Update app.py

app.py

@@ -68,6 +68,7 @@ QUOTE_VOTES_FILE = "quote_votes.md"
 MEDIA_VOTES_FILE = "media_votes.md"
 HISTORY_FILE = "chat_history.md"
 STATE_FILE = "user_state.txt"
+MEDIA_DIR = "media_base64"
 
 # Fancy digits - numbers got style! 🔢💃
 UNICODE_DIGITS = {i: f"{i}\uFE0F⃣" for i in range(10)}
@@ -267,6 +268,19 @@ async def save_chat_entry(username, message, is_markdown=False, quote_line=None,
         with open(CHAT_FILE, 'a') as f:
             f.write(f"{indent}[{timestamp}] Audio: {audio_filename}\n")
     if media_file:
+        if isinstance(media_file, Image.Image):
+            img_hash = compute_image_hash(media_file)
+            timestamp_prefix = format_timestamp_prefix(username)
+            media_filename = f"{timestamp_prefix}-{img_hash}.b64"
+            media_path = os.path.join(MEDIA_DIR, media_filename)
+            os.makedirs(MEDIA_DIR, exist_ok=True)
+            img_byte_arr = io.BytesIO()
+            media_file.save(img_byte_arr, format='PNG')
+            img_bytes = img_byte_arr.getvalue()
+            b64_data = base64.b64encode(img_bytes).decode()
+            with open(media_path, 'w') as f:
+                f.write(b64_data)
+            media_file = media_filename
         with open(CHAT_FILE, 'a') as f:
             f.write(f"{indent}[{timestamp}] Media: ![Media]({media_file})\n")
         with open(user_history_file, 'a') as f:
@@ -391,7 +405,7 @@ def play_and_download_audio(file_path):
             audio_data = f.read()
         b64 = base64.b64encode(audio_data).decode()
         audio_html = f'''
-        <audio controls style="display:inline; vertical-align:middle;">
+        <audio controls style="display:inline; vertical-align:middle; width:100px;">
             <source src="data:audio/mpeg;base64,{b64}" type="audio/mpeg">
         </audio>
         <a href="data:audio/mpeg;base64,{b64}" download="{os.path.basename(file_path)}" style="vertical-align:middle;">🎵</a>
@@ -399,19 +413,42 @@ def play_and_download_audio(file_path):
         return audio_html
     return ""
 
-# Image saver - pics preserved with naming! 📸💾
+# Image saver - pics preserved with naming as base64! 📸💾
 async def save_pasted_image(image, username):
     await asyncio.to_thread(log_action, username, "📸💾 - 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 = filename  # Top-level file
-    await asyncio.to_thread(image.save, filepath, "PNG")
+    voice = st.session_state.voice if username == st.session_state.username else FUN_USERNAMES.get(username, "en-US-AriaNeural")
+    filename = f"{timestamp}-{img_hash}-voice-{voice}.b64"
+    filepath = os.path.join(MEDIA_DIR, filename)
+    os.makedirs(MEDIA_DIR, exist_ok=True)
+    img_byte_arr = io.BytesIO()
+    image.save(img_byte_arr, format='PNG')
+    b64_data = base64.b64encode(img_byte_arr.getvalue()).decode()
+    with open(filepath, 'w') as f:
+        f.write(b64_data)
     st.session_state.image_hashes.add(img_hash)
     await save_chat_history_with_image(username, filepath)
-    return filepath
+    return filename
+
+# Display base64 image and audio
+def display_base64_media(media_file, width="100px"):
+    if os.path.exists(os.path.join(MEDIA_DIR, media_file)) and media_file.endswith('.b64'):
+        with open(os.path.join(MEDIA_DIR, media_file), 'r') as f:
+            b64_data = f.read()
+        img_data = base64.b64decode(b64_data)
+        img = Image.open(io.BytesIO(img_data))
+        st.image(img, use_container_width=True)
+        
+        # Find corresponding audio file
+        timestamp = media_file.split('-by-')[0] + '-by-' + media_file.split('-by-')[1].split('-')[0]
+        voice = media_file.split('-voice-')[1].split('.b64')[0]
+        audio_files = glob.glob(f"{timestamp}*-{voice}.mp3")
+        if audio_files:
+            audio_file = audio_files[0]
+            st.audio(audio_file)
 
