Create backup21.app.py

backup21.app.py

@@ -0,0 +1,947 @@
+import streamlit as st
+import anthropic, openai, base64, cv2, glob, json, math, os, pytz, random, re, requests, textract, time, zipfile
+import plotly.graph_objects as go
+import streamlit.components.v1 as components
+from datetime import datetime
+from audio_recorder_streamlit import audio_recorder
+from bs4 import BeautifulSoup
+from collections import defaultdict, deque, Counter
+from dotenv import load_dotenv
+from gradio_client import Client
+from huggingface_hub import InferenceClient
+from io import BytesIO
+from PIL import Image
+from PyPDF2 import PdfReader
+from urllib.parse import quote
+from xml.etree import ElementTree as ET
+from openai import OpenAI
+import extra_streamlit_components as stx
+from streamlit.runtime.scriptrunner import get_script_run_ctx
+import asyncio
+import edge_tts
+from streamlit_marquee import streamlit_marquee
+
+# ─────────────────────────────────────────────────────────
+# 1. CORE CONFIGURATION & SETUP
+# ─────────────────────────────────────────────────────────
+st.set_page_config(
+    page_title="🚲TalkingAIResearcher🏆",
+    page_icon="🚲🏆",
+    layout="wide",
+    initial_sidebar_state="auto",
+    menu_items={
+        'Get Help': 'https://huggingface.co/awacke1',
+        'Report a bug': 'https://huggingface.co/spaces/awacke1',
+        'About': "🚲TalkingAIResearcher🏆"
+    }
+)
+load_dotenv()
+
+# Available English voices for Edge TTS
+EDGE_TTS_VOICES = [
+    "en-US-AriaNeural",
+    "en-US-GuyNeural",
+    "en-US-JennyNeural",
+    "en-GB-SoniaNeural",
+    "en-GB-RyanNeural",
+    "en-AU-NatashaNeural",
+    "en-AU-WilliamNeural",
+    "en-CA-ClaraNeural",
+    "en-CA-LiamNeural"
+]
+
+# Session state variables
+if 'marquee_settings' not in st.session_state:
+    st.session_state['marquee_settings'] = {
+        "background": "#1E1E1E",
+        "color": "#FFFFFF",
+        "font-size": "14px",
+        "animationDuration": "20s",
+        "width": "100%",
+        "lineHeight": "35px"
+    }
+
+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 'transcript_history' not in st.session_state:
+    st.session_state['transcript_history'] = []
+
+if 'chat_history' not in st.session_state:
+    st.session_state['chat_history'] = []
+
+if 'openai_model' not in st.session_state:
+    st.session_state['openai_model'] = "gpt-4o-2024-05-13"
+
+if 'messages' not in st.session_state:
+    st.session_state['messages'] = []
+
+if 'last_voice_input' not in st.session_state:
+    st.session_state['last_voice_input'] = ""
+
+if 'editing_file' not in st.session_state:
+    st.session_state['editing_file'] = None
+
+if 'edit_new_name' not in st.session_state:
+    st.session_state['edit_new_name'] = ""
+
+if 'edit_new_content' not in st.session_state:
+    st.session_state['edit_new_content'] = ""
+
+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
+
+if 'old_val' not in st.session_state:
+    st.session_state['old_val'] = None
+
+if 'last_query' not in st.session_state:
+    st.session_state['last_query'] = ""
+
+if 'marquee_content' not in st.session_state:
+    st.session_state['marquee_content'] = "🚀 Welcome to TalkingAIResearcher | 🤖 Your Research Assistant"
+
+# API Keys
+openai_api_key = os.getenv('OPENAI_API_KEY', "")
+anthropic_key = os.getenv('ANTHROPIC_API_KEY_3', "")
+xai_key = os.getenv('xai',"")
+if 'OPENAI_API_KEY' in st.secrets:
+    openai_api_key = st.secrets['OPENAI_API_KEY']
+if 'ANTHROPIC_API_KEY' in st.secrets:
+    anthropic_key = st.secrets["ANTHROPIC_API_KEY"]
+
+openai.api_key = openai_api_key
+openai_client = OpenAI(api_key=openai.api_key, organization=os.getenv('OPENAI_ORG_ID'))
+HF_KEY = os.getenv('HF_KEY')
+API_URL = os.getenv('API_URL')
+
+# Helper constants
+FILE_EMOJIS = {
+    "md": "📝",
+    "mp3": "🎵",
+    "wav": "🔊"
+}
+
+# ─────────────────────────────────────────────────────────
+# 2. HELPER FUNCTIONS
+# ─────────────────────────────────────────────────────────
+
+def get_central_time():
+    """Get current time in US Central timezone."""
