()
st.session_state.displayed_chat_lines.clear()
return deleted_files
ASR_HTML = """
<html>
<head>
<title>Continuous Speech Demo</title>
<style>
body {
font-family: sans-serif;
padding: 20px;
max-width: 800px;
margin: 0 auto;
}
button {
padding: 10px 20px;
margin: 10px 5px;
font-size: 16px;
}
#status {
margin: 10px 0;
padding: 10px;
background: #e8f5e9;
border-radius: 4px;
}
#output {
white-space: pre-wrap;
padding: 15px;
background: #f5f5f5;
border-radius: 4px;
margin: 10px 0;
min-height: 100px;
max-height: 400px;
overflow-y: auto;
}
.controls {
margin: 10px 0;
}
</style>
</head>
<body>
<div class="controls">
<button id="start">Start Listening</button>
<button id="stop" disabled>Stop Listening</button>
<button id="clear">Clear Text</button>
</div>
<div id="status">Ready</div>
<div id="output"></div>
<script>
if (!('webkitSpeechRecognition' in window)) {
alert('Speech recognition not supported');
} else {
const recognition = new webkitSpeechRecognition();
const startButton = document.getElementById('start');
const stopButton = document.getElementById('stop');
const clearButton = document.getElementById('clear');
const status = document.getElementById('status');
const output = document.getElementById('output');
let fullTranscript = '';
let lastUpdateTime = Date.now();
recognition.continuous = true;
recognition.interimResults = true;
const startRecognition = () => {
try {
recognition.start();
status.textContent = 'Listening...';
startButton.disabled = true;
stopButton.disabled = false;
} catch (e) {
console.error(e);
status.textContent = 'Error: ' + e.message;
}
};
window.addEventListener('load', () => {
setTimeout(startRecognition, 1000);
});
startButton.onclick = startRecognition;
stopButton.onclick = () => {
recognition.stop();
status.textContent = 'Stopped';
startButton.disabled = false;
stopButton.disabled = true;
if (fullTranscript) {
sendDataToPython({
value: fullTranscript,
dataType: "json",
stopped: true
});
}
};
clearButton.onclick = () => {
fullTranscript = '';
output.textContent = '';
sendDataToPython({value: '', dataType: "json"});
};
recognition.onresult = (event) => {
let interimTranscript = '';
let finalTranscript = '';
for (let i = event.resultIndex; i < event.results.length; i++) {
const transcript = event.results[i][0].transcript;
if (event.results[i].isFinal) {
finalTranscript += transcript + '\\n';
} else {
interimTranscript += transcript;
}
}
if (finalTranscript || (Date.now() - lastUpdateTime > 5000)) {
if (finalTranscript) {
fullTranscript += finalTranscript;
}
lastUpdateTime = Date.now();
output.textContent = fullTranscript + (interimTranscript ? '... ' + interimTranscript : '');
output.scrollTop = output.scrollHeight;
sendDataToPython({
value: fullTranscript,
dataType: "json",
stopped: false
});
}
};
recognition.onend = () => {
if (!stopButton.disabled) {
try {
recognition.start();
console.log('Restarted recognition');
} catch (e) {
console.error('Failed to restart recognition:', e);
status.textContent = 'Error restarting: ' + e.message;
startButton.disabled = false;
stopButton.disabled = true;
}
}
};
recognition.onerror = (event) => {
console.error('Recognition error:', event.error);
status.textContent = 'Error: ' + event.error;
if (event.error === 'not-allowed' || event.error === 'service-not-allowed') {
startButton.disabled = false;
stopButton.disabled = true;
}
};
}
function sendDataToPython(data) {
window.parent.postMessage({
isStreamlitMessage: true,
type: "streamlit:setComponentValue",
...data
}, "*");
}
window.addEventListener('load', function() {
window.setTimeout(function() {
window.parent.postMessage({
isStreamlitMessage: true,
type: "streamlit:setFrameHeight",
height: document.documentElement.clientHeight
}, "*");
}, 0);
});
</script>
</body>
</html>
"""
def main():
NODE_NAME, port = get_node_name()
loop = asyncio.new_event_loop()
asyncio.set_event_loop(loop)
async def async_interface():
if 'username' not in st.session_state:
chat_content = await load_chat()
available_names = [name for name in FUN_USERNAMES if not any(f"{name} has joined" in line for line in chat_content.split('\n'))]
st.session_state.username = random.choice(available_names) if available_names else random.choice(list(FUN_USERNAMES.keys()))
st.session_state.voice = FUN_USERNAMES[st.session_state.username]
st.markdown(f"**πŸŽ™οΈ Voice Selected**: {st.session_state.voice} πŸ—£οΈ for {st.session_state.username}")