 # PDF saver and audio generator
 async def save_pdf_and_generate_audio(pdf_file, username, max_pages=10):
@@ -433,7 +470,7 @@ async def save_pdf_and_generate_audio(pdf_file, username, max_pages=10):
     for i in range(total_pages):
         text = reader.pages[i].extract_text()
         texts.append(text)
-        audio_filename = f"{timestamp}-page{i+1}-{file_hash}.mp3"
+        audio_filename = f"{timestamp}-page{i+1}-{file_hash}-voice-{voice}.mp3"
         audio_data = await audio_processor.create_audio(text, voice, audio_filename)
         if audio_data:
             audio_files.append(audio_filename)
@@ -441,7 +478,7 @@ async def save_pdf_and_generate_audio(pdf_file, username, max_pages=10):
     return pdf_filename, texts, audio_files
 
 # Video renderer - movies roll with autoplay! 🎥🎬
-def get_video_html(video_path, width="100%"):
+def get_video_html(video_path, width="100px"):
     video_url = f"data:video/mp4;base64,{base64.b64encode(open(video_path, 'rb').read()).decode()}"
     return f'''
     <video width="{width}" controls autoplay muted loop>
@@ -451,7 +488,7 @@ def get_video_html(video_path, width="100%"):
     '''
 
 # Audio renderer - sounds soar! 🎶✈️
-async def get_audio_html(audio_path, width="100%"):
+async def get_audio_html(audio_path, width="100px"):
     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()}"
@@ -510,7 +547,7 @@ async def run_websocket_server():
 
 # Delete all user files function
 def delete_user_files():
-    protected_files = {'app.py', 'requirements.txt', 'README.md', CHAT_FILE, QUOTE_VOTES_FILE, MEDIA_VOTES_FILE, HISTORY_FILE, STATE_FILE}
+    protected_files = {'app.py', 'requirements.txt', 'README.md', CHAT_FILE, QUOTE_VOTES_FILE, MEDIA_VOTES_FILE, HISTORY_FILE, STATE_FILE, MEDIA_DIR}
     deleted_files = []
     for file in os.listdir('.'):
         if file not in protected_files and not file.endswith('_history.md'):
@@ -519,6 +556,14 @@ def delete_user_files():
                 deleted_files.append(file)
             except Exception as e:
                 st.error(f"Failed to delete {file}: {e}")
+    for root, dirs, files in os.walk(MEDIA_DIR):
+        for file in files:
+            file_path = os.path.join(root, file)
+            try:
+                os.remove(file_path)
+                deleted_files.append(file_path)
+            except Exception as e:
+                st.error(f"Failed to delete {file_path}: {e}")
     st.session_state.image_hashes.clear()
     st.session_state.audio_cache.clear()
     st.session_state.base64_cache.clear()
@@ -532,9 +577,10 @@ def check_query_params():
     if q_value and q_value in FUN_USERNAMES:
         st.session_state.username = q_value
         st.session_state.voice = FUN_USERNAMES[q_value]
+        return q_value
     elif q_value:
         st.session_state.user_id = q_value  # Use as user_id if not a valid username
-    return q_value
+    return None
 
 # Mermaid graph generator
 def generate_mermaid_graph(chat_lines):
@@ -542,33 +588,30 @@ def generate_mermaid_graph(chat_lines):
     nodes = {}
     edges = []
     for i, line in enumerate(chat_lines):
-        if line.strip():
+        if line.strip() and not line.startswith('  '):
             timestamp = line.split('] ')[0][1:] if '] ' in line else "Unknown"
-            minute_key = timestamp[:16]  # Up to minute
             content = line.split(': ', 1)[1] if ': ' in line else line
-            user = content.split(' ')[0] if ' ' in content else "System"
+            user = content.split(' ')[0]
+            message = content.split(' ', 1)[1] if ' ' in content else ''
             node_id = f"{user}_{i}"
-            nodes[node_id] = f"{user} @ {timestamp}"
-            
-            if "Audio:" in line:
-                audio_file = line.split("Audio: ")[-1].strip()
-                nodes[f"audio_{i}"] = f"🎵 {os.path.basename(audio_file)}"
-                edges.append(f"{node_id} --> audio_{i}")
-            elif "Media:" in line:
-                media_file = line.split("Media: ")[-1].strip('![]()')
-                media_type = "🖼" if media_file.endswith(('.png', '.jpg')) else "🎥" if media_file.endswith('.mp4') else "📜"
-                nodes[f"media_{i}"] = f"{media_type} {os.path.basename(media_file)}"
-                edges.append(f"{node_id} --> media_{i}")
-            elif "> " in line:  # Quote
-                parent_line = chat_lines[i-1] if i > 0 else None
-                if parent_line:
-                    parent_id = f"{parent_line.split(': ')[1].split(' ')[0]}_{i-1}"
-                    edges.append(f"{parent_id} --> {node_id}")
+            nodes[node_id] = f"{user}: {message}"
+            if i + 1 < len(chat_lines) and "Audio:" in chat_lines[i + 1]:
+                audio_node = f"audio_{i}"
+                nodes[audio_node] = "🎵"
+                edges.append(f"{node_id} --> {audio_node}")
+            if i + 2 < len(chat_lines) and "Media:" in chat_lines[i + 2]:
+                media_node = f"media_{i}"
+                nodes[media_node] = "🖼"
+                edges.append(f"{node_id} --> {media_node}")
+            if i > 0 and "> " in line:
+                parent_user = chat_lines[i-1].split(': ')[1].split(' ')[0]
+                parent_id = f"{parent_user}_{i-1}"
+                edges.append(f"{parent_id} --> {node_id}")
 
