import streamlit as st import asyncio import websockets import uuid 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 anthropic import openai from PyPDF2 import PdfReader import threading import json import zipfile from gradio_client import Client from dotenv import load_dotenv from streamlit_marquee import streamlit_marquee from collections import defaultdict, Counter import pandas as pd # ๐Ÿ› ๏ธ Patch asyncio for nesting nest_asyncio.apply() # ๐ŸŽจ Page Config st.set_page_config( page_title="๐ŸšฒTalkingAIResearcher๐Ÿ†", page_icon="๐Ÿšฒ๐Ÿ†", layout="wide", initial_sidebar_state="auto" ) # ๐ŸŒŸ Static Config icons = '๐Ÿค–๐Ÿง ๐Ÿ”ฌ๐Ÿ“' Site_Name = '๐Ÿค–๐Ÿง Chat & Quote Node๐Ÿ“๐Ÿ”ฌ' START_ROOM = "Sector ๐ŸŒŒ" FUN_USERNAMES = { "CosmicJester ๐ŸŒŒ": "en-US-AriaNeural", "PixelPanda ๐Ÿผ": "en-US-JennyNeural", "QuantumQuack ๐Ÿฆ†": "en-GB-SoniaNeural", "StellarSquirrel ๐Ÿฟ๏ธ": "en-AU-NatashaNeural", "GizmoGuru โš™๏ธ": "en-CA-ClaraNeural", "NebulaNinja ๐ŸŒ ": "en-US-GuyNeural", "ByteBuster ๐Ÿ’พ": "en-GB-RyanNeural", "GalacticGopher ๐ŸŒ": "en-AU-WilliamNeural", "RocketRaccoon ๐Ÿš€": "en-CA-LiamNeural", "EchoElf ๐Ÿง": "en-US-AnaNeural", "PhantomFox ๐ŸฆŠ": "en-US-BrandonNeural", "WittyWizard ๐Ÿง™": "en-GB-ThomasNeural", "LunarLlama ๐ŸŒ™": "en-AU-FreyaNeural", "SolarSloth โ˜€๏ธ": "en-CA-LindaNeural", "AstroAlpaca ๐Ÿฆ™": "en-US-ChristopherNeural", "CyberCoyote ๐Ÿบ": "en-GB-ElliotNeural", "MysticMoose ๐ŸฆŒ": "en-AU-JamesNeural", "GlitchGnome ๐Ÿงš": "en-CA-EthanNeural", "VortexViper ๐Ÿ": "en-US-AmberNeural", "ChronoChimp ๐Ÿ’": "en-GB-LibbyNeural" } EDGE_TTS_VOICES = list(set(FUN_USERNAMES.values())) FILE_EMOJIS = {"md": "๐Ÿ“", "mp3": "๐ŸŽต", "png": "๐Ÿ–ผ๏ธ", "mp4": "๐ŸŽฅ"} # ๐Ÿ“ Directories (Media at Root) for d in ["chat_logs", "vote_logs", "audio_logs", "history_logs", "audio_cache"]: os.makedirs(d, exist_ok=True) CHAT_DIR = "chat_logs" VOTE_DIR = "vote_logs" MEDIA_DIR = "." AUDIO_CACHE_DIR = "audio_cache" AUDIO_DIR = "audio_logs" STATE_FILE = "user_state.txt" CHAT_FILE = os.path.join(CHAT_DIR, "global_chat.md") QUOTE_VOTES_FILE = os.path.join(VOTE_DIR, "quote_votes.md") IMAGE_VOTES_FILE = os.path.join(VOTE_DIR, "image_votes.md") HISTORY_FILE = os.path.join(VOTE_DIR, "vote_history.md") # ๐Ÿ”‘ API Keys load_dotenv() anthropic_key = os.getenv('ANTHROPIC_API_KEY', st.secrets.get('ANTHROPIC_API_KEY', "")) openai_api_key = os.getenv('OPENAI_API_KEY', st.secrets.get('OPENAI_API_KEY', "")) openai_client = openai.OpenAI(api_key=openai_api_key) # ๐Ÿ•’ Timestamp Helper def format_timestamp_prefix(username=""): central = pytz.timezone('US/Central') now = datetime.now(central) return f"{now.strftime('%Y%m%d_%H%M%S')}-by-{username}" # ๐Ÿ“ˆ Performance Timer class PerformanceTimer: def __init__(self, name): self.name, self.start = name, None def __enter__(self): self.start = time.time() return self def __exit__(self, *args): duration = time.time() - self.start st.session_state['operation_timings'][self.name] = duration st.session_state['performance_metrics'][self.name].append(duration) # ๐ŸŽ›๏ธ Session State Init def init_session_state(): defaults = { 'server_running': False, 'server_task': None, 'active_connections': {}, 'media_notifications': [], 'last_chat_update': 0, 'displayed_chat_lines': [], 'message_text': "", 'audio_cache': {}, 'pasted_image_data': None, 'quote_line': None, 'refresh_rate': 5, 'base64_cache': {}, 'transcript_history': [], 'last_transcript': "", 