+    central = pytz.timezone('US/Central')
+    return datetime.now(central)
+
+def format_timestamp_prefix():
+    """Generate timestamp prefix in format MM_dd_yy_hh_mm_AM/PM."""
+    ct = get_central_time()
+    return ct.strftime("%m_%d_%y_%I_%M_%p")
+
+def initialize_marquee_settings():
+    if 'marquee_settings' not in st.session_state:
+        st.session_state['marquee_settings'] = {
+            "background": "#1E1E1E",
+            "color": "#FFFFFF",
+            "font-size": "14px",
+            "animationDuration": "20s",
+            "width": "100%",
+            "lineHeight": "35px"
+        }
+
+def get_marquee_settings():
+    initialize_marquee_settings()
+    return st.session_state['marquee_settings']
+
+def update_marquee_settings_ui():
+    """Add color pickers & sliders for marquee config in sidebar."""
+    st.sidebar.markdown("### 🎯 Marquee Settings")
+    cols = st.sidebar.columns(2)
+    with cols[0]:
+        bg_color = st.color_picker("🎨 Background", 
+                                  st.session_state['marquee_settings']["background"], 
+                                  key="bg_color_picker")
+        text_color = st.color_picker("✍️ Text", 
+                                    st.session_state['marquee_settings']["color"], 
+                                    key="text_color_picker")
+    with cols[1]:
+        font_size = st.slider("📏 Size", 10, 24, 14, key="font_size_slider")
+        duration = st.slider("⏱️ Speed", 1, 20, 20, key="duration_slider")
+
+    st.session_state['marquee_settings'].update({
+        "background": bg_color,
+        "color": text_color,
+        "font-size": f"{font_size}px",
+        "animationDuration": f"{duration}s"
+    })
+
+def display_marquee(text, settings, key_suffix=""):
+    """Show marquee text with style from settings."""
+    truncated_text = text[:280] + "..." if len(text) > 280 else text
+    streamlit_marquee(
+        content=truncated_text,
+        **settings,
+        key=f"marquee_{key_suffix}"
+    )
+    st.write("")
+
+def get_high_info_terms(text: str, top_n=10) -> list:
+    """Extract top_n freq words or bigrams (excluding stopwords)."""
+    stop_words = set(['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:])]
+    combined = words + bi_grams
+    filtered = [term for term in combined if term not in stop_words and len(term.split()) <= 2]
+    counter = Counter(filtered)
+    return [term for term, freq in counter.most_common(top_n)]
+
+def clean_text_for_filename(text: str) -> str:
+    """Remove special chars, short words, etc. for filenames."""
+    text = text.lower()
+    text = re.sub(r'[^\w\s-]', '', text)
+    words = text.split()
+    # remove short or unhelpful words
+    stop_short = set(['the', 'and', 'for', 'with', 'this', 'that', 'ai', 'library'])
+    filtered = [w for w in words if len(w) > 3 and w not in stop_short]
+    return '_'.join(filtered)[:200]
+
+def generate_filename(prompt, response, file_type="md", max_length=200):
+    """
+    Generate a shortened filename by:
+      1) extracting high-info terms,
+      2) snippet from prompt+response,
+      3) remove duplicates,
+      4) truncate if needed.
+    """
+    prefix = format_timestamp_prefix() + "_"
+    combined_text = (prompt + " " + response)[:200]  
+    info_terms = get_high_info_terms(combined_text, top_n=5)  
+    snippet = (prompt[:40] + " " + response[:40]).strip()
+    snippet_cleaned = clean_text_for_filename(snippet)
+    
+    # remove duplicates
+    name_parts = info_terms + [snippet_cleaned]
+    seen = set()
+    unique_parts = []
+    for part in name_parts:
+        if part not in seen:
+            seen.add(part)
+            unique_parts.append(part)
+    
+    full_name = '_'.join(unique_parts).strip('_')
+    leftover_chars = max_length - len(prefix) - len(file_type) - 1
+    if len(full_name) > leftover_chars:
+        full_name = full_name[:leftover_chars]
+    
+    return f"{prefix}{full_name}.{file_type}"
+
+def create_file(prompt, response, file_type="md"):
+    """Create a text file from prompt + response with sanitized filename."""
+    filename = generate_filename(prompt.strip(), response.strip(), file_type)
+    with open(filename, 'w', encoding='utf-8') as f:
+        f.write(prompt + "\n\n" + response)
+    return filename
+
+def get_download_link(file, file_type="zip"):
+    """
+    Convert a file to base64 and return an HTML link for download.
