Update app.py

app.py

@@ -77,7 +77,7 @@ for d in ["chat_logs", "vote_logs", "audio_logs", "history_logs", "audio_cache"]
 
 CHAT_DIR = "chat_logs"
 VOTE_DIR = "vote_logs"
-MEDIA_DIR = "."  # Root directory
+MEDIA_DIR = "."
 AUDIO_CACHE_DIR = "audio_cache"
 AUDIO_DIR = "audio_logs"
 STATE_FILE = "user_state.txt"
@@ -101,11 +101,12 @@ def format_timestamp_prefix(username=""):
 
 # 📈 Performance Timer
 class PerformanceTimer:
-    def __init__(self, name): self.name, self.start = name, None
-    def __enter__(self): 
+    def __init__(self, name):
+        self.name, self.start = name, None
+    def __enter__(self):
         self.start = time.time()
         return self
-    def __exit__(self, *args): 
+    def __exit__(self, *args):
         duration = time.time() - self.start
         st.session_state['operation_timings'][self.name] = duration
         st.session_state['performance_metrics'][self.name].append(duration)
@@ -128,7 +129,8 @@ def init_session_state():
         'quote_source': "famous"
     }
     for k, v in defaults.items():
-        if k not in st.session_state: st.session_state[k] = v
+        if k not in st.session_state:
+            st.session_state[k] = v
 
 # 🖌️ Marquee Helpers
 def update_marquee_settings_ui():
@@ -147,8 +149,12 @@ def display_marquee(text, settings, key_suffix=""):
     st.write("")
 
 # 📝 Text & File Helpers
-def clean_text_for_tts(text): return re.sub(r'[#*!\[\]]+', '', ' '.join(text.split()))[:200] or "No text"
-def clean_text_for_filename(text): return '_'.join(re.sub(r'[^\w\s-]', '', text.lower()).split())[:200]
+def clean_text_for_tts(text):
+    return re.sub(r'[#*!\[\]]+', '', ' '.join(text.split()))[:200] or "No text"
+
+def clean_text_for_filename(text):
+    return '_'.join(re.sub(r'[^\w\s-]', '', text.lower()).split())[:200]
+
 def get_high_info_terms(text, top_n=10):
     stop_words = {'the', 'a', 'an', 'and', 'or', 'but', 'in', 'on', 'at', 'to', 'for', 'of', 'with'}
     words = re.findall(r'\b\w+(?:-\w+)*\b', text.lower())
@@ -163,7 +169,8 @@ def generate_filename(prompt, username, file_type="md"):
 
 def create_file(prompt, username, file_type="md"):
     filename = generate_filename(prompt, username, file_type)
-    with open(filename, 'w', encoding='utf-8') as f: f.write(prompt)
+    with open(filename, 'w', encoding='utf-8') as f:
+        f.write(prompt)
     return filename
 
 def get_download_link(file, file_type="mp3"):
@@ -202,10 +209,12 @@ def concatenate_markdown_files():
 # 🎶 Audio Processing
 async def async_edge_tts_generate(text, voice, username, rate=0, pitch=0, file_format="mp3"):
     cache_key = f"{text[:100]}_{voice}_{rate}_{pitch}_{file_format}"
-    if cache_key in st.session_state['audio_cache']: return st.session_state['audio_cache'][cache_key], 0
+    if cache_key in st.session_state['audio_cache']:
+        return st.session_state['audio_cache'][cache_key], 0
     start_time = time.time()
     text = clean_text_for_tts(text)
-    if not text: return None, 0
+    if not text:
+        return None, 0
     filename = f"{format_timestamp_prefix(username)}-{hashlib.md5(text.encode()).hexdigest()[:8]}.{file_format}"
     communicate = edge_tts.Communicate(text, voice, rate=f"{rate:+d}%", pitch=f"{pitch:+d}Hz")
     await communicate.save(filename)
@@ -228,11 +237,13 @@ async def save_chat_entry(username, message, is_markdown=False):
     central = pytz.timezone('US/Central')
     timestamp = datetime.now(central).strftime("%Y-%m-%d %H:%M:%S")
     entry = f"[{timestamp}] {username}: {message}" if not is_markdown else f"[{timestamp}] {username}:\n```markdown\n{message}\n```"
-    with open(CHAT_FILE, 'a') as f: f.write(f"{entry}\n")
+    with open(CHAT_FILE, 'a') as f:
+        f.write(f"{entry}\n")
     voice = FUN_USERNAMES.get(username, "en-US-AriaNeural")
     audio_file, _ = await async_edge_tts_generate(message, voice, username)
     if audio_file:
-        with open(HISTORY_FILE, 'a') as f: f.write(f"[{timestamp}] {username}: Audio - {audio_file}\n")
+        with open(HISTORY_FILE, 'a') as f:
+            f.write(f"[{timestamp}] {username}: Audio - {audio_file}\n")
         st.session_state['mp3_files'][os.path.basename(audio_file)] = audio_file
     await broadcast_message(f"{username}|{message}", "chat")
     st.session_state.last_chat_update = time.time()
@@ -241,8 +252,9 @@ async def save_chat_entry(username, message, is_markdown=False):
 