'image_hashes': set(), 'tts_voice': "en-US-AriaNeural", 'chat_history': [], 'marquee_settings': { "background": "#1E1E1E", "color": "#FFFFFF", "font-size": "14px", "animationDuration": "20s", "width": "100%", "lineHeight": "35px" }, 'operation_timings': {}, 'performance_metrics': defaultdict(list), 'enable_audio': True, 'download_link_cache': {}, 'username': None, 'autosend': True, 'autosearch': True, 'last_message': "", 'last_query': "", 'mp3_files': {}, 'timer_start': time.time(), 'quote_index': 0, 'quote_source': "famous", 'last_sent_transcript': "" } for k, v in defaults.items(): if k not in st.session_state: st.session_state[k] = v # ๐Ÿ–Œ๏ธ Marquee Helpers def update_marquee_settings_ui(): st.sidebar.markdown("### ๐ŸŽฏ Marquee Settings") cols = st.sidebar.columns(2) with cols[0]: st.session_state['marquee_settings']['background'] = st.color_picker("๐ŸŽจ Background", "#1E1E1E") st.session_state['marquee_settings']['color'] = st.color_picker("โœ๏ธ Text", "#FFFFFF") with cols[1]: st.session_state['marquee_settings']['font-size'] = f"{st.slider('๐Ÿ“ Size', 10, 24, 14)}px" st.session_state['marquee_settings']['animationDuration'] = f"{st.slider('โฑ๏ธ Speed', 1, 20, 20)}s" def display_marquee(text, settings, key_suffix=""): truncated = text[:280] + "..." if len(text) > 280 else text streamlit_marquee(content=truncated, **settings, key=f"marquee_{key_suffix}") st.write("") # ๐Ÿ“ Text & File Helpers def clean_text_for_tts(text): return re.sub(r'[#*!\[\]]+', '', ' '.join(text.split()))[:200] or "No text" def clean_text_for_filename(text): return '_'.join(re.sub(r'[^\w\s-]', '', text.lower()).split())[:200] def get_high_info_terms(text, top_n=10): stop_words = {'the', 'a', 'an', 'and', 'or', 'but', 'in', 'on', 'at', 'to', 'for', 'of', 'with'} words = re.findall(r'\b\w+(?:-\w+)*\b', text.lower()) bi_grams = [' '.join(pair) for pair in zip(words, words[1:])] filtered = [t for t in words + bi_grams if t not in stop_words and len(t.split()) <= 2] return [t for t, _ in Counter(filtered).most_common(top_n)] def generate_filename(prompt, username, file_type="md"): timestamp = format_timestamp_prefix(username) hash_val = hashlib.md5(prompt.encode()).hexdigest()[:8] return f"{timestamp}-{hash_val}.{file_type}" def create_file(prompt, username, file_type="md"): filename = generate_filename(prompt, username, file_type) with open(filename, 'w', encoding='utf-8') as f: f.write(prompt) return filename def get_download_link(file, file_type="mp3"): cache_key = f"dl_{file}" if cache_key not in st.session_state['download_link_cache']: with open(file, "rb") as f: b64 = base64.b64encode(f.read()).decode() mime_types = {"mp3": "audio/mpeg", "png": "image/png", "mp4": "video/mp4", "md": "text/markdown"} st.session_state['download_link_cache'][cache_key] = f'{FILE_EMOJIS.get(file_type, "Download")} Download {os.path.basename(file)}' return st.session_state['download_link_cache'][cache_key] def save_username(username): try: with open(STATE_FILE, 'w') as f: f.write(username) except Exception as e: print(f"Failed to save username: {e}") def load_username(): if os.path.exists(STATE_FILE): try: with open(STATE_FILE, 'r') as f: return f.read().strip() except Exception as e: print(f"Failed to load username: {e}") return None def concatenate_markdown_files(): md_files = sorted(glob.glob("*.md"), key=os.path.getmtime, reverse=True) all_md_content = "" for md_file