st.title(f"πŸ€–πŸ§ MMO {st.session_state.username}πŸ“πŸ”¬")
st.markdown(f"Welcome to {START_ROOM} - chat, vote, upload, paste images, and enjoy quoting! πŸŽ‰")
if not st.session_state.server_task:
st.session_state.server_task = loop.create_task(run_websocket_server())
audio_bytes = audio_recorder()
if audio_bytes:
audio_file = await process_voice_input(audio_bytes, st.session_state.username, st.session_state.voice)
if audio_file:
st.audio(audio_file)
st.rerun()
st.subheader("🎀 Continuous Speech Input")
asr_component = components.html(ASR_HTML, height=400)
if asr_component and isinstance(asr_component, dict) and 'value' in asr_component:
transcript = asr_component['value'].strip()
stopped = asr_component.get('stopped', False)
if transcript and transcript != st.session_state.last_transcript:
st.session_state.transcript_history.append(transcript)
st.session_state.last_transcript = transcript
if stopped:
audio_file = await save_audio_recording(audio_bytes, st.session_state.username, st.session_state.voice)
await save_chat_entry(st.session_state.username, f"Voice message: {transcript}\nAudio file: {os.path.basename(audio_file)}", is_markdown=True)
st.rerun()
st.subheader("🎡 Recorded Audio Files")
audio_files = glob.glob(f"{AUDIO_DIR}/rec_{st.session_state.username}_*.mp3")
if audio_files:
st.write(f"Found {len(audio_files)} recordings for {st.session_state.username}")
for audio_file in sorted(audio_files, key=os.path.getmtime, reverse=True):
col1, col2 = st.columns([3, 1])
with col1:
st.audio(audio_file)
st.write(f"File: {os.path.basename(audio_file)}")
with col2:
if audio_file not in st.session_state.base64_cache:
with open(audio_file, "rb") as f:
b64 = base64.b64encode(f.read()).decode()
st.session_state.base64_cache[audio_file] = b64
b64 = st.session_state.base64_cache[audio_file]
dl_link = f'<a href="data:audio/mpeg;base64,{b64}" download="{os.path.basename(audio_file)}">🎡 Download</a>'
st.markdown(dl_link, unsafe_allow_html=True)
st.subheader(f"{START_ROOM} Chat πŸ’¬")
chat_content = await load_chat()
chat_lines = chat_content.split('\n')
chat_votes = await load_votes(QUOTE_VOTES_FILE)
current_time = time.time()
if current_time - st.session_state.last_chat_update > 1 or not st.session_state.displayed_chat_lines:
new_lines = [line for line in chat_lines if line.strip() and ': ' in line and line not in st.session_state.displayed_chat_lines and not line.startswith('#')]
st.session_state.displayed_chat_lines.extend(new_lines)
st.session_state.last_chat_update = current_time
for i, line in enumerate(st.session_state.displayed_chat_lines):
col1, col2, col3, col4 = st.columns([3, 1, 1, 2])
with col1:
if "```markdown" in line:
markdown_content = re.search(r'```markdown\n(.*?)```', line, re.DOTALL)
if markdown_content:
st.markdown(markdown_content.group(1))
else:
st.markdown(line)
else:
st.markdown(line)
if "Pasted image:" in line or "Uploaded media:" in line:
file_path = line.split(': ')[-1].strip()
if os.path.exists(file_path):
if file_path not in st.session_state.base64_cache:
with open(file_path, "rb") as f:
b64 = base64.b64encode(f.read()).decode()
st.session_state.base64_cache[file_path] = b64
b64 = st.session_state.base64_cache[file_path]
mime_type = "image/png" if file_path.endswith(('.png', '.jpg')) else "video/mp4" if file_path.endswith('.mp4') else "audio/mpeg"
if file_path.endswith(('.png', '.jpg')):
st.image(file_path, use_container_width=True)
dl_link = f'<a href="data:{mime_type};base64,{b64}" download="{os.path.basename(file_path)}">πŸ“₯ Download {os.path.basename(file_path)}</a>'
st.markdown(dl_link, unsafe_allow_html=True)
elif file_path.endswith('.mp4'):
st.markdown(await get_video_html(file_path), unsafe_allow_html=True)
dl_link = f'<a href="data:{mime_type};base64,{b64}" download="{os.path.basename(file_path)}">πŸ“₯ Download {os.path.basename(file_path)}</a>'
st.markdown(dl_link, unsafe_allow_html=True)
elif file_path.endswith('.mp3'):
st.markdown(await get_audio_html(file_path), unsafe_allow_html=True)
dl_link = f'<a href="data:{mime_type};base64,{b64}" download="{os.path.basename(file_path)}">πŸ“₯ Download {os.path.basename(file_path)}</a>'
st.markdown(dl_link, unsafe_allow_html=True)
with col2:
vote_count = chat_votes.get(line.split('. ')[1] if '. ' in line else line, 0)
if st.button(f"πŸ‘ {vote_count}", key=f"chat_vote_{i}"):
