Spaces:

awacke1
/

SectorMultiplayerChatServer

Running

App Files Files Community

awacke1 commited on Mar 2

Commit

90747a3

verified ·

1 Parent(s): 4f5107e

Update app.py

Browse files

Files changed (1) hide show

app.py +247 -35

app.py CHANGED Viewed

@@ -10,14 +10,13 @@ import time
 import hashlib
 from PIL import Image
 import glob
-from urllib.parse import quote
 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  # Added back the re import for regular expressions
 from streamlit_paste_button import paste_image_button
 # Patch for nested async - sneaky fix! 🐍✨
@@ -129,15 +128,32 @@ if 'pasted_image_data' not in st.session_state:
 if 'quote_line' not in st.session_state:
     st.session_state.quote_line = None
 if 'refresh_rate' not in st.session_state:
-    st.session_state.refresh_rate = 5  # Default refresh rate
 if 'base64_cache' not in st.session_state:
-    st.session_state.base64_cache = {}  # Cache for base64 strings
 # Timestamp wizardry - clock ticks with flair! ⏰🎩
 def format_timestamp_prefix(username):
     now = datetime.now()
     return f"{now.strftime('%I-%M-%p-ct-%m-%d-%Y')}-by-{username}"
 # Node naming - christening the beast! 🌐🍼
 def get_node_name():
     parser = argparse.ArgumentParser(description='Start a chat node with a specific name')
@@ -168,10 +184,13 @@ def clean_text_for_tts(text):
     return cleaned[:200] if cleaned else "No text to speak"
 # Chat saver - words locked tight! 💬🔒
-async def save_chat_entry(username, message):
     await asyncio.to_thread(log_action, username, "💬🔒 - Chat saver - words locked tight!")
-    timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")  # Keep this for log consistency
-    entry = f"[{timestamp}] {username}: {message}"
     await asyncio.to_thread(lambda: open(CHAT_FILE, 'a').write(f"{entry}\n"))
     voice = FUN_USERNAMES.get(username, "en-US-AriaNeural")
     cleaned_message = clean_text_for_tts(message)
@@ -289,10 +308,14 @@ def play_and_download_audio(file_path):
 # Image saver - pics preserved with naming! 📸💾
 async def save_pasted_image(image, username):
     await asyncio.to_thread(log_action, username, "📸💾 - Image saver - pics preserved!")
     timestamp = format_timestamp_prefix(username)
-    filename = f"{timestamp}.png"
     filepath = os.path.join(MEDIA_DIR, filename)
     await asyncio.to_thread(image.save, filepath, "PNG")
     return filepath
 # Video renderer - movies roll with autoplay! 🎥🎬
@@ -375,6 +398,170 @@ async def perform_ai_lookup(query, vocal_summary=True, extended_refs=False, titl
         if audio_file:
             st.audio(audio_file)
 # Main execution - let’s roll! 🎲🚀
 def main():
     NODE_NAME, port = get_node_name()
@@ -399,6 +586,17 @@ def main():
             await process_voice_input(audio_bytes)
             st.rerun()
         # Load and display chat
         st.subheader(f"{START_ROOM} Chat 💬")
         chat_content = await load_chat()
@@ -414,7 +612,15 @@ def main():
         for i, line in enumerate(st.session_state.displayed_chat_lines):
             col1, col2, col3, col4 = st.columns([3, 1, 1, 2])
             with col1:
-                st.markdown(line)
                 if "Pasted image:" in line or "Uploaded media:" in line:
                     file_path = line.split(': ')[-1].strip()
                     if os.path.exists(file_path):
@@ -460,20 +666,23 @@ def main():
         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")
-            paste_result_quote = paste_image_button("📋 Paste Image with Quote", key="paste_button_quote")
             if paste_result_quote.image_data is not None:
-                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  # Store for use in quote submission