in md_files: with open(md_file, 'r', encoding='utf-8') as f: all_md_content += f.read() + "\n\n---\n\n" return all_md_content.strip() # ๐ŸŽถ Audio Processing async def async_edge_tts_generate(text, voice, username, rate=0, pitch=0, file_format="mp3"): cache_key = f"{text[:100]}_{voice}_{rate}_{pitch}_{file_format}" if cache_key in st.session_state['audio_cache']: return st.session_state['audio_cache'][cache_key], 0 start_time = time.time() text = clean_text_for_tts(text) if not text or text == "No text": print(f"Skipping audio generation for empty/invalid text: '{text}'") return None, 0 filename = f"{format_timestamp_prefix(username)}-{hashlib.md5(text.encode()).hexdigest()[:8]}.{file_format}" try: communicate = edge_tts.Communicate(text, voice, rate=f"{rate:+d}%", pitch=f"{pitch:+d}Hz") await communicate.save(filename) st.session_state['audio_cache'][cache_key] = filename return filename, time.time() - start_time except edge_tts.exceptions.NoAudioReceived as e: print(f"No audio received for text: '{text}' with voice: {voice}. Error: {e}") return None, 0 except Exception as e: print(f"Error generating audio for text: '{text}' with voice: {voice}. Error: {e}") return None, 0 def play_and_download_audio(file_path): if file_path and os.path.exists(file_path): st.audio(file_path) st.markdown(get_download_link(file_path), unsafe_allow_html=True) def load_mp3_viewer(): mp3_files = sorted(glob.glob(f"*.mp3"), key=os.path.getmtime, reverse=True) for mp3 in mp3_files: filename = os.path.basename(mp3) if filename not in st.session_state['mp3_files']: st.session_state['mp3_files'][filename] = mp3 async def save_chat_entry(username, message, is_markdown=False): central = pytz.timezone('US/Central') timestamp = datetime.now(central).strftime("%Y-%m-%d %H:%M:%S") entry = f"[{timestamp}] {username}: {message}" if not is_markdown else f"[{timestamp}] {username}:\n```markdown\n{message}\n```" with open(CHAT_FILE, 'a') as f: f.write(f"{entry}\n") voice = FUN_USERNAMES.get(username, "en-US-AriaNeural") audio_file, _ = await async_edge_tts_generate(message, voice, username) if audio_file: with open(HISTORY_FILE, 'a') as f: f.write(f"[{timestamp}] {username}: Audio - {audio_file}\n") st.session_state['mp3_files'][os.path.basename(audio_file)] = audio_file else: print(f"No audio generated for message: '{message}' by {username}") await broadcast_message(f"{username}|{message}", "chat") st.session_state.last_chat_update = time.time() st.session_state.chat_history.append(entry) return audio_file async def load_chat(): if not os.path.exists(CHAT_FILE): with open(CHAT_FILE, 'a') as f: f.write(f"# {START_ROOM} Chat\n\nWelcome to the cosmic hub! ๐ŸŽค\n") with open(CHAT_FILE, 'r') as f: content = f.read().strip() lines = content.split('\n') numbered_content = "\n".join(f"{i+1}. {line}" for i, line in enumerate(lines) if line.strip()) return numbered_content # ๐ŸŒ WebSocket Handling async def websocket_handler(websocket, path): client_id = str(uuid.uuid4()) room_id = "chat" if room_id not in st.session_state.active_connections: st.session_state.active_connections[room_id] = {} st.session_state.active_connections[room_id][client_id] = websocket username = st.session_state.get('username', random.choice(list(FUN_USERNAMES.keys()))) chat_content = await load_chat() if not any(f"Client-{client_id}" in line for line in chat_content.split('\n')): await save_chat_entry("System ๐ŸŒŸ", f"{username} has joined {START_ROOM}!") try: async for message in websocket: if '|' in message: username, content = message.split('|', 1) await save_chat_entry(username, content) else: await websocket.send("ERROR|Message format: username|content") except websockets.ConnectionClosed: await save_chat_entry("System ๐ŸŒŸ", f"{username} has left {START_ROOM}!") 