+    """
+    with open(file, "rb") as f:
+        b64 = base64.b64encode(f.read()).decode()
+    if file_type == "zip":
+        return f'<a href="data:application/zip;base64,{b64}" download="{os.path.basename(file)}">📂 Download {os.path.basename(file)}</a>'
+    elif file_type == "mp3":
+        return f'<a href="data:audio/mpeg;base64,{b64}" download="{os.path.basename(file)}">🎵 Download {os.path.basename(file)}</a>'
+    elif file_type == "wav":
+        return f'<a href="data:audio/wav;base64,{b64}" download="{os.path.basename(file)}">🔊 Download {os.path.basename(file)}</a>'
+    elif file_type == "md":
+        return f'<a href="data:text/markdown;base64,{b64}" download="{os.path.basename(file)}">📝 Download {os.path.basename(file)}</a>'
+    else:
+        return f'<a href="data:application/octet-stream;base64,{b64}" download="{os.path.basename(file)}">Download {os.path.basename(file)}</a>'
+
+def clean_for_speech(text: str) -> str:
+    """Clean up text for TTS output."""
+    text = text.replace("\n", " ")
+    text = text.replace("</s>", " ")
+    text = text.replace("#", "")
+    text = re.sub(r"\(https?:\/\/[^\)]+\)", "", text)
+    text = re.sub(r"\s+", " ", text).strip()
+    return text
+
+async def edge_tts_generate_audio(text, voice="en-US-AriaNeural", rate=0, pitch=0, file_format="mp3"):
+    """Async TTS generation with edge-tts library."""
+    text = clean_for_speech(text)
+    if not text.strip():
+        return None
+    rate_str = f"{rate:+d}%"
+    pitch_str = f"{pitch:+d}Hz"
+    communicate = edge_tts.Communicate(text, voice, rate=rate_str, pitch=pitch_str)
+    out_fn = generate_filename(text, text, file_type=file_format)
+    await communicate.save(out_fn)
+    return out_fn
+
+def speak_with_edge_tts(text, voice="en-US-AriaNeural", rate=0, pitch=0, file_format="mp3"):
+    """Wrapper for the async TTS generate call."""
+    return asyncio.run(edge_tts_generate_audio(text, voice, rate, pitch, file_format))
+
+def play_and_download_audio(file_path, file_type="mp3"):
+    """Streamlit audio + a quick download link."""
+    if file_path and os.path.exists(file_path):
+        st.audio(file_path)
+        dl_link = get_download_link(file_path, file_type=file_type)
+        st.markdown(dl_link, unsafe_allow_html=True)
+
+def save_qa_with_audio(question, answer, voice=None):
+    """Save Q&A to markdown and also generate audio."""
+    if not voice:
+        voice = st.session_state['tts_voice']
+    
+    combined_text = f"# Question\n{question}\n\n# Answer\n{answer}"
+    md_file = create_file(question, answer, "md")
+    audio_text = f"{question}\n\nAnswer: {answer}"
+    audio_file = speak_with_edge_tts(
+        audio_text,
+        voice=voice,
+        file_format=st.session_state['audio_format']
+    )
+    return md_file, audio_file
+
+# ─────────────────────────────────────────────────────────
+# 3. PAPER PARSING & DISPLAY
+# ─────────────────────────────────────────────────────────
+
+def parse_arxiv_refs(ref_text: str):
+    """
+    Given a multi-line markdown with arxiv references, parse them into
+    a list of dicts: {date, title, url, authors, summary, ...}.
+    """
+    if not ref_text:
+        return []
+
+    results = []
+    current_paper = {}
+    lines = ref_text.split('\n')
+    
+    for i, line in enumerate(lines):
+        if line.count('|') == 2:
+            # Found a new paper line
+            if current_paper:
+                results.append(current_paper)
+                if len(results) >= 20:
+                    break
+            try:
+                header_parts = line.strip('* ').split('|')
+                date = header_parts[0].strip()
+                title = header_parts[1].strip()
+                url_match = re.search(r'(https://arxiv.org/\S+)', line)
+                url = url_match.group(1) if url_match else f"paper_{len(results)}"
+                
+                current_paper = {
+                    'date': date,
+                    'title': title,
+                    'url': url,
+                    'authors': '',
+                    'summary': '',
+                    'full_audio': None,
+                    'download_base64': '',
+                }
+            except Exception as e:
+                st.warning(f"Error parsing paper header: {str(e)}")
+                current_paper = {}
+                continue
+        
+        elif current_paper:
+            # If authors not set, fill it; otherwise, fill summary
+            if not current_paper['authors']:
+                current_paper['authors'] = line.strip('* ')
+            else:
+                if current_paper['summary']:
+                    current_paper['summary'] += ' ' + line.strip()
+                else:
+                    current_paper['summary'] = line.strip()
+    
+    if current_paper:
+        results.append(current_paper)
+    
+    return results[:20]
+
+def create_paper_links_md(papers):
+    """Creates a minimal .md content linking to each paper's arxiv URL."""