             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}")
                     st.session_state.pasted_image_data = None
-                await save_chat_entry(st.session_state.username, markdown_response)
                 st.session_state.quote_line = None
                 st.session_state.message_text = ''
                 st.rerun()
@@ -485,53 +694,56 @@ def main():
             st.rerun()
         message = st.text_input(f"Message as {st.session_state.username}", key="message_input", value=st.session_state.message_text, on_change=lambda: st.session_state.update(message_text=st.session_state.message_input))
-        paste_result_msg = paste_image_button("📋 Paste Image with Message", key="paste_button_msg")
         if paste_result_msg.image_data is not None:
-            st.image(paste_result_msg.image_data, caption="Received Image for Message")
-            filename = await save_pasted_image(paste_result_msg.image_data, st.session_state.username)
-            if filename:
-                st.session_state.pasted_image_data = filename  # Store for use in message submission
         if st.button("Send 🚀", key="send_button") and (message.strip() or st.session_state.pasted_image_data):
             if message.strip():
-                await save_chat_entry(st.session_state.username, message)
             if st.session_state.pasted_image_data:
                 await save_chat_entry(st.session_state.username, f"Pasted image: {st.session_state.pasted_image_data}")
                 st.session_state.pasted_image_data = None
             st.session_state.message_text = ''
             st.rerun()
-        # Main action tabs and model use choices
         tab_main = st.radio("Action:", ["🎤 Voice", "📸 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)
-        # Enhanced Media Gallery with Image, Audio, Video
         st.subheader("Upload Media 🎨🎶🎥")
         uploaded_file = st.file_uploader("Upload Media", type=['png', 'jpg', 'mp4', 'mp3'])
         if uploaded_file:
             timestamp = format_timestamp_prefix(st.session_state.username)
             username = st.session_state.username
             ext = uploaded_file.name.split('.')[-1]
-            filename = f"{timestamp}.{ext}"
-            file_path = os.path.join(MEDIA_DIR, filename)
-            await asyncio.to_thread(lambda: open(file_path, 'wb').write(uploaded_file.getbuffer()))
-            st.success(f"Uploaded {filename}")
-            await save_chat_entry(username, f"Uploaded media: {file_path}")
-            if file_path.endswith('.mp4'):
-                st.session_state.media_notifications.append(file_path)
         st.subheader("Media Gallery 🎨🎶🎥")
         media_files = glob.glob(f"{MEDIA_DIR}/*.png") + glob.glob(f"{MEDIA_DIR}/*.jpg") + glob.glob(f"{MEDIA_DIR}/*.mp4") + glob.glob(f"{MEDIA_DIR}/*.mp3")
         if media_files:
             media_votes = await load_votes(MEDIA_VOTES_FILE)
             st.write("### All Media Uploads")
-            seen_files = set()  # Track unique file paths to prevent duplicates
             for media_file in sorted(media_files, key=os.path.getmtime, reverse=True):
-                if media_file not in seen_files:  # Only add if not already seen
                     seen_files.add(media_file)
                     filename = os.path.basename(media_file)
                     vote_count = media_votes.get(media_file, 0)
                     col1, col2 = st.columns([3, 1])
                     with col1:
                         st.markdown(f"**{filename}**")
@@ -554,7 +766,7 @@ def main():
             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)  # Use synchronous sleep
         st.rerun()
         st.sidebar.subheader("Chat History 📜")