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): 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: if client_id in st.session_state.active_connections[room_id]: del st.session_state.active_connections[room_id][client_id] async def run_websocket_server(): 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() # ๐Ÿ“š PDF to Audio class AudioProcessor: def __init__(self): self.cache_dir = AUDIO_CACHE_DIR os.makedirs(self.cache_dir, exist_ok=True) self.metadata = json.load(open(f"{self.cache_dir}/metadata.json")) if os.path.exists(f"{self.cache_dir}/metadata.json") else {} def _save_metadata(self): with open(f"{self.cache_dir}/metadata.json", 'w') as f: json.dump(self.metadata, f) async def create_audio(self, text, voice='en-US-AriaNeural'): cache_key = hashlib.md5(f"{text}:{voice}".encode()).hexdigest() cache_path = f"{self.cache_dir}/{cache_key}.mp3" if cache_key in self.metadata and os.path.exists(cache_path): return cache_path text = clean_text_for_tts(text) if not text: return None communicate = edge_tts.Communicate(text, voice) await communicate.save(cache_path) self.metadata[cache_key] = {'timestamp': datetime.now().isoformat(), 'text_length': len(text), 'voice': voice} self._save_metadata() return cache_path def process_pdf(pdf_file, max_pages, voice, audio_processor): reader = PdfReader(pdf_file) total_pages = min(len(reader.pages), max_pages) texts, audios = [], {} async def process_page(i, text): audio_path = await audio_processor.create_audio(text, voice) if audio_path: audios[i] = audio_path for i in range(total_pages): text = reader.pages[i].extract_text() texts.append(text) threading.Thread(target=lambda: asyncio.run(process_page(i, text))).start() return texts, audios, total_pages # ๐Ÿ” ArXiv & AI Lookup def parse_arxiv_refs(ref_text): if not ref_text: return [] papers = [] current = {} for line in ref_text.split('\n'): if line.count('|') == 2: if current: papers.append(current) date, title, *_ = line.strip('* ').split('|') url = re.search(r'(https://arxiv.org/\S+)', line).group(1) if re.search(r'(https://arxiv.org/\S+)', line) else f"paper_{len(papers)}" current = {'date': date, 'title': title, 'url': url, 'authors': '', 'summary': '', 'full_audio': None, 'download_base64': ''} elif current: if not current['authors']: current['authors'] = line.strip('* ') else: current['summary'] += ' ' + line.strip() if current['summary'] else line.strip() if current: papers.append(current) return papers[:20] def generate_5min_feature_markdown(paper): title, summary, authors, date, url = paper['title'], paper['summary'], paper['authors'], paper['date'], paper['url'] pdf_url = url.replace("abs", "pdf") + (".pdf" if not url.endswith(".pdf") else "") wct, sw = len(title.split()), len(summary.split()) terms = get_high_info_terms(summary, 15) rouge = round((len(terms) / max(sw, 1)) * 100, 2) mermaid = "```mermaid\nflowchart TD\n" + "\n".join(f' T{i+1}["{t}"] --> T{i+2}["{terms[i+1]}"]' for i in range(len(terms)-1)) + "\n```" return f""" ## ๐Ÿ“„ {title} **Authors:** {authors} | **Date:** {date} | **Words:** Title: {wct}, Summary: {sw} **Links:** [Abstract]({url}) | [PDF]({pdf_url}) **Terms:** {', '.join(terms)} | **ROUGE:** {rouge}% ### ๐ŸŽค TTF Read Aloud - **Title:** {title} | **Terms:** {', '.join(terms)} | **ROUGE:** {rouge}% #### Concepts Graph {mermaid} --- """ def create_detailed_paper_md(papers): return "# Detailed Summary\n" + "\n".join(generate_5min_feature_markdown(p) for p in papers) async def create_paper_audio_files(papers, query): for p in papers: audio_text = clean_text_for_tts(f"{p['title']} by {p['authors']}. {p['summary']}") p['full_audio'], _ = await async_edge_tts_generate(audio_text, st.session_state['tts_voice'], p['authors']) if p['full_audio']: p['download_base64'] = get_download_link(p['full_audio']) async def perform_ai_lookup(q, useArxiv=True, useArxivAudio=False): client = anthropic.Anthropic(api_key=anthropic_key) response = client.messages.create(model="claude-3-sonnet-20240229", max_tokens=1000, messages=[{"role": "user", "content": q}]) result = response.content[0].text st.markdown("### Claude's Reply ๐Ÿง \n" + result) md_file = create_file(result, "System", "md") audio_file, _ = await async_edge_tts_generate(result, st.session_state['tts_voice'], "System") play_and_download_audio(audio_file) if useArxiv: q += result gradio_client = Client("awacke1/Arxiv-Paper-Search-And-QA-RAG-Pattern") refs = gradio_client.predict(q, 10, "Semantic Search", "mistralai/Mixtral-8x7B-Instruct-v0.1", api_name="/update_with_rag_md")[0] result = f"๐Ÿ”Ž {q}\n\n{refs}" md_file, audio_file = create_file(result, "System", "md"), (await async_edge_tts_generate(result, st.session_state['tts_voice'], "System"))[0] play_and_download_audio(audio_file) papers = parse_arxiv_refs(refs) if papers and useArxivAudio: await create_paper_audio_files(papers, q) return result, papers return result, [] def save_vote(file, item, user_hash): timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S") entry = f"[{timestamp}] {user_hash} voted for {item}" try: with open(file, 'a') as f: f.write(f"{entry}\n") with open(HISTORY_FILE, 'a') as f: f.write(f"- {timestamp} - User {user_hash} voted for {item}\n") return True except Exception as e: print(f"Vote save flop: {e}") return False def load_votes(file): if not os.path.exists(file): with open(file, 'w') as f: f.write("# Vote Tally\n\nNo votes yet - get clicking! ๐Ÿ–ฑ๏ธ\n") try: with open(file, 'r') as f: lines = f.read().strip().split('\n') votes = {} for line in lines[2:]: # Skip header if line.strip() and 'voted for' in line: item = line.split('voted for ')[1] votes[item] = votes.get(item, 0) + 1 return votes except Exception as e: print(f"Vote load oopsie: {e}") return {} def generate_user_hash(): if 'user_hash' not in st.session_state: session_id = str(random.getrandbits(128)) hash_object = hashlib.md5(session_id.encode()) st.session_state['user_hash'] = hash_object.hexdigest()[:8] return st.session_state['user_hash'] async def save_pasted_image(image, username): img_hash = hashlib.md5(image.tobytes()).hexdigest()[:8] if img_hash in st.session_state.image_hashes: return None timestamp = format_timestamp_prefix(username) filename = f"{timestamp}-{img_hash}.png" filepath = filename image.save(filepath, "PNG") st.session_state.image_hashes.add(img_hash) return filepath # ๐Ÿ“ฆ Zip Files def create_zip_of_files(md_files, mp3_files, png_files, mp4_files, query): all_files = md_files + mp3_files + png_files + mp4_files if not all_files: return None terms = get_high_info_terms(" ".join([open(f, 'r', encoding='utf-8').read() if f.endswith('.md') else os.path.splitext(os.path.basename(f))[0].replace('_', ' ') for f in all_files] + [query]), 5) zip_name = f"{format_timestamp_prefix()}_{'-'.join(terms)[:20]}.zip" with zipfile.ZipFile(zip_name, 'w') as z: [z.write(f) for f in all_files] return zip_name # ๐ŸŽฎ Main Interface async def async_interface(): init_session_state() load_mp3_viewer() saved_username = load_username() if saved_username and saved_username in FUN_USERNAMES: st.session_state.username = saved_username if not st.session_state.username: available = [n for n in FUN_USERNAMES if not any(f"{n} has joined" in l for l in (await load_chat()).split('\n'))] st.session_state.username = random.choice(available or list(FUN_USERNAMES.keys())) st.session_state.tts_voice = FUN_USERNAMES[st.session_state.username] await save_chat_entry("System ๐ŸŒŸ", f"{st.session_state.username} has joined {START_ROOM}!") save_username(st.session_state.username) st.title(f"{Site_Name} for {st.session_state.username}") update_marquee_settings_ui() display_marquee(f"๐Ÿš€ Welcome to {START_ROOM} | ๐Ÿค– {st.session_state.username}", st.session_state['marquee_settings'], "welcome") if not st.session_state.server_task: st.session_state.server_task = asyncio.create_task(run_websocket_server()) tab_main = st.radio("Action:", ["๐ŸŽค Chat & Voice", "๐Ÿ“ธ Media", "๐Ÿ” ArXiv", "๐Ÿ“š PDF to Audio"], horizontal=True) useArxiv, useArxivAudio = st.checkbox("Search ArXiv", True), st.checkbox("ArXiv Audio", False) st.session_state.autosend = st.checkbox("Autosend Chat", value=True) st.session_state.autosearch = st.checkbox("Autosearch ArXiv", value=True) # ๐ŸŽค Chat & Voice if tab_main == "๐ŸŽค Chat & Voice": st.subheader(f"{START_ROOM} Chat ๐Ÿ’ฌ") chat_content = await load_chat() chat_container = st.container() with chat_container: lines = chat_content.split('\n') for i, line in enumerate(lines): if line.strip(): col1, col2 = st.columns([5, 1]) with col1: st.markdown(line) for mp3_name, mp3_path in st.session_state['mp3_files'].items(): if st.session_state.username in mp3_name and any(word in mp3_name for word in line.split()): st.audio(mp3_path) break with col2: if st.button(f"๐Ÿ‘", key=f"chat_vote_{i}"): user_hash = generate_user_hash() save_vote(QUOTE_VOTES_FILE, line, user_hash) st.session_state.timer_start = time.time() save_username(st.session_state.username) st.rerun() message = st.text_input(f"Message as {st.session_state.username}", key="message_input") paste_result = paste_image_button("๐Ÿ“‹ Paste Image or Text", key="paste_button_msg") if paste_result.image_data is not None: if isinstance(paste_result.image_data, str): st.session_state.message_text = paste_result.image_data message = st.text_input(f"Message as {st.session_state.username}", key="message_input_paste", value=st.session_state.message_text) else: st.image(paste_result.image_data, caption="Pasted Image") filename = await save_pasted_image(paste_result.image_data, st.session_state.username) if filename: st.session_state.pasted_image_data = filename if (message and message != st.session_state.last_message) or st.session_state.pasted_image_data: st.session_state.last_message = message if st.session_state.autosend or st.button("Send ๐Ÿš€"): if message.strip(): await