 async def load_chat():
     if not os.path.exists(CHAT_FILE):
-        with open(CHAT_FILE, 'a') as f: f.write(f"# {START_ROOM} Chat\n\nWelcome to the cosmic hub! 🎤\n")
-    with open(CHAT_FILE, 'r') as f: 
+        with open(CHAT_FILE, 'a') as f:
+            f.write(f"# {START_ROOM} Chat\n\nWelcome to the cosmic hub! 🎤\n")
+    with open(CHAT_FILE, 'r') as f:
         content = f.read().strip()
     lines = content.split('\n')
     numbered_content = "\n".join(f"{i+1}. {line}" for i, line in enumerate(lines) if line.strip())
@@ -298,7 +310,8 @@ class AudioProcessor:
         self.metadata = json.load(open(f"{self.cache_dir}/metadata.json")) if os.path.exists(f"{self.cache_dir}/metadata.json") else {}
 
     def _save_metadata(self):
-        with open(f"{self.cache_dir}/metadata.json", 'w') as f: json.dump(self.metadata, f)
+        with open(f"{self.cache_dir}/metadata.json", 'w') as f:
+            json.dump(self.metadata, f)
 
     async def create_audio(self, text, voice='en-US-AriaNeural'):
         cache_key = hashlib.md5(f"{text}:{voice}".encode()).hexdigest()
@@ -306,7 +319,8 @@ class AudioProcessor:
         if cache_key in self.metadata and os.path.exists(cache_path):
             return cache_path
         text = clean_text_for_tts(text)
-        if not text: return None
+        if not text:
+            return None
         communicate = edge_tts.Communicate(text, voice)
         await communicate.save(cache_path)
         self.metadata[cache_key] = {'timestamp': datetime.now().isoformat(), 'text_length': len(text), 'voice': voice}
@@ -317,7 +331,7 @@ def process_pdf(pdf_file, max_pages, voice, audio_processor):
     reader = PdfReader(pdf_file)
     total_pages = min(len(reader.pages), max_pages)
     texts, audios = [], {}
-    async def process_page(i, text): 
+    async def process_page(i, text):
         audio_path = await audio_processor.create_audio(text, voice)
         if audio_path:
             audios[i] = audio_path
@@ -329,19 +343,24 @@ def process_pdf(pdf_file, max_pages, voice, audio_processor):
 
 # 🔍 ArXiv & AI Lookup
 def parse_arxiv_refs(ref_text):
-    if not ref_text: return []
+    if not ref_text:
+        return []
     papers = []
     current = {}
     for line in ref_text.split('\n'):
         if line.count('|') == 2:
-            if current: papers.append(current)
+            if current:
+                papers.append(current)
             date, title, *_ = line.strip('* ').split('|')
             url = re.search(r'(https://arxiv.org/\S+)', line).group(1) if re.search(r'(https://arxiv.org/\S+)', line) else f"paper_{len(papers)}"
             current = {'date': date, 'title': title, 'url': url, 'authors': '', 'summary': '', 'full_audio': None, 'download_base64': ''}
         elif current:
-            if not current['authors']: current['authors'] = line.strip('* ')
-            else: current['summary'] += ' ' + line.strip() if current['summary'] else line.strip()
-    if current: papers.append(current)
+            if not current['authors']:
+                current['authors'] = line.strip('* ')
+            else:
+                current['summary'] += ' ' + line.strip() if current['summary'] else line.strip()
+    if current:
+        papers.append(current)
     return papers[:20]
 
 def generate_5min_feature_markdown(paper):
@@ -363,13 +382,15 @@ def generate_5min_feature_markdown(paper):
 ---
 """
 