+    lines = ["# Paper Links\n"]
+    for i, p in enumerate(papers, start=1):
+        lines.append(f"{i}. **{p['title']}** — [Arxiv]({p['url']})")
+    return "\n".join(lines)
+
+def create_paper_audio_files(papers, input_question):
+    """
+    For each paper, generate TTS audio summary, store the path in `paper['full_audio']`,
+    and also store a base64 link for stable downloading.
+    """
+    for paper in papers:
+        try:
+            audio_text = f"{paper['title']} by {paper['authors']}. {paper['summary']}"
+            audio_text = clean_for_speech(audio_text)
+            file_format = st.session_state['audio_format']
+            audio_file = speak_with_edge_tts(
+                audio_text, 
+                voice=st.session_state['tts_voice'], 
+                file_format=file_format
+            )
+            paper['full_audio'] = audio_file
+
+            if audio_file:
+                with open(audio_file, "rb") as af:
+                    b64_data = base64.b64encode(af.read()).decode()
+                download_filename = os.path.basename(audio_file)
+                mime_type = "mpeg" if file_format == "mp3" else "wav"
+                paper['download_base64'] = (
+                    f'<a href="data:audio/{mime_type};base64,{b64_data}" '
+                    f'download="{download_filename}">🎵 Download {download_filename}</a>'
+                )
+
+        except Exception as e:
+            st.warning(f"Error processing paper {paper['title']}: {str(e)}")
+            paper['full_audio'] = None
+            paper['download_base64'] = ''
+
+            
+def display_file_history_in_sidebar():
+    """
+    Shows a history of files grouped by query, with lazy loading of audio and content.
+    """
+    st.sidebar.markdown("---")
+    st.sidebar.markdown("### 📂 File History")
+
+    # Gather all files
+    md_files = glob.glob("*.md")
+    mp3_files = glob.glob("*.mp3")
+    wav_files = glob.glob("*.wav")
+    all_files = md_files + mp3_files + wav_files
+
+    if not all_files:
+        st.sidebar.write("No files found.")
+        return
+
+    # Group files by their query prefix (timestamp_query)
+    grouped_files = {}
+    for f in all_files:
+        fname = os.path.basename(f)
+        prefix = '_'.join(fname.split('_')[:6])  # Get timestamp part
+        if prefix not in grouped_files:
+            grouped_files[prefix] = {'md': [], 'audio': [], 'loaded': False}
+        
+        ext = os.path.splitext(fname)[1].lower()
+        if ext == '.md':
+            grouped_files[prefix]['md'].append(f)
+        elif ext in ['.mp3', '.wav']:
+            grouped_files[prefix]['audio'].append(f)
+
+    # Sort groups by timestamp (newest first)
+    sorted_groups = sorted(grouped_files.items(), key=lambda x: x[0], reverse=True)
+
+    # 🗑⬇️ Sidebar delete all and zip all download
+    col1, col4 = st.sidebar.columns(2)
+    with col1:
+        if st.button("🗑 Delete All"):
+            for f in all_files:
+                os.remove(f)
+            st.session_state.should_rerun = True
+    with col4:
+        if st.button("⬇️ Zip All"):
+            zip_name = create_zip_of_files(md_files, mp3_files, wav_files, 
+                                         st.session_state.get('last_query', ''))
+            if zip_name:
+                st.sidebar.markdown(get_download_link(zip_name, "zip"), 
+                                  unsafe_allow_html=True)
+
+    # Display grouped files
+    for prefix, files in sorted_groups:
+        # Get a preview of content from first MD file
+        preview = ""
+        if files['md']:
+            with open(files['md'][0], "r", encoding="utf-8") as f:
+                preview = f.read(200).replace("\n", " ")
+                if len(preview) > 200:
+                    preview += "..."
+
+        # Create unique key for this group
+        group_key = f"group_{prefix}"
+        if group_key not in st.session_state:
+            st.session_state[group_key] = False
+
+        # Display group expander
+        with st.sidebar.expander(f"📑 Query Group: {prefix}"):
+            st.write("**Preview:**")
+            st.write(preview)
+            
+            # Load full content button
+            if st.button("📖 View Full Content", key=f"btn_{prefix}"):
+                st.session_state[group_key] = True
+
+            # Only show full content and audio if button was clicked
+            if st.session_state[group_key]:
+                # Display markdown files
+                for md_file in files['md']:
+                    with open(md_file, "r", encoding="utf-8") as f:
+                        content = f.read()
+                    st.markdown("**Full Content:**")
+                    st.markdown(content)
+                    st.markdown(get_download_link(md_file, file_type="md"), 
+                              unsafe_allow_html=True)
+
+                # Display audio files
+                usePlaySidebar=False
+                if usePlaySidebar:
+                    for audio_file in files['audio']:
+                        ext = os.path.splitext(audio_file)[1].replace('.', '')
+                        st.audio(audio_file)
+                        st.markdown(get_download_link(audio_file, file_type=ext), 
+                                  unsafe_allow_html=True)
+
+def display_papers(papers, marquee_settings):
+    """Display paper info with both abs and PDF links."""