save_chat_entry(st.session_state.username, message, True) 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.timer_start = time.time() save_username(st.session_state.username) st.rerun() st.subheader("๐ŸŽค Speech-to-Chat") from mycomponent import mycomponent transcript_data = mycomponent(default_value=st.session_state.get('last_transcript', ''), key="speech_input") if transcript_data and 'value' in transcript_data: transcript = transcript_data['value'].strip() if transcript and transcript != st.session_state.get('last_transcript', ''): st.session_state.last_transcript = transcript else: transcript = st.session_state.get('last_transcript', '') edited_transcript = st.text_area("โœ๏ธ Edit Transcript:", value=transcript, height=100, key="transcript_input") if transcript: st.write(f"๐ŸŽ™๏ธ You said: {transcript}") if edited_transcript and edited_transcript != st.session_state.get('last_sent_transcript', ''): if st.session_state.autosend: await save_chat_entry(st.session_state.username, edited_transcript, True) st.session_state.last_sent_transcript = edited_transcript st.session_state.timer_start = time.time() save_username(st.session_state.username) with chat_container: st.markdown(await load_chat()) elif st.button("Send to Chat", key="send_transcript"): await save_chat_entry(st.session_state.username, edited_transcript, True) st.session_state.last_sent_transcript = edited_transcript st.session_state.timer_start = time.time() save_username(st.session_state.username) st.rerun() if not transcript: st.write("๐ŸŽ™๏ธ Speak to transcribe your message...") # ๐Ÿ“ธ Media elif tab_main == "๐Ÿ“ธ Media": st.header("๐Ÿ“ธ Media Gallery") all_files = sorted(glob.glob("*.md") + glob.glob("*.mp3") + glob.glob("*.png") + glob.glob("*.mp4"), key=os.path.getmtime, reverse=True) md_files = [f for f in all_files if f.endswith('.md')] mp3_files = [f for f in all_files if f.endswith('.mp3')] png_files = [f for f in all_files if f.endswith('.png')] mp4_files = [f for f in all_files if f.endswith('.mp4')] st.subheader("All Submitted Text") all_md_content = concatenate_markdown_files() st.markdown(all_md_content) st.subheader("๐ŸŽต Audio (MP3)") for mp3 in mp3_files: with st.expander(os.path.basename(mp3)): st.audio(mp3) st.markdown(get_download_link(mp3, "mp3"), unsafe_allow_html=True) st.subheader("๐Ÿ–ผ๏ธ Images (PNG)") for png in png_files: with st.expander(os.path.basename(png)): st.image(png, use_container_width=True) st.markdown(get_download_link(png, "png"), unsafe_allow_html=True) st.subheader("๐ŸŽฅ Videos (MP4)") for mp4 in mp4_files: with st.expander(os.path.basename(mp4)): st.video(mp4) st.markdown(get_download_link(mp4, "mp4"), unsafe_allow_html=True) uploaded_file = st.file_uploader("Upload Media", type=['png', 'mp4', 'mp3']) if uploaded_file: filename = f"{format_timestamp_prefix(st.session_state.username)}-{hashlib.md5(uploaded_file.getbuffer()).hexdigest()[:8]}.{uploaded_file.name.split('.')