 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
 # Patch for nested async - sneaky fix! 🐍✨
 if 'quote_line' not in st.session_state:
     st.session_state.quote_line = None
 if 'refresh_rate' not in st.session_state:
+    st.session_state.refresh_rate = 5
 if 'base64_cache' not in st.session_state:
+    st.session_state.base64_cache = {}
+if 'transcript_history' not in st.session_state:
+    st.session_state.transcript_history = []
+if 'last_transcript' not in st.session_state:
+    st.session_state.last_transcript = ""
+if 'image_hashes' not in st.session_state:
+    st.session_state.image_hashes = set()  # Track unique image hashes
 # Timestamp wizardry - clock ticks with flair! ⏰🎩
 def format_timestamp_prefix(username):
     now = datetime.now()
     return f"{now.strftime('%I-%M-%p-ct-%m-%d-%Y')}-by-{username}"
+# Compute image hash from binary data
+def compute_image_hash(image_data):
+    if isinstance(image_data, Image.Image):
+        # Convert PIL Image to bytes
+        img_byte_arr = io.BytesIO()
+        image_data.save(img_byte_arr, format='PNG')
+        img_bytes = img_byte_arr.getvalue()
+    else:
+        img_bytes = image_data  # Assume raw bytes if not PIL Image
+    return hashlib.md5(img_bytes).hexdigest()[:8]  # Shortened hash for filename
 # Node naming - christening the beast! 🌐🍼
 def get_node_name():
     parser = argparse.ArgumentParser(description='Start a chat node with a specific name')
     return cleaned[:200] if cleaned else "No text to speak"
 # Chat saver - words locked tight! 💬🔒
+async def save_chat_entry(username, message, is_markdown=False):
     await asyncio.to_thread(log_action, username, "💬🔒 - Chat saver - words locked tight!")
+    timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
+    if is_markdown:
+        entry = f"[{timestamp}] {username}:\n```markdown\n{message}\n```"
+    else:
+        entry = f"[{timestamp}] {username}: {message}"
     await asyncio.to_thread(lambda: open(CHAT_FILE, 'a').write(f"{entry}\n"))
     voice = FUN_USERNAMES.get(username, "en-US-AriaNeural")
     cleaned_message = clean_text_for_tts(message)
 # Image saver - pics preserved with naming! 📸💾
 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  # Suppress duplicate
     timestamp = format_timestamp_prefix(username)
+    filename = f"{timestamp}-{img_hash}.png"
     filepath = os.path.join(MEDIA_DIR, filename)
     await asyncio.to_thread(image.save, filepath, "PNG")
+    st.session_state.image_hashes.add(img_hash)
     return filepath
 # Video renderer - movies roll with autoplay! 🎥🎬
         if audio_file:
             st.audio(audio_file)
+# ASR Component HTML
+ASR_HTML = """
+<html>
+<head>
+    <title>Continuous Speech Demo</title>
+    <style>
+        body {
+            font-family: sans-serif;
+            padding: 20px;
+            max-width: 800px;
+            margin: 0 auto;
+        }
+        button {
+            padding: 10px 20px;
+            margin: 10px 5px;
+            font-size: 16px;
+        }
+        #status {
+            margin: 10px 0;
+            padding: 10px;
+            background: #e8f5e9;
+            border-radius: 4px;
+        }
+        #output {
+            white-space: pre-wrap;
+            padding: 15px;
+            background: #f5f5f5;
+            border-radius: 4px;
+            margin: 10px 0;
+            min-height: 100px;
+            max-height: 400px;
+            overflow-y: auto;
+        }
+        .controls {
+            margin: 10px 0;
+        }
+    </style>
+</head>
+<body>
+    <div class="controls">
+        <button id="start">Start Listening</button>
+        <button id="stop" disabled>Stop Listening</button>
+        <button id="clear">Clear Text</button>
+    </div>
+    <div id="status">Ready</div>
+    <div id="output"></div>
+    <script>
+        if (!('webkitSpeechRecognition' in window)) {
+            alert('Speech recognition not supported');
+        } else {
+            const recognition = new webkitSpeechRecognition();
+            const startButton = document.getElementById('start');
+            const stopButton = document.getElementById('stop');
+            const clearButton = document.getElementById('clear');
+            const status = document.getElementById('status');
+            const output = document.getElementById('output');
+            let fullTranscript = '';
+            let lastUpdateTime = Date.now();
+            recognition.continuous = true;
+            recognition.interimResults = true;
+            const startRecognition = () => {
+                try {
+                    recognition.start();