-def create_detailed_paper_md(papers): return "# Detailed Summary\n" + "\n".join(generate_5min_feature_markdown(p) for p in papers)
+def create_detailed_paper_md(papers):
+    return "# Detailed Summary\n" + "\n".join(generate_5min_feature_markdown(p) for p in papers)
 
 async def create_paper_audio_files(papers, query):
     for p in papers:
         audio_text = clean_text_for_tts(f"{p['title']} by {p['authors']}. {p['summary']}")
         p['full_audio'], _ = await async_edge_tts_generate(audio_text, st.session_state['tts_voice'], p['authors'])
-        if p['full_audio']: p['download_base64'] = get_download_link(p['full_audio'])
+        if p['full_audio']:
+            p['download_base64'] = get_download_link(p['full_audio'])
 
 async def perform_ai_lookup(q, useArxiv=True, useArxivAudio=False):
     client = anthropic.Anthropic(api_key=anthropic_key)
@@ -388,7 +409,8 @@ async def perform_ai_lookup(q, useArxiv=True, useArxivAudio=False):
         md_file, audio_file = create_file(result, "System", "md"), (await async_edge_tts_generate(result, st.session_state['tts_voice'], "System"))[0]
         play_and_download_audio(audio_file)
         papers = parse_arxiv_refs(refs)
-        if papers and useArxivAudio: await create_paper_audio_files(papers, q)
+        if papers and useArxivAudio:
+            await create_paper_audio_files(papers, q)
         return result, papers
     return result, []
 
@@ -435,7 +457,7 @@ async def save_pasted_image(image, username):
         return None
     timestamp = format_timestamp_prefix(username)
     filename = f"{timestamp}-{img_hash}.png"
-    filepath = filename  # Save at root
+    filepath = filename
     image.save(filepath, "PNG")
     st.session_state.image_hashes.add(img_hash)
     return filepath
@@ -443,10 +465,12 @@ async def save_pasted_image(image, username):
 # 📦 Zip Files
 def create_zip_of_files(md_files, mp3_files, png_files, mp4_files, query):
     all_files = md_files + mp3_files + png_files + mp4_files
-    if not all_files: return None
+    if not all_files:
+        return None
     terms = get_high_info_terms(" ".join([open(f, 'r', encoding='utf-8').read() if f.endswith('.md') else os.path.splitext(os.path.basename(f))[0].replace('_', ' ') for f in all_files] + [query]), 5)
     zip_name = f"{format_timestamp_prefix()}_{'-'.join(terms)[:20]}.zip"
-    with zipfile.ZipFile(zip_name, 'w') as z: [z.write(f) for f in all_files]
+    with zipfile.ZipFile(zip_name, 'w') as z:
+        [z.write(f) for f in all_files]
     return zip_name
 
 # 🎮 Main Interface
@@ -498,7 +522,7 @@ async def async_interface():
                             st.session_state.timer_start = time.time()
                             save_username(st.session_state.username)
                             st.rerun()
-        
+
         message = st.text_input(f"Message as {st.session_state.username}", key="message_input")
         paste_result = paste_image_button("📋 Paste Image or Text", key="paste_button_msg")
         if paste_result.image_data is not None:
@@ -524,22 +548,28 @@ async def async_interface():
 
