Update app.py

app.py

@@ -20,6 +20,8 @@ from streamlit_paste_button import paste_image_button
 import pytz
 import shutil
 from urllib.parse import urlencode
+from PyPDF2 import PdfReader
+import json
 
 # Patch for nested async - sneaky fix! 🐍✨
 nest_asyncio.apply()
@@ -87,7 +89,7 @@ UNICODE_FONTS = [
     ("Sans Serif Bold Italic", lambda x: "".join(chr(ord(c) + 0x1D63C - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D656 - 0x61) if 'a' <= c <= 'z' else c for c in x)),
     ("Monospace", lambda x: "".join(chr(ord(c) + 0x1D670 - 0x41) if 'A' <= c <= 'Z' else chr(ord(c) + 0x1D68A - 0x61) if 'a' <= c <= 'z' else c for c in x)),
     ("Circled", lambda x: "".join(chr(ord(c) - 0x41 + 0x24B6) if 'A' <= c <= 'Z' else chr(ord(c) - 0x61 + 0x24D0) if 'a' <= c <= 'z' else c for c in x)),
-    ("Squared", lambda x: "".join(chr(ord(c) - 0x41 + 0x1F130) if 'A' <= c <= 'Z' else c for c in x)),
+    ("Squared", lambda x: "".join(chr(ord(c) - 0x1F130 - 0x41) if 'A' <= c <= 'Z' else c for c in x)),
     ("Negative Circled", lambda x: "".join(chr(ord(c) - 0x41 + 0x1F150) if 'A' <= c <= 'Z' else c for c in x)),
     ("Negative Squared", lambda x: "".join(chr(ord(c) - 0x41 + 0x1F170) if 'A' <= c <= 'Z' else c for c in x)),
     ("Regional Indicator", lambda x: "".join(chr(ord(c) - 0x41 + 0x1F1E6) if 'A' <= c <= 'Z' else c for c in x)),
@@ -106,10 +108,6 @@ if 'last_chat_update' not in st.session_state:
     st.session_state.last_chat_update = 0
 if 'displayed_chat_lines' not in st.session_state:
     st.session_state.displayed_chat_lines = []
-if 'old_val' not in st.session_state:
-    st.session_state.old_val = ""
-if 'last_query' not in st.session_state:
-    st.session_state.last_query = ""
 if 'message_text' not in st.session_state:
     st.session_state.message_text = ""
 if 'audio_cache' not in st.session_state:
@@ -125,7 +123,7 @@ if 'base64_cache' not in st.session_state:
 if 'image_hashes' not in st.session_state:
     st.session_state.image_hashes = set()
 if 'gallery_columns' not in st.session_state:
-    st.session_state.gallery_columns = 1  # Default gallery tiles
+    st.session_state.gallery_columns = 1
 if 'user_id' not in st.session_state:
     st.session_state.user_id = None
 if 'user_hash' not in st.session_state:
@@ -177,6 +175,48 @@ def clean_text_for_tts(text):
     cleaned = ' '.join(cleaned.split())
     return cleaned[:200] if cleaned else "No text to speak"
 