comment = st.session_state.message_text
await save_vote(QUOTE_VOTES_FILE, line.split('. ')[1] if '. ' in line else line, await generate_user_hash(), st.session_state.username, comment)
st.session_state.message_text = ''
st.rerun()
with col3:
if st.button("πŸ“’ Quote", key=f"quote_{i}"):
st.session_state.quote_line = line
st.rerun()
with col4:
username = line.split(': ')[1].split(' ')[0]
cache_key = f"{line}_{FUN_USERNAMES.get(username, 'en-US-AriaNeural')}"
if cache_key not in st.session_state.audio_cache:
cleaned_text = clean_text_for_tts(line.split(': ', 1)[1])
audio_file = await async_edge_tts_generate(cleaned_text, FUN_USERNAMES.get(username, "en-US-AriaNeural"))
st.session_state.audio_cache[cache_key] = audio_file
audio_file = st.session_state.audio_cache.get(cache_key)
if audio_file:
play_and_download_audio(audio_file)
if st.session_state.quote_line:
st.markdown(f"### Quoting: {st.session_state.quote_line}")
quote_response = st.text_area("Add your response", key="quote_response", value=st.session_state.message_text)
paste_result_quote = paste_image_button("πŸ“‹ Paste Image or Text with Quote", key="paste_button_quote")
if paste_result_quote.image_data is not None:
if isinstance(paste_result_quote.image_data, str):
st.session_state.message_text = paste_result_quote.image_data
st.text_area("Add your response", key="quote_response", value=st.session_state.message_text)
else:
st.image(paste_result_quote.image_data, caption="Received Image for Quote")
filename = await save_pasted_image(paste_result_quote.image_data, st.session_state.username)
if filename:
st.session_state.pasted_image_data = filename
if st.button("Send Quote πŸš€", key="send_quote"):
markdown_response = f"### Quote Response\n- **Original**: {st.session_state.quote_line}\n- **{st.session_state.username} Replies**: {quote_response}"
if st.session_state.pasted_image_data:
markdown_response += f"\n- **Image**: ![Pasted Image]({st.session_state.pasted_image_data})"
await save_chat_entry(st.session_state.username, f"Pasted image: {st.session_state.pasted_image_data}")
st.session_state.pasted_image_data = None
await save_chat_entry(st.session_state.username, markdown_response, is_markdown=True)
st.session_state.quote_line = None
st.session_state.message_text = ''
st.rerun()
current_selection = st.session_state.username if st.session_state.username in FUN_USERNAMES else ""
new_username = st.selectbox("Change Name and Voice", [""] + list(FUN_USERNAMES.keys()), index=(list(FUN_USERNAMES.keys()).index(current_selection) + 1 if current_selection else 0), format_func=lambda x: f"{x} ({FUN_USERNAMES.get(x, 'No Voice')})" if x else "Select a name")
if new_username and new_username != st.session_state.username:
await save_chat_entry("System 🌟", f"{st.session_state.username} changed name to {new_username}")
st.session_state.username = new_username
st.session_state.voice = FUN_USERNAMES[new_username]
st.markdown(f"**πŸŽ™οΈ Voice Changed**: {st.session_state.voice} πŸ—£οΈ for {st.session_state.username}")
st.rerun()
message = st.text_input(f"Message as {st.session_state.username} (Voice: {st.session_state.voice})", key="message_input", value=st.session_state.message_text)
paste_result_msg = paste_image_button("πŸ“‹ Paste Image or Text with Message", key="paste_button_msg")
if paste_result_msg.image_data is not None:
if isinstance(paste_result_msg.image_data, str):
st.session_state.message_text = paste_result_msg.image_data
st.text_input(f"Message as {st.session_state.username} (Voice: {st.session_state.voice})", key="message_input_paste", value=st.session_state.message_text)
else:
st.image(paste_result_msg.image_data, caption="Received Image for Message")
filename = await save_pasted_image(paste_result_msg.image_data, st.session_state.username)
if filename:
st.session_state.pasted_image_data = filename
if st.button("Send πŸš€", key="send_button") and (message.strip() or st.session_state.pasted_image_data):
if message.strip():
audio_file = await save_chat_entry(st.session_state.username, message, is_markdown=True)
if audio_file:
st.session_state.audio_cache[f"{message}_{FUN_USERNAMES[st.session_state.username]}"] = audio_file
if st.session_state.pasted_image_data:
await save_chat_entry(st.session_state.username, f"Pasted image: {st.session_state.pasted_image_data}")
st.session_state.pasted_image_data = None
st.session_state.message_text = ''
st.rerun()