Update app.py

app.py

@@ -50,12 +50,11 @@ EDGE_TTS_VOICES = [
 
 # Initialize session state variables
 if 'marquee_settings' not in st.session_state:
-    # Default to 20s animationDuration instead of 10s:
     st.session_state['marquee_settings'] = {
         "background": "#1E1E1E",
         "color": "#FFFFFF",
         "font-size": "14px",
-        "animationDuration": "20s",  # <- changed to 20s
+        "animationDuration": "20s",
         "width": "100%",
         "lineHeight": "35px"
     }
@@ -91,7 +90,6 @@ if 'last_query' not in st.session_state:
 if 'marquee_content' not in st.session_state:
     st.session_state['marquee_content'] = "🚀 Welcome to TalkingAIResearcher | 🤖 Your Research Assistant"
 
-# 🔑 2. API Setup & Clients
 openai_api_key = os.getenv('OPENAI_API_KEY', "")
 anthropic_key = os.getenv('ANTHROPIC_API_KEY_3', "")
 xai_key = os.getenv('xai',"")
@@ -105,7 +103,6 @@ openai_client = OpenAI(api_key=openai.api_key, organization=os.getenv('OPENAI_OR
 HF_KEY = os.getenv('HF_KEY')
 API_URL = os.getenv('API_URL')
 
-# Constants
 FILE_EMOJIS = {
     "md": "📝",
     "mp3": "🎵",
@@ -113,17 +110,14 @@ FILE_EMOJIS = {
 }
 
 def get_central_time():
-    """Get current time in US Central timezone"""
     central = pytz.timezone('US/Central')
     return datetime.now(central)
 
 def format_timestamp_prefix():
-    """Generate timestamp prefix in format MM_dd_yy_hh_mm_AM/PM"""
     ct = get_central_time()
     return ct.strftime("%m_%d_%y_%I_%M_%p")
 
 def initialize_marquee_settings():
-    """Initialize marquee settings in session state"""
     if 'marquee_settings' not in st.session_state:
         st.session_state['marquee_settings'] = {
             "background": "#1E1E1E",
@@ -135,13 +129,10 @@ def initialize_marquee_settings():
         }
 
 def get_marquee_settings():
-    """Get or update marquee settings from session state"""
     initialize_marquee_settings()
     return st.session_state['marquee_settings']
 
 def update_marquee_settings_ui():
-    """Update marquee settings via UI controls"""
-    initialize_marquee_settings()
     st.sidebar.markdown("### 🎯 Marquee Settings")
     cols = st.sidebar.columns(2)
     with cols[0]:
@@ -163,7 +154,6 @@ def update_marquee_settings_ui():
     })
 
 def display_marquee(text, settings, key_suffix=""):
-    """Display marquee with given text and settings"""
     truncated_text = text[:280] + "..." if len(text) > 280 else text
     streamlit_marquee(
         content=truncated_text,
@@ -173,9 +163,6 @@ def display_marquee(text, settings, key_suffix=""):
     st.write("")
 
 def get_high_info_terms(text: str, top_n=10) -> list:
-    """
-    Finds the top_n frequent words or bigrams (excluding some common stopwords).
-    """
     stop_words = set(['the', 'a', 'an', 'and', 'or', 'but', 'in', 'on', 'at', 'to', 'for', 'of', 'with'])
     words = re.findall(r'\b\w+(?:-\w+)*\b', text.lower())
     bi_grams = [' '.join(pair) for pair in zip(words, words[1:])]
@@ -185,9 +172,6 @@ def get_high_info_terms(text: str, top_n=10) -> list:
     return [term for term, freq in counter.most_common(top_n)]
 
 def clean_text_for_filename(text: str) -> str:
-    """
-    Cleans a text so it can be used in a filename.
-    """
     text = text.lower()
     text = re.sub(r'[^\w\s-]', '', text)
     words = text.split()
@@ -196,23 +180,13 @@ def clean_text_for_filename(text: str) -> str:
     filtered = [w for w in words if len(w) > 3 and w not in stop_short]
     return '_'.join(filtered)[:200]
 
-
 def generate_filename(prompt, response, file_type="md", max_length=200):
-    """
-    Generate a shortened filename by:
-      1. Extracting high-info terms
-      2. Creating a smaller snippet
-      3. Cleaning & joining them
-      4. Removing duplicates
-      5. Truncating if needed
-    """
     prefix = format_timestamp_prefix() + "_"
     combined_text = (prompt + " " + response)[:200]  
     info_terms = get_high_info_terms(combined_text, top_n=5)  
     snippet = (prompt[:40] + " " + response[:40]).strip()
     snippet_cleaned = clean_text_for_filename(snippet)
-    
-    # Combine info terms + snippet, remove duplicates
+    # remove duplicates
     name_parts = info_terms + [snippet_cleaned]
     seen = set()
     unique_parts = []
@@ -228,20 +202,13 @@ def generate_filename(prompt, response, file_type="md", max_length=200):
     
     return f"{prefix}{full_name}.{file_type}"
 