     for node_id, label in nodes.items():
         mermaid_code += f"    {node_id}[\"{label}\"]\n"
     mermaid_code += "\n".join(f"    {edge}" for edge in edges)
-    return f"```mermaid\n{mermaid_code}\n```"
+    return mermaid_code
 
 # Main execution - let’s roll! 🎲🚀
 def main():
@@ -582,12 +625,13 @@ def main():
         if not st.session_state.user_id:
             st.session_state.user_id = str(uuid.uuid4())
             st.session_state.user_hash = await generate_user_hash()
-        if 'username' not in st.session_state:
-            q_value = check_query_params()
-            if not q_value or q_value not in FUN_USERNAMES:
-                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()))
+
+        # Check query params first to override username
+        q_value = check_query_params()
+        if not q_value and '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"**🎙️ Voice Selected**: {st.session_state.voice} 🗣️ for {st.session_state.username}")
 
@@ -598,229 +642,254 @@ def main():
                 st.session_state.displayed_chat_lines = f.read().split('\n')
 
         user_url = f"/q={st.session_state.username}"
-        st.markdown(f"Your unique URL path: [{user_url}]({user_url})", unsafe_allow_html=True)
+        with st.container():
+            st.markdown(f"<small>Your unique URL path: [{user_url}]({user_url})</small>", unsafe_allow_html=True)
 
-        st.title(f"🤖🧠MMO {st.session_state.username}📝🔬")
-        st.markdown(f"Welcome to {START_ROOM} - chat, vote, upload, paste images, and enjoy quoting! 🎉 User ID: {st.session_state.user_id}")
+        with st.container():
+            st.markdown(f"#### 🤖🧠MMO {st.session_state.username}📝🔬")
+            st.markdown(f"<small>Welcome to {START_ROOM} - chat, vote, upload, paste images, and enjoy quoting! 🎉 User ID: {st.session_state.user_id}</small>", unsafe_allow_html=True)
 
         if not st.session_state.server_task:
             st.session_state.server_task = loop.create_task(run_websocket_server())
 