+# Audio Processor Class from your code
+class AudioProcessor:
+    def __init__(self):
+        self.cache_dir = "audio_cache"
+        os.makedirs(self.cache_dir, exist_ok=True)
+        self.metadata = self._load_metadata()
+
+    def _load_metadata(self):
+        metadata_file = os.path.join(self.cache_dir, "metadata.json")
+        return json.load(open(metadata_file)) if os.path.exists(metadata_file) else {}
+
+    def _save_metadata(self):
+        metadata_file = os.path.join(self.cache_dir, "metadata.json")
+        with open(metadata_file, 'w') as f:
+            json.dump(self.metadata, f)
+
+    async def create_audio(self, text, voice='en-US-AriaNeural', filename=None):
+        cache_key = hashlib.md5(f"{text}:{voice}".encode()).hexdigest()
+        cache_path = filename or os.path.join(self.cache_dir, f"{cache_key}.mp3")
+
+        if cache_key in self.metadata and os.path.exists(cache_path):
+            return open(cache_path, 'rb').read()
+
+        # Clean text for speech
+        text = text.replace("\n", " ").replace("</s>", " ").strip()
+        if not text:
+            return None
+
+        # Generate audio with edge_tts
+        communicate = edge_tts.Communicate(text, voice)
+        await communicate.save(cache_path)
+
+        # Update metadata
+        self.metadata[cache_key] = {
+            'timestamp': datetime.now().isoformat(),
+            'text_length': len(text),
+            'voice': voice
+        }
+        self._save_metadata()
+
+        return open(cache_path, 'rb').read()
+
 # Chat saver - words locked tight! 💬🔒
 async def save_chat_entry(username, message, is_markdown=False, quote_line=None):
     await asyncio.to_thread(log_action, username, "💬🔒 - Chat saver - words locked tight!")
@@ -203,23 +243,28 @@ async def save_chat_entry(username, message, is_markdown=False, quote_line=None)
         f.write(f"{entry}\n")
 
     cleaned_message = clean_text_for_tts(message)
-    audio_file = await async_edge_tts_generate(cleaned_message, voice)
-    if audio_file:
+    audio_processor = AudioProcessor()
+    audio_filename = f"{format_timestamp_prefix(username)}-{hashlib.md5(cleaned_message.encode()).hexdigest()[:8]}.mp3"
+    audio_data = await audio_processor.create_audio(cleaned_message, voice, audio_filename)
+    if audio_data:
         with open(HISTORY_FILE, 'a') as f:
-            f.write(f"[{timestamp}] {username} ({voice}): Audio generated - {audio_file}\n")
+            f.write(f"[{timestamp}] {username} ({voice}): Audio generated - {audio_filename}\n")
         with open(user_history_file, 'a') as f:
-            f.write(f"{indent}[{timestamp}] Audio: {audio_file}\n")
+            f.write(f"{indent}[{timestamp}] Audio: {audio_filename}\n")
         
-        # Embed audio in global chat if it's an image or media reference
-        if message.startswith("Pasted image:") or message.startswith("Uploaded media:"):
+        # Embed audio and image in global chat if applicable
+        if message.startswith("Pasted image:") or message.startswith("Uploaded media:") or message.startswith("Uploaded PDF:"):
             with open(CHAT_FILE, 'a') as f:
-                f.write(f"{indent}[{timestamp}] Audio: {audio_file}\n")
+                f.write(f"{indent}[{timestamp}] Audio: {audio_filename}\n")
+                file_path = message.split(": ")[1]
+                if os.path.exists(file_path) and file_path.endswith(('.png', '.jpg', '.mp4', '.pdf')):
+                    f.write(f"{indent}[{timestamp}] Media: ![Media]({file_path})\n")
 
     await broadcast_message(f"{username}|{message}", "chat")
     st.session_state.last_chat_update = time.time()
-    return audio_file
+    return audio_filename
 
-# Save chat history with image
+# Save chat history with image or PDF
 async def save_chat_history_with_image(username, image_path):
     central = pytz.timezone('US/Central')
     timestamp = datetime.now(central).strftime("%Y-%m-%d_%H-%M-%S")
@@ -230,7 +275,7 @@ async def save_chat_history_with_image(username, image_path):
         with open(user_history_file, 'w') as f:
             f.write(f"# Chat History for {username} (Voice: {voice})\n\n")
     with open(user_history_file, 'a') as f:
-        f.write(f"[{timestamp}] {username} (Voice: {voice}) Shared Image: {os.path.basename(image_path)}\n")
+        f.write(f"[{timestamp}] {username} (Voice: {voice}) Shared Media: {os.path.basename(image_path)}\n")
         f.write(f"```markdown\n{chat_content}\n```\n")
 
