app.py

import streamlit as st
import anthropic, openai, base64, cv2, glob, json, math, os, pytz, random, re, requests, time, zipfile
from datetime import datetime
from audio_recorder_streamlit import audio_recorder
from collections import defaultdict, Counter
from dotenv import load_dotenv
from gradio_client import Client
from huggingface_hub import InferenceClient
from PIL import Image
from openai import OpenAI
import asyncio
import edge_tts
from streamlit_marquee import streamlit_marquee

st.set_page_config(
    page_title="🚲TalkingAIResearcher🏆",
    page_icon="🚲🏆",
    layout="wide"
)

EDGE_TTS_VOICES = [
    "en-US-AriaNeural",
    "en-US-GuyNeural", 
    "en-US-JennyNeural",
    "en-GB-SoniaNeural"
]

FILE_EMOJIS = {
    "md": "📝",
    "mp3": "🎵",
    "wav": "🔊",
    "txt": "📄",
    "pdf": "📑"
}

# Initialize session states
if 'tts_voice' not in st.session_state:
    st.session_state['tts_voice'] = EDGE_TTS_VOICES[0]
if 'audio_format' not in st.session_state:
    st.session_state['audio_format'] = 'mp3'
if 'messages' not in st.session_state:
    st.session_state['messages'] = []
if 'chat_history' not in st.session_state:
    st.session_state['chat_history'] = []
if 'viewing_prefix' not in st.session_state:
    st.session_state['viewing_prefix'] = None
if 'should_rerun' not in st.session_state:
    st.session_state['should_rerun'] = False

# API Setup
openai_client = OpenAI(api_key=os.getenv('OPENAI_API_KEY'))
claude_client = anthropic.Anthropic(api_key=os.getenv('ANTHROPIC_API_KEY'))

@st.cache_resource
def get_cached_audio_b64(file_path):
    with open(file_path, "rb") as f:
        return base64.b64encode(f.read()).decode()

def beautify_filename(filename):
    name = os.path.splitext(filename)[0]
    return name.replace('_', ' ').replace('.', ' ')

def create_zip_of_files(md_files, mp3_files, wav_files, query=''):
    all_files = md_files + mp3_files + wav_files
    if not all_files: return None
    
    timestamp = datetime.now().strftime("%y%m_%H%M")
    zip_name = f"{timestamp}_archive.zip"
    with zipfile.ZipFile(zip_name, 'w') as z:
        for f in all_files:
            z.write(f)
    return zip_name

def get_download_link(file_path, file_type="zip"):
    with open(file_path, "rb") as f:
        b64 = base64.b64encode(f.read()).decode()
    ext_map = {'zip': '📦', 'mp3': '🎵', 'wav': '🔊', 'md': '📝'}
    emoji = ext_map.get(file_type, '')
    return f'<a href="data:application/{file_type};base64,{b64}" download="{os.path.basename(file_path)}">{emoji} Download {os.path.basename(file_path)}</a>'

def load_files_for_sidebar():
    files = [f for f in glob.glob("*.*") if not f.lower().endswith('readme.md')]
    groups = defaultdict(list)
    for f in files:
        basename = os.path.basename(f)
        group_name = basename[:9] if len(basename) >= 9 else 'Other'
        groups[group_name].append(f)
    return sorted(groups.items(), 
                 key=lambda x: max(os.path.getmtime(f) for f in x[1]), 
                 reverse=True)

def display_marquee_controls():
    st.sidebar.markdown("### 🎯 Marquee Settings")
    cols = st.sidebar.columns(2)
    with cols[0]:
        bg_color = st.color_picker("🎨 Background", "#1E1E1E")
        text_color = st.color_picker("✍️ Text", "#FFFFFF")
    with cols[1]:
        font_size = st.slider("📏 Size", 10, 24, 14)
        duration = st.slider("⏱️ Speed", 1, 20, 10)
    
    return {
        "background": bg_color,
        "color": text_color,
        "font-size": f"{font_size}px",
        "animationDuration": f"{duration}s",
        "width": "100%",
        "lineHeight": "35px"
    }

def display_file_manager_sidebar(groups_sorted):
    st.sidebar.title("📚 File Manager")
    all_files = {'md': [], 'mp3': [], 'wav': []}
    
    for _, files in groups_sorted:
        for f in files:
            ext = os.path.splitext(f)[1].lower().strip('.')
            if ext in all_files:
                all_files[ext].append(f)

    cols = st.sidebar.columns(4)
    for i, (ext, files) in enumerate(all_files.items()):
        with cols[i]:
            if st.button(f"🗑️ {ext.upper()}"):
                [os.remove(f) for f in files]
                st.session_state.should_rerun = True
                
    if st.sidebar.button("📦 Zip All"):
        zip_name = create_zip_of_files(
            all_files['md'], all_files['mp3'], all_files['wav']
        )
        if zip_name:
            st.sidebar.markdown(get_download_link(zip_name), unsafe_allow_html=True)

    for group_name, files in groups_sorted:
        timestamp = (datetime.strptime(group_name, "%y%m_%H%M").strftime("%Y-%m-%d %H:%M") 
                    if len(group_name) == 9 else group_name)
        