[-1]}" with open(filename, 'wb') as f: f.write(uploaded_file.getbuffer()) await save_chat_entry(st.session_state.username, f"Uploaded: {filename}") st.session_state.timer_start = time.time() save_username(st.session_state.username) st.rerun() # ๐Ÿ” ArXiv elif tab_main == "๐Ÿ” ArXiv": q = st.text_input("๐Ÿ” Query:", key="arxiv_query") if q and q != st.session_state.last_query: st.session_state.last_query = q if st.session_state.autosearch or st.button("๐Ÿ” Run"): result, papers = await perform_ai_lookup(q, useArxiv, useArxivAudio) for i, p in enumerate(papers, 1): with st.expander(f"{i}. ๐Ÿ“„ {p['title']}"): st.markdown(f"**{p['date']} | {p['title']}** โ€” [Link]({p['url']})") st.markdown(generate_5min_feature_markdown(p)) if p.get('full_audio'): play_and_download_audio(p['full_audio']) # ๐Ÿ“š PDF to Audio elif tab_main == "๐Ÿ“š PDF to Audio": audio_processor = AudioProcessor() pdf_file = st.file_uploader("Choose PDF", "pdf") max_pages = st.slider('Pages', 1, 100, 10) if pdf_file: with st.spinner('Processing...'): texts, audios, total = process_pdf(pdf_file, max_pages, st.session_state['tts_voice'], audio_processor) for i, text in enumerate(texts): with st.expander(f"Page {i+1}"): st.markdown(text) while i not in audios: time.sleep(0.1) if audios.get(i): st.audio(audios[i]) st.markdown(get_download_link(audios[i], "mp3"), unsafe_allow_html=True) await save_chat_entry(st.session_state.username, f"PDF Page {i+1} converted to audio: {audios[i]}") # ๐Ÿ—‚๏ธ Sidebar with Dialog and Audio st.sidebar.subheader("Voice Settings") new_username = st.sidebar.selectbox("Change Name/Voice", list(FUN_USERNAMES.keys()), index=list(FUN_USERNAMES.keys()).index(st.session_state.username)) if new_username != st.session_state.username: await save_chat_entry("System ๐ŸŒŸ", f"{st.session_state.username} changed to {new_username}") st.session_state.username, st.session_state.tts_voice = new_username, FUN_USERNAMES[new_username] st.session_state.timer_start = time.time() save_username(st.session_state.username) st.rerun() st.sidebar.markdown("### ๐Ÿ’ฌ Chat Dialog & Media") chat_content = await load_chat() lines = chat_content.split('\n') all_files = sorted(glob.glob("*.md") + glob.glob("*.mp3") + glob.glob("*.png") + glob.glob("*.mp4"), key=os.path.getmtime, reverse=True) for line in lines[-10:]: if line.strip(): st.sidebar.markdown(f"**{line}**") for f in all_files: f_name = os.path.basename(f) if st.session_state.username in f_name and any(word in f_name for word in line.split()): if f.endswith('.mp3'): st.sidebar.audio(f) st.sidebar.markdown(get_download_link(f, "mp3"), unsafe_allow_html=True) elif f.endswith('.png'): st.sidebar.image(f, use_container_width=True) st.sidebar.markdown(get_download_link(f, "png"), unsafe_allow_html=True) elif f.endswith('.mp4'): st.sidebar.video(f) st.sidebar.markdown(get_download_link(f, "mp4"), unsafe_allow_html=True) break st.sidebar.subheader("Vote Totals") chat_votes = load_votes(QUOTE_VOTES_FILE) image_votes = load_votes(IMAGE_VOTES_FILE) for item, count in chat_votes.items(): st.sidebar.write(f"{item}: {count} votes") for image, count in image_votes.items(): st.sidebar.write(f"{image}: {count} votes") md_files = [f for f in all_files if f.endswith('.md')] mp3_files = [f for f in all_files if f.endswith('.mp3')] png_files = [f for f in all_files if f.endswith('.png')] mp4_files = [f for f in all_files if f.endswith('.mp4')] st.sidebar.markdown("### ๐Ÿ“‚ File History") for f in all_files[:10]: st.sidebar.write(f"{FILE_EMOJIS.get(f.split('.')[-1], '๐Ÿ“„')} {os.path.basename(f)}") if st.sidebar.button("โฌ‡๏ธ Zip All"): zip_name = create_zip_of_files(md_files, mp3_files, png_files, mp4_files, "latest_query") if zip_name: st.sidebar.markdown(get_download_link(zip_name, "zip"), unsafe_allow_html=True) def main(): asyncio.run(async_interface()) if __name__ == "__main__": main()