         # Unified Chat History at Top with Markdown Emoji Output
-        st.subheader(f"{START_ROOM} Chat History 💬")
-        chat_content = await load_chat()
-        chat_lines = [line for line in chat_content.split('\n') if line.strip() and not line.startswith('#')]
-        if chat_lines:
-            chat_by_minute = {}
-            for line in reversed(chat_lines):
-                timestamp = line.split('] ')[0][1:] if '] ' in line else "Unknown"
-                minute_key = timestamp[:16]  # Up to minute
-                if minute_key not in chat_by_minute:
-                    chat_by_minute[minute_key] = []
-                chat_by_minute[minute_key].append(line)
-
-            markdown_output = ""
-            for minute, lines in chat_by_minute.items():
-                minute_output = f"### {minute}\n"
-                for line in lines:
-                    if ': ' in line and not line.startswith('  '):
-                        user_message = line.split(': ', 1)[1]
-                        minute_output += f"- 💬 **{user_message.split(' ')[0]}**: {user_message.split(' ', 1)[1] if ' ' in user_message else ''}\n"
-                    elif "Audio:" in line:
-                        audio_file = line.split("Audio: ")[-1].strip()
-                        audio_html = play_and_download_audio(audio_file)
-                        minute_output += f"  - 🎵 {audio_html}\n"
-                    elif "Media:" in line:
-                        media_file = line.split("Media: ")[-1].strip('![]()')
-                        if media_file.endswith(('.png', '.jpg')):
-                            minute_output += f"  - 🖼 <img src='file://{media_file}' width='100'>\n"
-                        elif media_file.endswith('.mp4'):
-                            minute_output += f"  - 🎥 <video src='file://{media_file}' width='100' controls autoplay muted loop></video>\n"
-                        elif media_file.endswith('.pdf'):
-                            minute_output += f"  - 📜 {os.path.basename(media_file)}\n"
-                markdown_output += minute_output
-            st.markdown(markdown_output, unsafe_allow_html=True)
+        with st.container():
+            st.markdown(f"##### {START_ROOM} Chat History 💬")
+            chat_content = await load_chat()
+            chat_lines = [line for line in chat_content.split('\n') if line.strip() and not line.startswith('#')]
+            if chat_lines:
+                chat_by_minute = {}
+                for line in reversed(chat_lines):
+                    timestamp = line.split('] ')[0][1:] if '] ' in line else "Unknown"
+                    minute_key = timestamp[:16]  # Up to minute
+                    if minute_key not in chat_by_minute:
+                        chat_by_minute[minute_key] = []
+                    chat_by_minute[minute_key].append(line)
+
+                markdown_output = ""
+                for minute, lines in chat_by_minute.items():
+                    minute_output = f"###### {minute[-5:]}\n"  # Show only HH:MM
+                    for line in lines:
+                        if ': ' in line and not line.startswith('  '):
+                            user_message = line.split(': ', 1)[1]
+                            user = user_message.split(' ')[0]
+                            msg = user_message.split(' ', 1)[1] if ' ' in user_message else ''
+                            audio_html = ""
+                            media_content = ""
+                            next_lines = chat_lines[chat_lines.index(line)+1:chat_lines.index(line)+3]
+                            for nl in next_lines:
+                                if "Audio:" in nl:
+                                    audio_file = nl.split("Audio: ")[-1].strip()
+                                    audio_html = play_and_download_audio(audio_file)
+                                elif "Media:" in nl:
+                                    media_file = nl.split("Media: ")[-1].strip('![]()')
+                                    if media_file.endswith('.b64'):
+                                        media_content = display_base64_media(media_file, width="100px")
+                                    elif media_file.endswith(('.png', '.jpg')):
+                                        media_content = f"<img src='file://{media_file}' width='100'>"
+                                    elif media_file.endswith('.mp4'):
+                                        media_content = get_video_html(media_file)
+                                    elif media_file.endswith('.pdf'):
+                                        media_content = f"📜 {os.path.basename(media_file)}"
+                            minute_output += f"- 💬 **{user}**: {msg} {audio_html} {media_content}\n"
+                    markdown_output += minute_output
+                st.markdown(markdown_output, unsafe_allow_html=True)
 