         st.subheader("🎤 Speech-to-Chat")
         from mycomponent import speech_component
-        transcript_data = speech_component(default_value=st.session_state.get('last_transcript', ''))
+        transcript_data = speech_component(default_value=st.session_state.get('last_transcript', ''), key="speech_input")
         if transcript_data and 'value' in transcript_data:
             transcript = transcript_data['value'].strip()
-            st.write(f"🎙️ You said: {transcript}")
-            if transcript and transcript != st.session_state.last_transcript:
-                st.session_state.last_transcript = transcript
-                if st.session_state.autosend:
-                    await save_chat_entry(st.session_state.username, transcript, True)
-                    st.session_state.timer_start = time.time()
-                    save_username(st.session_state.username)
-                    st.rerun()
-                elif st.button("Send to Chat"):
-                    await save_chat_entry(st.session_state.username, transcript, True)
-                    st.session_state.timer_start = time.time()
-                    save_username(st.session_state.username)
-                    st.rerun()
+            if transcript:
+                st.write(f"🎙️ You said: {transcript}")
+                if transcript != st.session_state.last_transcript:
+                    st.session_state.last_transcript = transcript
+                    if st.session_state.autosend:
+                        await save_chat_entry(st.session_state.username, transcript, True)
+                        st.session_state.timer_start = time.time()
+                        save_username(st.session_state.username)
+                        # Update chat display without full rerun to reduce flicker
+                        with chat_container:
+                            st.markdown(await load_chat())
+                    else:
+                        if st.button("Send to Chat", key="send_transcript"):
+                            await save_chat_entry(st.session_state.username, transcript, True)
+                            st.session_state.timer_start = time.time()
+                            save_username(st.session_state.username)
+                            st.rerun()
+        else:
+            st.write("🎙️ Speak to transcribe your message...")
 
     # 📸 Media
     elif tab_main == "📸 Media":
@@ -550,12 +580,10 @@ async def async_interface():
         png_files = [f for f in all_files if f.endswith('.png')]
         mp4_files = [f for f in all_files if f.endswith('.mp4')]
 
-        # Display concatenated Markdown
         st.subheader("All Submitted Text")
         all_md_content = concatenate_markdown_files()
         st.markdown(all_md_content)
 
-        # Display Media by Type at End
         st.subheader("🎵 Audio (MP3)")
         for mp3 in mp3_files:
             with st.expander(os.path.basename(mp3)):
@@ -577,7 +605,7 @@ async def async_interface():
         uploaded_file = st.file_uploader("Upload Media", type=['png', 'mp4', 'mp3'])
         if uploaded_file:
             filename = f"{format_timestamp_prefix(st.session_state.username)}-{hashlib.md5(uploaded_file.getbuffer()).hexdigest()[:8]}.{uploaded_file.name.split('.')[-1]}"
-            with open(filename, 'wb') as f:  # Save at root
+            with open(filename, 'wb') as f:
                 f.write(uploaded_file.getbuffer())
             await save_chat_entry(st.session_state.username, f"Uploaded: {filename}")
             st.session_state.timer_start = time.time()
@@ -595,7 +623,8 @@ async def async_interface():
                     with st.expander(f"{i}. 📄 {p['title']}"):
                         st.markdown(f"**{p['date']} | {p['title']}** — [Link]({p['url']})")
                         st.markdown(generate_5min_feature_markdown(p))
-                        if p.get('full_audio'): play_and_download_audio(p['full_audio'])
+                        if p.get('full_audio'):
+                            play_and_download_audio(p['full_audio'])
 
     # 📚 PDF to Audio
     elif tab_main == "📚 PDF to Audio":
@@ -608,7 +637,8 @@ async def async_interface():
                 for i, text in enumerate(texts):
                     with st.expander(f"Page {i+1}"):
                         st.markdown(text)
-                        while i not in audios: time.sleep(0.1)
+                        while i not in audios:
+                            time.sleep(0.1)
                         if audios.get(i):
                             st.audio(audios[i])
                             st.markdown(get_download_link(audios[i], "mp3"), unsafe_allow_html=True)
@@ -662,7 +692,8 @@ async def async_interface():
         st.sidebar.write(f"{FILE_EMOJIS.get(f.split('.')[-1], '📄')} {os.path.basename(f)}")
     if st.sidebar.button("⬇️ Zip All"):
         zip_name = create_zip_of_files(md_files, mp3_files, png_files, mp4_files, "latest_query")
-        if zip_name: st.sidebar.markdown(get_download_link(zip_name, "zip"), unsafe_allow_html=True)
+        if zip_name:
+            st.sidebar.markdown(get_download_link(zip_name, "zip"), unsafe_allow_html=True)
 
 def main():
     asyncio.run(async_interface())