+    st.write("## Research Papers")
+    for i, paper in enumerate(papers, start=1):
+        marquee_text = f"📄 {paper['title']} | 👤 {paper['authors'][:120]}"
+        display_marquee(marquee_text, marquee_settings, key_suffix=f"paper_{i}")
+        
+        with st.expander(f"{i}. 📄 {paper['title']}", expanded=True):
+            # Create PDF link by replacing 'abs' with 'pdf' in arxiv URL
+            pdf_url = paper['url'].replace('/abs/', '/pdf/')
+            st.markdown(f"""
+                **{paper['date']} | {paper['title']}**  
+                📄 [Abstract]({paper['url']}) | 📑 [PDF]({pdf_url})
+            """)
+            st.markdown(f"*Authors:* {paper['authors']}")
+            st.markdown(paper['summary'])
+            if paper.get('full_audio'):
+                st.write("📚 Paper Audio")
+                st.audio(paper['full_audio'])
+                if paper['download_base64']:
+                    st.markdown(paper['download_base64'], unsafe_allow_html=True)
+
+def display_papers_in_sidebar(papers):
+    """Mirrors the paper listing in sidebar with lazy loading."""
+    st.sidebar.title("🎶 Papers & Audio")
+    for i, paper in enumerate(papers, start=1):
+        paper_key = f"paper_{paper['url']}"
+        if paper_key not in st.session_state:
+            st.session_state[paper_key] = False
+
+        with st.sidebar.expander(f"{i}. {paper['title']}"):
+            # Create PDF link
+            pdf_url = paper['url'].replace('/abs/', '/pdf/')
+            st.markdown(f"📄 [Abstract]({paper['url']}) | 📑 [PDF]({pdf_url})")
+            
+            # Preview of authors and summary
+            st.markdown(f"**Authors:** {paper['authors'][:100]}...")
+            if paper['summary']:
+                st.markdown(f"**Summary:** {paper['summary'][:200]}...")
+            
+            # Load audio button
+            if paper['full_audio'] and st.button("🎵 Load Audio", 
+                                               key=f"btn_{paper_key}"):
+                st.session_state[paper_key] = True
+            
+            # Show audio player and download only if requested
+            if st.session_state[paper_key] and paper['full_audio']:
+                st.audio(paper['full_audio'])
+                if paper['download_base64']:
+                    st.markdown(paper['download_base64'], unsafe_allow_html=True)
+                    
+# ─────────────────────────────────────────────────────────
+# 4. ZIP FUNCTION
+# ─────────────────────────────────────────────────────────
+
+def create_zip_of_files(md_files, mp3_files, wav_files, input_question):
+    """
+    Zip up all relevant files, limiting the final zip name to ~20 chars
+    to avoid overly long base64 strings. 
+    """
+    md_files = [f for f in md_files if os.path.basename(f).lower() != 'readme.md']
+    all_files = md_files + mp3_files + wav_files
+    if not all_files:
+        return None
+
+    all_content = []
+    for f in all_files:
+        if f.endswith('.md'):
+            with open(f, 'r', encoding='utf-8') as file:
+                all_content.append(file.read())
+        elif f.endswith('.mp3') or f.endswith('.wav'):
+            basename = os.path.splitext(os.path.basename(f))[0]
+            words = basename.replace('_', ' ')
+            all_content.append(words)
+    
+    all_content.append(input_question)
+    combined_content = " ".join(all_content)
+    info_terms = get_high_info_terms(combined_content, top_n=10)
+    
+    timestamp = format_timestamp_prefix()
+    name_text = '-'.join(term for term in info_terms[:5])  
+    short_zip_name = (timestamp + "_" + name_text)[:20] + ".zip"
+
+    with zipfile.ZipFile(short_zip_name, 'w') as z:
+        for f in all_files:
+            z.write(f)
+    return short_zip_name
+
+# ─────────────────────────────────────────────────────────
+# 5. MAIN LOGIC: AI LOOKUP & VOICE INPUT
+# ─────────────────────────────────────────────────────────
+
+def perform_ai_lookup(q, vocal_summary=True, extended_refs=False, 
+                     titles_summary=True, full_audio=False):
+    """Main routine that uses Anthropic (Claude) + Gradio ArXiv RAG pipeline."""