             # Mermaid Graph Visualization
-            st.subheader("Chat Relationship Tree 🌳")
-            mermaid_graph = generate_mermaid_graph(chat_lines)
-            st.markdown(mermaid_graph)
-
-        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("📋 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
-                        await save_chat_entry(st.session_state.username, f"Pasted image: {filename}", quote_line=st.session_state.quote_line, media_file=filename)
-            if st.button("Send Quote 🚀", 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}", quote_line=st.session_state.quote_line, media_file=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, quote_line=st.session_state.quote_line)
-                st.session_state.quote_line = None
-                st.session_state.message_text = ''
+            st.markdown("###### Chat Relationship Tree 🌳")
+            mermaid_code = generate_mermaid_graph(chat_lines)
+            mermaid_html = f"""
+            <script src="https://cdn.jsdelivr.net/npm/mermaid/dist/mermaid.min.js"></script>
+            <div class="mermaid" style="height: 200px; overflow: auto;">{mermaid_code}</div>
+            <script>mermaid.initialize({{startOnLoad:true}});</script>
+            """
+            components.html(mermaid_html, height=250)
+
+        with st.container():
+            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("📋 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
+                            await save_chat_entry(st.session_state.username, f"Pasted image: {filename}", quote_line=st.session_state.quote_line, media_file=paste_result_quote.image_data)
+                if st.button("Send Quote 🚀", 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}", quote_line=st.session_state.quote_line, media_file=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, quote_line=st.session_state.quote_line)
+                    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"**🎙️ Voice Changed**: {st.session_state.voice} 🗣️ for {st.session_state.username}")
                 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"**🎙️ Voice Changed**: {st.session_state.voice} 🗣️ for {st.session_state.username}")
-            st.rerun()
-
-        # Message input with Send button on the right
-        col_input, col_send = st.columns([5, 1])
-        with col_input:
-            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)
-        with col_send:
-            if st.button("Send 🚀", key="send_button"):
-                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
-                        st.audio(audio_file)  # Immediate preview
-                if st.session_state.pasted_image_data:
-                    await save_chat_entry(st.session_state.username, f"Pasted image: {st.session_state.pasted_image_data}", media_file=st.session_state.pasted_image_data)
-                    st.session_state.pasted_image_data = None
-                st.session_state.message_text = ''
-                st.rerun()
-
-        paste_result_msg = paste_image_button("📋 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:
-                    await save_chat_entry(st.session_state.username, f"Pasted image: {filename}", media_file=filename)
-                    st.session_state.pasted_image_data = None
-
-        tab_main = st.radio("Action:", ["📸 Media", "🔍 ArXiv", "📝 Editor"], horizontal=True)
-        useArxiv = st.checkbox("Search Arxiv for Research Paper Answers", value=True)
-        useArxivAudio = st.checkbox("Generate Audio File for Research Paper Answers", value=False)
-
-        st.subheader("Upload Media 🎨🎶📜🎥")
-        uploaded_file = st.file_uploader("Upload Media", type=['png', 'jpg', 'mp4', 'mp3', 'wav', 'pdf', 'txt', 'md', 'py'])
-        if uploaded_file:
-            timestamp = format_timestamp_prefix(st.session_state.username)
-            username = st.session_state.username
-            ext = uploaded_file.name.split('.')[-1]
-            file_hash = hashlib.md5(uploaded_file.getbuffer()).hexdigest()[:8]
-            if file_hash not in st.session_state.image_hashes:
-                filename = f"{timestamp}-{file_hash}.{ext}"
-                file_path = filename  # Top-level file
-                await asyncio.to_thread(lambda: open(file_path, 'wb').write(uploaded_file.getbuffer()))
-                st.success(f"Uploaded {filename}")
-                if ext == 'pdf':
-                    pdf_filename, texts, audio_files = await save_pdf_and_generate_audio(uploaded_file, username)
-                    await save_chat_entry(username, f"Uploaded PDF: {pdf_filename}", media_file=pdf_filename)
-                    for i, (text, audio_file) in enumerate(zip(texts, audio_files)):
+            # Message input with Send button on the right
+            col_input, col_send = st.columns([5, 1])
+            with col_input:
+                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)
+            with col_send:
+                if st.button("Send 🚀", key="send_button"):
+                    if message.strip():