 # Chat loader - history unleashed! 📜🚀
@@ -315,7 +360,7 @@ async def async_edge_tts_generate(text, voice, rate=0, pitch=0, file_format="mp3
     username = st.session_state.get('username', 'System 🌟')
     await asyncio.to_thread(log_action, username, "🎶🌟 - Audio maker - voices come alive!")
     timestamp = format_timestamp_prefix(username)
-    filename = f"{timestamp}.{file_format}"
+    filename = f"{timestamp}-{hashlib.md5(text.encode()).hexdigest()[:8]}.{file_format}"
     filepath = filename  # Top-level file
     communicate = edge_tts.Communicate(text, voice, rate=f"{rate:+d}%", pitch=f"{pitch:+d}Hz")
     try:
@@ -353,7 +398,34 @@ async def save_pasted_image(image, username):
     await save_chat_history_with_image(username, filepath)
     return filepath
 
-# Video renderer verbatim from your last version - movies roll with autoplay! 🎥🎬
+# PDF saver and audio generator
+async def save_pdf_and_generate_audio(pdf_file, username, max_pages=10):
+    await asyncio.to_thread(log_action, username, "📜🎶 - PDF saver and audio generator!")
+    timestamp = format_timestamp_prefix(username)
+    file_hash = hashlib.md5(pdf_file.getbuffer()).hexdigest()[:8]
+    pdf_filename = f"{timestamp}-{file_hash}.pdf"
+    with open(pdf_filename, 'wb') as f:
+        f.write(pdf_file.getbuffer())
+
+    reader = PdfReader(pdf_filename)
+    total_pages = min(len(reader.pages), max_pages)
+    texts = []
+    audio_files = []
+
+    audio_processor = AudioProcessor()
+    voice = st.session_state.voice if username == st.session_state.username else FUN_USERNAMES.get(username, "en-US-AriaNeural")
+
+    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_data = await audio_processor.create_audio(text, voice, audio_filename)
+        if audio_data:
+            audio_files.append(audio_filename)
+
+    return pdf_filename, texts, audio_files
+
+# Video renderer - movies roll with autoplay! 🎥🎬
 def get_video_html(video_path, width="100%"):
     video_url = f"data:video/mp4;base64,{base64.b64encode(open(video_path, 'rb').read()).decode()}"
     return f'''
@@ -443,13 +515,16 @@ def delete_user_files():
     st.session_state.displayed_chat_lines.clear()
     return deleted_files
 
-# Query parameter checker
+# Query parameter checker - parse q for username
 def check_query_params():
     query_params = st.query_params if hasattr(st, 'query_params') else st.experimental_get_query_params()
-    user_id = query_params.get("q", [None])[0]
-    if user_id:
-        st.session_state.user_id = user_id
-    return user_id
+    q_value = query_params.get("q", [None])[0]
+    if q_value and q_value in FUN_USERNAMES:
+        st.session_state.username = q_value
+        st.session_state.voice = FUN_USERNAMES[q_value]
+    elif q_value:
+        st.session_state.user_id = q_value  # Use as user_id if not a valid username
+    return q_value
 
 # Main execution - let’s roll! 🎲🚀
 def main():
@@ -464,15 +539,21 @@ def main():
             st.session_state.user_id = str(uuid.uuid4())
             st.session_state.user_hash = await generate_user_hash()
         if '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()))
+            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()))
             st.session_state.voice = FUN_USERNAMES[st.session_state.username]
             st.markdown(f"**🎙️ Voice Selected**: {st.session_state.voice} 🗣️ for {st.session_state.username}")
 