+    start = time.time()
+    ai_constitution = """
+    You are a talented AI coder and songwriter...
+    """
+
+    # --- 1) Claude API
+    client = anthropic.Anthropic(api_key=anthropic_key)
+    user_input = q
+    response = client.messages.create(
+        model="claude-3-sonnet-20240229",
+        max_tokens=1000,
+        messages=[
+            {"role": "user", "content": user_input}
+        ])
+    st.write("Claude's reply 🧠:")
+    st.markdown(response.content[0].text)
+
+    # Save & produce audio
+    result = response.content[0].text
+    create_file(q, result) 
+    md_file, audio_file = save_qa_with_audio(q, result)
+    st.subheader("📝 Main Response Audio")
+    play_and_download_audio(audio_file, st.session_state['audio_format'])
+
+    # --- 2) Arxiv RAG
+    #st.write("Arxiv's AI this Evening is Mixtral 8x7B...")
+    st.write('Running Arxiv RAG with Claude inputs.')
+    client = Client("awacke1/Arxiv-Paper-Search-And-QA-RAG-Pattern")
+    refs = client.predict(
+        q, 
+        10, 
+        "Semantic Search", 
+        "mistralai/Mixtral-8x7B-Instruct-v0.1",
+        api_name="/update_with_rag_md"
+    )[0]
+
+    #r2 = client.predict(
+    #    q, 
+    #    "mistralai/Mixtral-8x7B-Instruct-v0.1", 
+    #    True, 
+    #    api_name="/ask_llm"
+    #)
+
+    # --- 3) Claude API with arxiv list of papers to app.py
+    client = anthropic.Anthropic(api_key=anthropic_key)
+    user_input = q + '\n\n' + 'Use the paper list below to answer the question thinking through step by step how to create a streamlit app.py and requirements.txt for the solution that answers the questions with a working app to demonstrate.'+ '\n\n'
+    response = client.messages.create(
+        model="claude-3-sonnet-20240229",
+        max_tokens=1000,
+        messages=[
+            {"role": "user", "content": user_input}
+        ])
+    r2 = response.content[0].text
+    st.write("Claude's reply 🧠:")
+    st.markdown(r2)
+    
+    #result = f"### 🔎 {q}\n\n{r2}\n\n{refs}"
+    result = f"🔎 {r2}\n\n{refs}"
+    md_file, audio_file = save_qa_with_audio(q, result)
+    st.subheader("📝 Main Response Audio")
+    play_and_download_audio(audio_file, st.session_state['audio_format'])
+
+    # --- 3) Parse + handle papers
+    papers = parse_arxiv_refs(refs)
+    if papers:
+        # Create minimal links page first
+        paper_links = create_paper_links_md(papers)
+        links_file = create_file(q, paper_links, "md")
+        st.markdown(paper_links)
+
+        # Then create audio for each paper
+        create_paper_audio_files(papers, input_question=q)
+        display_papers(papers, get_marquee_settings())
+        display_papers_in_sidebar(papers)
+    else:
+        st.warning("No papers found in the response.")
+
+    elapsed = time.time() - start
+    st.write(f"**Total Elapsed:** {elapsed:.2f} s")
+    return result
+
+def process_voice_input(text):
+    """When user sends voice query, we run the AI lookup + Q&A with audio."""
+    if not text:
+        return
+    st.subheader("🔍 Search Results")
+    result = perform_ai_lookup(
+        text, 
+        vocal_summary=True,
+        extended_refs=False,
+        titles_summary=True,
+        full_audio=True
+    )
+    md_file, audio_file = save_qa_with_audio(text, result)
+    st.subheader("📝 Generated Files")
+    st.write(f"Markdown: {md_file}")
+    st.write(f"Audio: {audio_file}")
+    play_and_download_audio(audio_file, st.session_state['audio_format'])
+
+# ─────────────────────────────────────────────────────────
+# 6. FILE HISTORY SIDEBAR
+# ─────────────────────────────────────────────────────────
+
+def display_file_history_in_sidebar():
+    """
+    Shows a history of each local .md, .mp3, .wav file in descending
+    order of modification time, with quick icons and optional download links.
+    """
+    st.sidebar.markdown("---")
+    st.sidebar.markdown("### 📂 File History")
+
+    # Gather all files
+    md_files = glob.glob("*.md")
+    mp3_files = glob.glob("*.mp3")
+    wav_files = glob.glob("*.wav")
+    all_files = md_files + mp3_files + wav_files
+
+    if not all_files:
+        st.sidebar.write("No files found.")