+                        audio_file = await save_chat_entry(st.session_state.username, message, is_markdown=True)
                         if audio_file:
-                            with open(CHAT_FILE, 'a') as f:
-                                f.write(f"  [{timestamp}] Page {i+1} Audio: {audio_file}\n")
+                            st.session_state.audio_cache[f"{message}_{FUN_USERNAMES[st.session_state.username]}"] = audio_file
+                            st.audio(audio_file)  # Immediate preview
+                    if st.session_state.pasted_image_data:
+                        await save_chat_entry(st.session_state.username, f"Pasted image: {st.session_state.pasted_image_data}", media_file=st.session_state.pasted_image_data)
+                        st.session_state.pasted_image_data = None
+                    st.session_state.message_text = ''
+                    st.rerun()
+
+            paste_result_msg = paste_image_button("📋 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:
-                    await save_chat_entry(username, f"Uploaded media: {file_path}", media_file=file_path)
-                await save_chat_history_with_image(username, file_path)
-                st.session_state.image_hashes.add(file_hash)
-                if file_path.endswith('.mp4'):
-                    st.session_state.media_notifications.append(file_path)
-
-        # Big Red Delete Button
-        st.subheader("🛑 Danger Zone")
-        if st.button("Try Not To Delete It All On Your First Day", key="delete_all", help="Deletes all user-added files!", type="primary", use_container_width=True):
-            deleted_files = delete_user_files()
-            if deleted_files:
-                st.markdown("### 🗑️ Deleted Files:\n" + "\n".join([f"- `{file}`" for file in deleted_files]))
-            else:
-                st.markdown("### 🗑️ Nothing to Delete!")
-            st.rerun()
+                    st.image(paste_result_msg.image_data, caption="Received Image for Quote")
+                    filename = await save_pasted_image(paste_result_msg.image_data, st.session_state.username)
+                    if filename:
+                        await save_chat_entry(st.session_state.username, f"Pasted image: {filename}", media_file=paste_result_msg.image_data)
+                        st.session_state.pasted_image_data = None
+
+        with st.container():
+            tab_main = st.radio("Action:", ["📸 Media", "🔍 ArXiv", "📝 Editor"], horizontal=True, label_visibility="collapsed")
+            useArxiv = st.checkbox("Search Arxiv for Research Paper Answers", value=True, label_visibility="collapsed")
+            useArxivAudio = st.checkbox("Generate Audio File for Research Paper Answers", value=False, label_visibility="collapsed")
+
+            st.markdown("###### Upload Media 🎨🎶📜🎥")
+            uploaded_file = st.file_uploader("Upload Media", type=['png', 'jpg', 'mp4', 'mp3', 'wav', 'pdf', 'txt', 'md', 'py'])
+            if uploaded_file:
+                timestamp = format_timestamp_prefix(st.session_state.username)
+                username = st.session_state.username
+                ext = uploaded_file.name.split('.')[-1]
+                file_hash = hashlib.md5(uploaded_file.getbuffer()).hexdigest()[:8]
+                if file_hash not in st.session_state.image_hashes:
+                    filename = f"{timestamp}-{file_hash}.{ext}"
+                    file_path = filename  # Top-level file
+                    await asyncio.to_thread(lambda: open(file_path, 'wb').write(uploaded_file.getbuffer()))
+                    st.success(f"Uploaded {filename}")
+                    if ext == 'pdf':
+                        pdf_filename, texts, audio_files = await save_pdf_and_generate_audio(uploaded_file, username)
+                        await save_chat_entry(username, f"Uploaded PDF: {pdf_filename}", media_file=pdf_filename)
+                        for i, (text, audio_file) in enumerate(zip(texts, audio_files)):
+                            if audio_file:
+                                with open(CHAT_FILE, 'a') as f:
+                                    f.write(f"  [{timestamp}] Page {i+1} Audio: {audio_file}\n")
+                    else:
+                        await save_chat_entry(username, f"Uploaded media: {file_path}", media_file=file_path)
+                    await save_chat_history_with_image(username, file_path)
+                    st.session_state.image_hashes.add(file_hash)
+                    if file_path.endswith('.mp4'):
+                        st.session_state.media_notifications.append(file_path)
+
+            # Big Red Delete Button
+            st.markdown("###### 🛑 Danger Zone")
+            if st.button("Try Not To Delete It All On Your First Day", key="delete_all", help="Deletes all user-added files!", type="primary", use_container_width=True, label_visibility="collapsed"):
+                deleted_files = delete_user_files()
+                if deleted_files:
+                    st.markdown("### 🗑️ Deleted Files:\n" + "\n".join([f"- `{file}`" for file in deleted_files]))
+                else:
+                    st.markdown("### 🗑️ Nothing to Delete!")
+                st.rerun()
 