        with st.sidebar.expander(f"📁 {timestamp} ({len(files)})", expanded=True):
            c1, c2 = st.columns(2)
            with c1:
                if st.button("👀", key=f"view_{group_name}"):
                    st.session_state.viewing_prefix = group_name
            with c2:
                if st.button("🗑️", key=f"del_{group_name}"):
                    [os.remove(f) for f in files]
                    st.session_state.should_rerun = True

            for f in files:
                ext = os.path.splitext(f)[1].lower().strip('.')
                emoji = FILE_EMOJIS.get(ext, '📄')
                pretty_name = beautify_filename(os.path.basename(f))
                st.write(f"{emoji} **{pretty_name}**")
                
                if ext in ['mp3', 'wav']:
                    st.audio(f)
                    if st.button("🔄", key=f"loop_{f}"):
                        audio_b64 = get_cached_audio_b64(f)
                        st.components.v1.html(
                            f'''
                            <audio id="player_{f}" loop>
                                <source src="data:audio/{ext};base64,{audio_b64}">
                            </audio>
                            <script>
                                document.getElementById("player_{f}").play();
                            </script>
                            ''',
                            height=0
                        )

async def edge_tts_generate(text, voice, file_format="mp3"):
    text = re.sub(r'\s+', ' ', text).strip()
    if not text: return None
    communicate = edge_tts.Communicate(text, voice)
    filename = f"{datetime.now().strftime('%y%m_%H%M')}_{voice}.{file_format}"
    await communicate.save(filename)
    return filename

def parse_arxiv_refs(text):
    papers = []
    current_paper = None
    
    for line in text.split('\n'):
        if '|' in line:
            if current_paper:
                papers.append(current_paper)
            parts = line.strip('* ').split('|')
            current_paper = {
                'date': parts[0].strip(),
                'title': parts[1].strip(),
                'authors': '',
                'summary': '',
                'id': re.search(r'(\d{4}\.\d{5})', line).group(1) if re.search(r'(\d{4}\.\d{5})', line) else ''
            }
        elif current_paper:
            if not current_paper['authors']:
                current_paper['authors'] = line.strip('* ')
            else:
                current_paper['summary'] += ' ' + line.strip()
    
    if current_paper:
        papers.append(current_paper)
    return papers

def perform_ai_lookup(query):
    client = Client("awacke1/Arxiv-Paper-Search-And-QA-RAG-Pattern")
    response = client.predict(
        query, 20, "Semantic Search", 
        "mistralai/Mixtral-8x7B-Instruct-v0.1",
        api_name="/update_with_rag_md"
    )
    
    papers = parse_arxiv_refs(response[0])
    marquee_settings = display_marquee_controls()
    
    for paper in papers:
        content = f"📄 {paper['title']} | 👤 {paper['authors']} | 📝 {paper['summary']}"
        streamlit_marquee(
            content=content,
            **marquee_settings,
            key=f"paper_{paper['id'] or random.randint(1000,9999)}"
        )
        st.write("")  # Spacing
    
    return papers

def main():
    marquee_settings = display_marquee_controls()
    
    streamlit_marquee(
        content="🚀 Welcome to TalkingAIResearcher | 🤖 Your Research Assistant",
        **marquee_settings,
        key="welcome"
    )
    
    tab = st.radio("Action:", ["🎤 Voice", "🔍 ArXiv", "📝 Editor"], horizontal=True)
    
    if tab == "🔍 ArXiv":
        query = st.text_input("🔍 Search:")
        if query:
            papers = perform_ai_lookup(query)
            st.write(f"Found {len(papers)} papers")
    
    groups = load_files_for_sidebar()
    display_file_manager_sidebar(groups)
    
    if st.session_state.should_rerun:
        st.session_state.should_rerun = False
        st.rerun()

# Condensed sidebar markdown
sidebar_md = """# 📚 Research Papers

## 🧠 AGI Levels
L0 ❌ No AI
L1 🌱 ChatGPT/Bard [2303.08774v1](https://arxiv.org/abs/2303.08774) [PDF](https://arxiv.org/pdf/2303.08774.pdf)
L2 💪 Watson [2201.11903v1](https://arxiv.org/abs/2201.11903) [PDF](https://arxiv.org/pdf/2201.11903.pdf)
L3 🎯 DALL·E [2204.06125v1](https://arxiv.org/abs/2204.06125) [PDF](https://arxiv.org/pdf/2204.06125.pdf)
L4 🏆 AlphaGo [1712.01815v1](https://arxiv.org/abs/1712.01815) [PDF](https://arxiv.org/pdf/1712.01815.pdf) 
L5 🚀 AlphaFold [2203.15556v1](https://arxiv.org/abs/2203.15556) [PDF](https://arxiv.org/pdf/2203.15556.pdf)

## 🧬 AlphaFold2 
[2203.15556v1](https://arxiv.org/abs/2203.15556) [PDF](https://arxiv.org/pdf/2203.15556.pdf)
1. 🧬 Input Seq → 2. 🔍 DB Search → 3. 🧩 MSA
4. 📑 Templates → 5. 🔄 Evoformer → 6. 🧱 Structure
7. 🎯 3D Predict → 8. ♻️ Recycle x3"""

st.sidebar.markdown(sidebar_md)

if __name__ == "__main__":
    main()