+                    status.textContent = 'Listening...';
+                    startButton.disabled = true;
+                    stopButton.disabled = false;
+                } catch (e) {
+                    console.error(e);
+                    status.textContent = 'Error: ' + e.message;
+                }
+            };
+            window.addEventListener('load', () => {
+                setTimeout(startRecognition, 1000);
+            });
+            startButton.onclick = startRecognition;
+            stopButton.onclick = () => {
+                recognition.stop();
+                status.textContent = 'Stopped';
+                startButton.disabled = false;
+                stopButton.disabled = true;
+            };
+            clearButton.onclick = () => {
+                fullTranscript = '';
+                output.textContent = '';
+                sendDataToPython({value: '', dataType: "json"});
+            };
+            recognition.onresult = (event) => {
+                let interimTranscript = '';
+                let finalTranscript = '';
+                for (let i = event.resultIndex; i < event.results.length; i++) {
+                    const transcript = event.results[i][0].transcript;
+                    if (event.results[i].isFinal) {
+                        finalTranscript += transcript + '\\n';
+                    } else {
+                        interimTranscript += transcript;
+                    }
+                }
+                if (finalTranscript || (Date.now() - lastUpdateTime > 5000)) {
+                    if (finalTranscript) {
+                        fullTranscript += finalTranscript;
+                    }
+                    lastUpdateTime = Date.now();
+                    output.textContent = fullTranscript + (interimTranscript ? '... ' + interimTranscript : '');
+                    output.scrollTop = output.scrollHeight;
+                    sendDataToPython({value: fullTranscript, dataType: "json"});
+                }
+            };
+            recognition.onend = () => {
+                if (!stopButton.disabled) {
+                    try {
+                        recognition.start();
+                        console.log('Restarted recognition');
+                    } catch (e) {
+                        console.error('Failed to restart recognition:', e);
+                        status.textContent = 'Error restarting: ' + e.message;
+                        startButton.disabled = false;
+                        stopButton.disabled = true;
+                    }
+                }
+            };
+            recognition.onerror = (event) => {
+                console.error('Recognition error:', event.error);
+                status.textContent = 'Error: ' + event.error;
+                if (event.error === 'not-allowed' || event.error === 'service-not-allowed') {
+                    startButton.disabled = false;
+                    stopButton.disabled = true;
+                }
+            };
+        }
+        function sendDataToPython(data) {
+            window.parent.postMessage({
+                isStreamlitMessage: true,
+                type: "streamlit:setComponentValue",
+                ...data
+            }, "*");
+        }
+        window.addEventListener('load', function() {
+            window.setTimeout(function() {
+                window.parent.postMessage({
+                    isStreamlitMessage: true,
+                    type: "streamlit:setFrameHeight",
+                    height: document.documentElement.clientHeight
+                }, "*");
+            }, 0);
+        });
+    </script>
+</body>
+</html>
+"""
 # Main execution - let’s roll! 🎲🚀
 def main():
     NODE_NAME, port = get_node_name()
             await process_voice_input(audio_bytes)
             st.rerun()
+        # Continuous Speech Input (ASR)
+        st.subheader("🎤 Continuous Speech Input")
+        asr_component = components.html(ASR_HTML, height=400)
+        if asr_component and isinstance(asr_component, dict) and 'value' in asr_component:
+            transcript = asr_component['value'].strip()
+            if transcript and transcript != st.session_state.last_transcript:
+                st.session_state.transcript_history.append(transcript)
+                await save_chat_entry(st.session_state.username, transcript, is_markdown=True)
+                st.session_state.last_transcript = transcript
+                st.rerun()
         # Load and display chat
         st.subheader(f"{START_ROOM} Chat 💬")
         chat_content = await load_chat()
         for i, line in enumerate(st.session_state.displayed_chat_lines):