-        st.subheader("Refresh ⏳")
-        refresh_rate = st.slider("Refresh Rate", 1, 300, st.session_state.refresh_rate)
-        st.session_state.refresh_rate = refresh_rate
-        timer_placeholder = st.empty()
-        for i in range(st.session_state.refresh_rate, -1, -1):
-            font_name, font_func = random.choice(UNICODE_FONTS)
-            countdown_str = "".join(UNICODE_DIGITS[int(d)] for d in str(i)) if i < 10 else font_func(str(i))
-            timer_placeholder.markdown(f"<p class='timer'>⏳ {font_func('Refresh in:')} {countdown_str}</p>", unsafe_allow_html=True)
-            time.sleep(1)
-        st.rerun()
+            st.markdown("###### Refresh ⏳")
+            refresh_rate = st.slider("Refresh Rate", 1, 300, st.session_state.refresh_rate, label_visibility="collapsed")
+            st.session_state.refresh_rate = refresh_rate
+            timer_placeholder = st.empty()
+            for i in range(st.session_state.refresh_rate, -1, -1):
+                font_name, font_func = random.choice(UNICODE_FONTS)
+                countdown_str = "".join(UNICODE_DIGITS[int(d)] for d in str(i)) if i < 10 else font_func(str(i))
+                timer_placeholder.markdown(f"<small>⏳ {font_func('Refresh in:')} {countdown_str}</small>", unsafe_allow_html=True)
+                time.sleep(1)
+            st.rerun()
 