+        return
+
+    # 🗑⬇️ Sidebar delete all and zip all download
+    col1, col4 = st.sidebar.columns(2)
+    with col1:
+        if st.button("🗑 Delete All"):
+            for f in all_md:
+                os.remove(f)
+            for f in all_mp3:
+                os.remove(f)
+            for f in all_wav:
+                os.remove(f)
+            st.session_state.should_rerun = True
+    with col4:
+        if st.button("⬇️ Zip All"):
+            zip_name = create_zip_of_files(md_files, mp3_files, wav_files, st.session_state.get('last_query', ''))
+            if zip_name:
+                st.sidebar.markdown(get_download_link(zip_name, "zip"), unsafe_allow_html=True)
+    
+    # Sort newest first
+    all_files = sorted(all_files, key=os.path.getmtime, reverse=True)
+
+    for f in all_files:
+        fname = os.path.basename(f)
+        ext = os.path.splitext(fname)[1].lower().strip('.')
+        emoji = FILE_EMOJIS.get(ext, '📦')
+        time_str = datetime.fromtimestamp(os.path.getmtime(f)).strftime("%Y-%m-%d %H:%M:%S")
+
+        with st.sidebar.expander(f"{emoji} {fname}"):
+            st.write(f"**Modified:** {time_str}")
+            if ext == "md":
+                with open(f, "r", encoding="utf-8") as file_in:
+                    snippet = file_in.read(200).replace("\n", " ")
+                if len(snippet) == 200:
+                    snippet += "..."
+                st.write(snippet)
+                st.markdown(get_download_link(f, file_type="md"), unsafe_allow_html=True)
+            elif ext in ["mp3","wav"]:
+                st.audio(f)
+                st.markdown(get_download_link(f, file_type=ext), unsafe_allow_html=True)
+            else:
+                st.markdown(get_download_link(f), unsafe_allow_html=True)
+
+# ─────────────────────────────────────────────────────────
+# 7. MAIN APP
+# ─────────────────────────────────────────────────────────
+
+def main():
+    # 1) Setup marquee UI in the sidebar
+    update_marquee_settings_ui()
+    marquee_settings = get_marquee_settings()
+
+    # 2) Display the marquee welcome
+    display_marquee(st.session_state['marquee_content'], 
+                    {**marquee_settings, "font-size": "28px", "lineHeight": "50px"},
+                    key_suffix="welcome")
+
+    # 3) Main action tabs
+    tab_main = st.radio("Action:", ["🎤 Voice", "📸 Media", "🔍 ArXiv", "📝 Editor"], 
+                        horizontal=True)
+
+    # Example custom component usage
+    mycomponent = components.declare_component("mycomponent", path="mycomponent")
+    val = mycomponent(my_input_value="Hello")
+
+    if val:
+        val_stripped = val.replace('\\n', ' ')
+        edited_input = st.text_area("✏️ Edit Input:", value=val_stripped, height=100)
+        run_option = st.selectbox("Model:", ["Arxiv"])
+        col1, col2 = st.columns(2)
+        with col1:
+            autorun = st.checkbox("⚙ AutoRun", value=True)
+        with col2:
+            full_audio = st.checkbox("📚FullAudio", value=False)
+
+        input_changed = (val != st.session_state.old_val)
+
+        if autorun and input_changed:
+            st.session_state.old_val = val
+            st.session_state.last_query = edited_input
+            perform_ai_lookup(edited_input, 
+                              vocal_summary=True, 
+                              extended_refs=False, 
+                              titles_summary=True, 
+                              full_audio=full_audio)
+        else:
+            if st.button("▶ Run"):
+                st.session_state.old_val = val
+                st.session_state.last_query = edited_input
+                perform_ai_lookup(edited_input, 
+                                  vocal_summary=True, 
+                                  extended_refs=False, 
+                                  titles_summary=True, 
+                                  full_audio=full_audio)
+    
+    # ─────────────────────────────────────────────────────────
+    # TAB: ArXiv
+    # ─────────────────────────────────────────────────────────
+    if tab_main == "🔍 ArXiv":
+        st.subheader("🔍 Query ArXiv")
+        q = st.text_input("🔍 Query:", key="arxiv_query")
+        
+        st.markdown("### 🎛 Options")
+        vocal_summary = st.checkbox("🎙ShortAudio", value=True, key="option_vocal_summary")
+        extended_refs = st.checkbox("📜LongRefs", value=False, key="option_extended_refs")
+        titles_summary = st.checkbox("🔖TitlesOnly", value=True, key="option_titles_summary")
+        full_audio = st.checkbox("📚FullAudio", value=False, key="option_full_audio")
+        full_transcript = st.checkbox("🧾FullTranscript", value=False, key="option_full_transcript")