-        # Check query params and set user ID
-        check_query_params()
-        user_url = f"/q={st.session_state.user_id}"
+        # Check existing history file for content
+        user_history_file = f"{st.session_state.username}_history.md"
+        if os.path.exists(user_history_file):
+            with open(user_history_file, 'r') as f:
+                st.session_state.displayed_chat_lines = f.read().split('\n')
+
+        user_url = f"/q={st.session_state.username}"
         st.write(f"Your unique URL path: [{user_url}]({user_url})")
 
         st.title(f"🤖🧠MMO {st.session_state.username}📝🔬")
@@ -484,7 +565,6 @@ def main():
         # Unified Chat History at Top
         st.subheader(f"{START_ROOM} Chat History 💬")
         chat_content = await load_chat()
-        chat_lines = chat_content.split('\n')
         chat_lines = [line for line in chat_content.split('\n') if line.strip() and not line.startswith('#')]
         if chat_lines:
             col1, col2 = st.columns([2, 1])
@@ -501,6 +581,15 @@ def main():
                                 st.markdown(line)
                         else:
                             st.markdown(line)
+                        if "Media:" in line:
+                            file_path = line.split("Media: ")[-1].strip('![]()')
+                            if os.path.exists(file_path):
+                                if file_path.endswith(('.png', '.jpg')):
+                                    st.image(file_path, use_container_width=True)
+                                elif file_path.endswith('.mp4'):
+                                    st.markdown(get_video_html(file_path), unsafe_allow_html=True)
+                                elif file_path.endswith('.pdf'):
+                                    st.write(f"PDF: {os.path.basename(file_path)}")
                     with col_audio:
                         username = line.split(': ')[1].split(' ')[0] if ': ' in line else "Unknown"
                         cache_key = f"{line}_{FUN_USERNAMES.get(username, 'en-US-AriaNeural')}"
@@ -530,7 +619,8 @@ def main():
                     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 �<|control678|>", key="send_quote"):
+                        await save_chat_entry(st.session_state.username, f"Pasted image: {filename}", quote_line=st.session_state.quote_line)
+            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})"
@@ -576,14 +666,15 @@ def main():
                 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
+                    await save_chat_entry(st.session_state.username, f"Pasted image: {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'])
+        st.subheader("Upload Media 🎨🎶📜🎥")
+        uploaded_file = st.file_uploader("Upload Media", type=['png', 'jpg', 'mp4', 'mp3', 'wav', 'pdf'])
         if uploaded_file:
             timestamp = format_timestamp_prefix(st.session_state.username)
             username = st.session_state.username
@@ -594,7 +685,14 @@ def main():
                 file_path = filename  # Top-level file
                 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 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}")
+                    for i, (text, audio_file) in enumerate(zip(texts, audio_files)):
+                        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}")
                 await save_chat_history_with_image(username, file_path)
                 st.session_state.image_hashes.add(file_hash)
                 if file_path.endswith('.mp4'):
@@ -622,10 +720,10 @@ def main():
         st.rerun()
 
         # Gallery with Adjustable Tiles
-        st.subheader("Media Gallery 🎨🎶🎥")
+        st.subheader("Media Gallery 🎨🎶📜🎥")
         gallery_columns = st.slider("Number of Gallery Tiles", 1, 20, st.session_state.gallery_columns)
         st.session_state.gallery_columns = gallery_columns
-        media_files = glob.glob("*.png") + glob.glob("*.jpg") + glob.glob("*.mp4")
+        media_files = glob.glob("*.png") + glob.glob("*.jpg") + glob.glob("*.mp4") + glob.glob("*.pdf")
         if media_files:
             media_votes = await load_votes(MEDIA_VOTES_FILE)
             seen_files = set()
@@ -642,6 +740,8 @@ def main():
                             st.image(media_file, use_container_width=True)
                         elif media_file.endswith('.mp4'):
                             st.markdown(get_video_html(media_file), unsafe_allow_html=True)
+                        elif media_file.endswith('.pdf'):
+                            st.write(f"PDF: {filename}")
                         if st.button(f"👍 {vote_count}", key=f"media_vote_{media_file}"):
                             await save_vote(MEDIA_VOTES_FILE, media_file, await generate_user_hash(), st.session_state.username)
                             st.rerun()