         # Separate Galleries for Own and Shared Files
-        all_files = glob.glob("*.md") + glob.glob("*.pdf") + glob.glob("*.txt") + glob.glob("*.py") + glob.glob("*.png") + glob.glob("*.jpg") + glob.glob("*.mp3") + glob.glob("*.mp4")
-        own_files = [f for f in all_files if st.session_state.user_id in f]
-        shared_files = [f for f in all_files if f not in own_files and not f in [CHAT_FILE, QUOTE_VOTES_FILE, MEDIA_VOTES_FILE, HISTORY_FILE, STATE_FILE]]
-
-        st.subheader("Your Files 📂")
-        st.subheader("Image Gallery 🖼")
-        own_image_files = [f for f in own_files if f.endswith(('.png', '.jpg'))]
-        image_cols = st.slider("Image Gallery Columns 🖼 (Own)", min_value=1, max_value=15, value=5)
-        cols = st.columns(image_cols)
-        for idx, image_file in enumerate(own_image_files):
-            with cols[idx % image_cols]:
-                st.image(image_file, use_container_width=True)
-
-        st.subheader("Video Gallery 🎥")
-        own_video_files = [f for f in own_files if f.endswith('.mp4')]
-        video_cols = st.slider("Video Gallery Columns 🎬 (Own)", min_value=1, max_value=5, value=3)
-        cols = st.columns(video_cols)
-        for idx, video_file in enumerate(own_video_files):
-            with cols[idx % video_cols]:
-                st.markdown(get_video_html(video_file, width="100%"), unsafe_allow_html=True)
-
-        st.subheader("Audio Gallery 🎧")
-        own_audio_files = [f for f in own_files if f.endswith(('.mp3', '.wav'))]
-        audio_cols = st.slider("Audio Gallery Columns 🎶 (Own)", min_value=1, max_value=15, value=5)
-        cols = st.columns(audio_cols)
-        for idx, audio_file in enumerate(own_audio_files):
-            with cols[idx % audio_cols]:
-                st.markdown(await get_audio_html(audio_file, width="100%"), unsafe_allow_html=True)
-
-        st.subheader("Shared Files 📤")
-        st.subheader("Image Gallery 🖼")
-        shared_image_files = [f for f in shared_files if f.endswith(('.png', '.jpg'))]
-        image_cols = st.slider("Image Gallery Columns 🖼 (Shared)", min_value=1, max_value=15, value=5)
-        cols = st.columns(image_cols)
-        for idx, image_file in enumerate(shared_image_files):
-            with cols[idx % image_cols]:
-                st.image(image_file, use_container_width=True)
-
-        st.subheader("Video Gallery 🎥")
-        shared_video_files = [f for f in shared_files if f.endswith('.mp4')]
-        video_cols = st.slider("Video Gallery Columns 🎬 (Shared)", min_value=1, max_value=5, value=3)
-        cols = st.columns(video_cols)
-        for idx, video_file in enumerate(shared_video_files):
-            with cols[idx % video_cols]:
-                st.markdown(get_video_html(video_file, width="100%"), unsafe_allow_html=True)
-
-        st.subheader("Audio Gallery 🎧")
-        shared_audio_files = [f for f in shared_files if f.endswith(('.mp3', '.wav'))]
-        audio_cols = st.slider("Audio Gallery Columns 🎶 (Shared)", min_value=1, max_value=15, value=5)
-        cols = st.columns(audio_cols)
-        for idx, audio_file in enumerate(shared_audio_files):
-            with cols[idx % audio_cols]:
-                st.markdown(await get_audio_html(audio_file, width="100%"), unsafe_allow_html=True)
+        with st.container():
+            all_files = glob.glob("*.md") + glob.glob("*.pdf") + glob.glob("*.txt") + glob.glob("*.py") + glob.glob("*.png") + glob.glob("*.jpg") + glob.glob("*.mp3") + glob.glob("*.mp4") + glob.glob(os.path.join(MEDIA_DIR, "*.b64"))
+            own_files = [f for f in all_files if st.session_state.user_id in os.path.basename(f)]
+            shared_files = [f for f in all_files if f not in own_files and not f in [CHAT_FILE, QUOTE_VOTES_FILE, MEDIA_VOTES_FILE, HISTORY_FILE, STATE_FILE, os.path.join(MEDIA_DIR, "*")]]
+
+            st.markdown("###### Your Files 📂")
+            st.markdown("###### Image Gallery 🖼")
+            own_image_files = [f for f in own_files if f.endswith(('.png', '.jpg', '.b64'))]
+            image_cols = st.slider("Image Gallery Columns 🖼 (Own)", min_value=1, max_value=15, value=5, label_visibility="collapsed")
+            cols = st.columns(image_cols)
+            for idx, image_file in enumerate(own_image_files):
+                with cols[idx % image_cols]:
+                    if image_file.endswith('.b64'):
+                        display_base64_media(os.path.basename(image_file))
+                    else:
+                        st.image(image_file, use_container_width=True)
+
+            st.markdown("###### Video Gallery 🎥")
+            own_video_files = [f for f in own_files if f.endswith('.mp4')]
+            video_cols = st.slider("Video Gallery Columns 🎬 (Own)", min_value=1, max_value=5, value=3, label_visibility="collapsed")
+            cols = st.columns(video_cols)
+            for idx, video_file in enumerate(own_video_files):
+                with cols[idx % video_cols]:
+                    st.markdown(get_video_html(video_file), unsafe_allow_html=True)
+
+            st.markdown("###### Audio Gallery 🎧")
+            own_audio_files = [f for f in own_files if f.endswith(('.mp3', '.wav'))]
+            audio_cols = st.slider("Audio Gallery Columns 🎶 (Own)", min_value=1, max_value=15, value=5, label_visibility="collapsed")
+            cols = st.columns(audio_cols)
+            for idx, audio_file in enumerate(own_audio_files):
+                with cols[idx % audio_cols]:
+                    st.markdown(await get_audio_html(audio_file), unsafe_allow_html=True)
+
+            st.markdown("###### Shared Files 📤")
+            st.markdown("###### Image Gallery 🖼")
+            shared_image_files = [f for f in shared_files if f.endswith(('.png', '.jpg', '.b64'))]
+            image_cols = st.slider("Image Gallery Columns 🖼 (Shared)", min_value=1, max_value=15, value=5, label_visibility="collapsed")
+            cols = st.columns(image_cols)
+            for idx, image_file in enumerate(shared_image_files):
+                with cols[idx % image_cols]:
+                    if image_file.endswith('.b64'):
+                        display_base64_media(os.path.basename(image_file))
+                    else:
+                        st.image(image_file, use_container_width=True)
+
+            st.markdown("###### Video Gallery 🎥")
+            shared_video_files = [f for f in shared_files if f.endswith('.mp4')]
+            video_cols = st.slider("Video Gallery Columns 🎬 (Shared)", min_value=1, max_value=5, value=3, label_visibility="collapsed")
+            cols = st.columns(video_cols)
+            for idx, video_file in enumerate(shared_video_files):
+                with cols[idx % video_cols]:
+                    st.markdown(get_video_html(video_file), unsafe_allow_html=True)
+
+            st.markdown("###### Audio Gallery 🎧")
+            shared_audio_files = [f for f in shared_files if f.endswith(('.mp3', '.wav'))]
+            audio_cols = st.slider("Audio Gallery Columns 🎶 (Shared)", min_value=1, max_value=15, value=5, label_visibility="collapsed")
+            cols = st.columns(audio_cols)
+            for idx, audio_file in enumerate(shared_audio_files):
+                with cols[idx % audio_cols]:
+                    st.markdown(await get_audio_html(audio_file), unsafe_allow_html=True)
 
         # Full Log at End with Download
-        st.subheader("Full Chat Log 📜")
-        with open(CHAT_FILE, 'r') as f:
-            history_content = f.read()
-        st.markdown(history_content)
-        st.download_button("Download Chat Log as .md", history_content, file_name=f"chat_{st.session_state.user_id}.md", mime="text/markdown")
+        with st.container():
+            st.markdown("###### Full Chat Log 📜")
+            with open(CHAT_FILE, 'r') as f:
+                history_content = f.read()
+            st.markdown(history_content)
+            st.download_button("Download Chat Log as .md", history_content, file_name=f"chat_{st.session_state.user_id}.md", mime="text/markdown", label_visibility="collapsed")
 
     loop.run_until_complete(async_interface())