+        
+        if q and st.button("🔍Run"):
+            st.session_state.last_query = q
+            result = perform_ai_lookup(q, vocal_summary=vocal_summary, extended_refs=extended_refs, 
+                                       titles_summary=titles_summary, full_audio=full_audio)
+            if full_transcript:
+                create_file(q, result, "md")
+
+    # ─────────────────────────────────────────────────────────
+    # TAB: Voice
+    # ─────────────────────────────────────────────────────────
+    elif tab_main == "🎤 Voice":
+        st.subheader("🎤 Voice Input")
+        
+        st.markdown("### 🎤 Voice Settings")
+        selected_voice = st.selectbox(
+            "Select TTS Voice:",
+            options=EDGE_TTS_VOICES,
+            index=EDGE_TTS_VOICES.index(st.session_state['tts_voice'])
+        )
+        
+        st.markdown("### 🔊 Audio Format")
+        selected_format = st.radio(
+            "Choose Audio Format:",
+            options=["MP3", "WAV"],
+            index=0
+        )
+
+        # Update session state if voice/format changes
+        if selected_voice != st.session_state['tts_voice']:
+            st.session_state['tts_voice'] = selected_voice
+            st.rerun()
+        if selected_format.lower() != st.session_state['audio_format']:
+            st.session_state['audio_format'] = selected_format.lower()
+            st.rerun()
+
+        # Input text
+        user_text = st.text_area("💬 Message:", height=100)
+        user_text = user_text.strip().replace('\n', ' ')
+
+        if st.button("📨 Send"):
+            process_voice_input(user_text)
+
+        st.subheader("📜 Chat History")
+        for c in st.session_state.chat_history:
+            st.write("**You:**", c["user"])
+            st.write("**Response:**", c["claude"])
+
+    # ─────────────────────────────────────────────────────────
+    # TAB: Media
+    # ─────────────────────────────────────────────────────────
+    elif tab_main == "📸 Media":
+        st.header("📸 Media Gallery")
+
+        # By default, show audio first
+        tabs = st.tabs(["🎵 Audio", "🖼 Images", "🎥 Video"])
+        
+        # AUDIO sub-tab
+        with tabs[0]:
+            st.subheader("🎵 Audio Files")
+            audio_files = glob.glob("*.mp3") + glob.glob("*.wav")
+            if audio_files:
+                for a in audio_files:
+                    with st.expander(os.path.basename(a)):
+                        st.audio(a)
+                        ext = os.path.splitext(a)[1].replace('.', '')
+                        dl_link = get_download_link(a, file_type=ext)
+                        st.markdown(dl_link, unsafe_allow_html=True)
+            else:
+                st.write("No audio files found.")
+        
+        # IMAGES sub-tab
+        with tabs[1]:
+            st.subheader("🖼 Image Files")
+            imgs = glob.glob("*.png") + glob.glob("*.jpg") + glob.glob("*.jpeg")
+            if imgs:
+                c = st.slider("Cols", 1, 5, 3, key="cols_images")
+                cols = st.columns(c)
+                for i, f in enumerate(imgs):
+                    with cols[i % c]:
+                        st.image(Image.open(f), use_container_width=True)
+            else:
+                st.write("No images found.")
+        
+        # VIDEO sub-tab
+        with tabs[2]:
+            st.subheader("🎥 Video Files")
+            vids = glob.glob("*.mp4") + glob.glob("*.mov") + glob.glob("*.avi")
+            if vids:
+                for v in vids:
+                    with st.expander(os.path.basename(v)):
+                        st.video(v)
+            else:
+                st.write("No videos found.")
+
+    # ─────────────────────────────────────────────────────────
+    # TAB: Editor
+    # ─────────────────────────────────────────────────────────
+    elif tab_main == "📝 Editor":
+        st.write("Select or create a file to edit. (Currently minimal demo)")
+
+    # ─────────────────────────────────────────────────────────
+    # SIDEBAR: FILE HISTORY
+    # ─────────────────────────────────────────────────────────
+    display_file_history_in_sidebar()
+
+    # Some light CSS styling
+    st.markdown("""
+    <style>
+        .main { background: linear-gradient(to right, #1a1a1a, #2d2d2d); color: #fff; }
+        .stMarkdown { font-family: 'Helvetica Neue', sans-serif; }
+        .stButton>button { margin-right: 0.5rem; }
+    </style>
+    """, unsafe_allow_html=True)
+
+    # Rerun if needed
+    if st.session_state.should_rerun:
+        st.session_state.should_rerun = False
+        st.rerun()
+
+if __name__ == "__main__":
+    main()