             col1, col2, col3, col4 = st.columns([3, 1, 1, 2])
             with col1:
+                if "```markdown" in line:
+                    # Extract Markdown content
+                    markdown_content = re.search(r'```markdown\n(.*?)```', line, re.DOTALL)
+                    if markdown_content:
+                        st.markdown(markdown_content.group(1))
+                    else:
+                        st.markdown(line)
+                else:
+                    st.markdown(line)
                 if "Pasted image:" in line or "Uploaded media:" in line:
                     file_path = line.split(': ')[-1].strip()
                     if os.path.exists(file_path):
         if 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):  # Text from clipboard
+                    st.session_state.message_text = paste_result_quote.image_data
+                else:  # Image
+                    st.image(paste_result_quote.image_data, caption="Received Image for Quote")
+                    filename = await save_pasted_image(paste_result_quote.image_data, st.session_state.username)
+                    if filename:
+                        st.session_state.pasted_image_data = filename
             if st.button("Send Quote 🚀", key="send_quote"):
                 markdown_response = f"### Quote Response\n- **Original**: {st.session_state.quote_line}\n- **{st.session_state.username} Replies**: {quote_response}"
                 if st.session_state.pasted_image_data:
                     markdown_response += f"\n- **Image**: ![Pasted Image]({st.session_state.pasted_image_data})"
                     await save_chat_entry(st.session_state.username, f"Pasted image: {st.session_state.pasted_image_data}")
                     st.session_state.pasted_image_data = None
+                await save_chat_entry(st.session_state.username, markdown_response, is_markdown=True)
                 st.session_state.quote_line = None
                 st.session_state.message_text = ''
                 st.rerun()
             st.rerun()
         message = st.text_input(f"Message as {st.session_state.username}", key="message_input", value=st.session_state.message_text, on_change=lambda: st.session_state.update(message_text=st.session_state.message_input))
+        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):  # Text from clipboard
+                st.session_state.message_text = paste_result_msg.image_data
+            else:  # Image
+                st.image(paste_result_msg.image_data, caption="Received Image for Message")
+                filename = await save_pasted_image(paste_result_msg.image_data, st.session_state.username)
+                if filename:
+                    st.session_state.pasted_image_data = filename
         if st.button("Send 🚀", key="send_button") and (message.strip() or st.session_state.pasted_image_data):
             if message.strip():
+                await save_chat_entry(st.session_state.username, message, is_markdown=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.message_text = ''
             st.rerun()
         tab_main = st.radio("Action:", ["🎤 Voice", "📸 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'])
         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 = os.path.join(MEDIA_DIR, filename)
+                await asyncio.to_thread(lambda: open(file_path, 'wb').write(uploaded_file.getbuffer()))
+                st.success(f"Uploaded {filename}")
+                await save_chat_entry(username, f"Uploaded media: {file_path}")
+                st.session_state.image_hashes.add(file_hash)
+                if file_path.endswith('.mp4'):
+                    st.session_state.media_notifications.append(file_path)
         st.subheader("Media Gallery 🎨🎶🎥")
         media_files = glob.glob(f"{MEDIA_DIR}/*.png") + glob.glob(f"{MEDIA_DIR}/*.jpg") + glob.glob(f"{MEDIA_DIR}/*.mp4") + glob.glob(f"{MEDIA_DIR}/*.mp3")
         if media_files:
             media_votes = await load_votes(MEDIA_VOTES_FILE)
             st.write("### All Media Uploads")
+            seen_files = set()
             for media_file in sorted(media_files, key=os.path.getmtime, reverse=True):
+                if media_file not in seen_files:
                     seen_files.add(media_file)
                     filename = os.path.basename(media_file)
                     vote_count = media_votes.get(media_file, 0)
                     col1, col2 = st.columns([3, 1])
                     with col1:
                         st.markdown(f"**{filename}**")
             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.sidebar.subheader("Chat History 📜")