-
 def create_file(prompt, response, file_type="md"):
-    """
-    Create a file using the shortened filename from generate_filename().
-    """
     filename = generate_filename(prompt.strip(), response.strip(), file_type)
     with open(filename, 'w', encoding='utf-8') as f:
         f.write(prompt + "\n\n" + response)
     return filename
 
 def get_download_link(file, file_type="zip"):
-    """
-    Returns an HTML anchor tag for downloading the specified file (base64-encoded).
-    """
     with open(file, "rb") as f:
         b64 = base64.b64encode(f.read()).decode()
     if file_type == "zip":
@@ -256,9 +223,6 @@ def get_download_link(file, file_type="zip"):
         return f'<a href="data:application/octet-stream;base64,{b64}" download="{os.path.basename(file)}">Download {os.path.basename(file)}</a>'
 
 def clean_for_speech(text: str) -> str:
-    """
-    Cleans text to make TTS output more coherent.
-    """
     text = text.replace("\n", " ")
     text = text.replace("</s>", " ")
     text = text.replace("#", "")
@@ -281,36 +245,25 @@ def speak_with_edge_tts(text, voice="en-US-AriaNeural", rate=0, pitch=0, file_fo
     return asyncio.run(edge_tts_generate_audio(text, voice, rate, pitch, file_format))
 
 def play_and_download_audio(file_path, file_type="mp3"):
-    """Play audio and show a direct download link in the main area."""
     if file_path and os.path.exists(file_path):
         st.audio(file_path)
         dl_link = get_download_link(file_path, file_type=file_type)
         st.markdown(dl_link, unsafe_allow_html=True)
 
 def save_qa_with_audio(question, answer, voice=None):
-    """Save Q&A to markdown and generate audio file."""
     if not voice:
         voice = st.session_state['tts_voice']
-    
-    # Create markdown file
     combined_text = f"# Question\n{question}\n\n# Answer\n{answer}"
     md_file = create_file(question, answer, "md")
-    
-    # Generate audio file
     audio_text = f"{question}\n\nAnswer: {answer}"
     audio_file = speak_with_edge_tts(
         audio_text,
         voice=voice,
         file_format=st.session_state['audio_format']
     )
-    
     return md_file, audio_file
 
 def parse_arxiv_refs(ref_text: str):
-    """
-    Given a multi-line markdown with arxiv references, parse them into
-    a structure: [{date, title, url, authors, summary}, ...]
-    """
     if not ref_text:
         return []
 
@@ -320,19 +273,16 @@ def parse_arxiv_refs(ref_text: str):
     
     for i, line in enumerate(lines):
         if line.count('|') == 2:
-            # We found a new paper header line
             if current_paper:
                 results.append(current_paper)
                 if len(results) >= 20:
                     break
-            
             try:
                 header_parts = line.strip('* ').split('|')
                 date = header_parts[0].strip()
                 title = header_parts[1].strip()
                 url_match = re.search(r'(https://arxiv.org/\S+)', line)
                 url = url_match.group(1) if url_match else f"paper_{len(results)}"
-                
                 current_paper = {
                     'date': date,
                     'title': title,
@@ -346,9 +296,7 @@ def parse_arxiv_refs(ref_text: str):
                 st.warning(f"Error parsing paper header: {str(e)}")
                 current_paper = {}
                 continue
-        
         elif current_paper:
-            # Fill authors if empty, else fill summary
             if not current_paper['authors']:
                 current_paper['authors'] = line.strip('* ')
             else:
@@ -356,31 +304,20 @@ def parse_arxiv_refs(ref_text: str):
                     current_paper['summary'] += ' ' + line.strip()
                 else:
                     current_paper['summary'] = line.strip()
-    
     if current_paper:
         results.append(current_paper)
     
     return results[:20]
 
 def create_paper_links_md(papers):
-    """
-    Creates a minimal markdown list of paper titles + arxiv links
-    (and if you store PDF links, you could also include them).
-    """
     lines = ["# Paper Links\n"]
     for i, p in enumerate(papers, start=1):
-        # Basic link
         lines.append(f"{i}. **{p['title']}** — [Arxiv]({p['url']})")
     return "\n".join(lines)
 
 def create_paper_audio_files(papers, input_question):
-    """
-    Generate TTS audio for each paper, store base64 link for stable download, 
-    and attach to each paper dict.
-    """
     for paper in papers:
         try:
-            # Just a short version for TTS
             audio_text = f"{paper['title']} by {paper['authors']}. {paper['summary']}"
             audio_text = clean_for_speech(audio_text)
 
@@ -392,11 +329,9 @@ def create_paper_audio_files(papers, input_question):
             )
             paper['full_audio'] = audio_file
 
-            # Store a base64 link with consistent name
             if audio_file:
                 with open(audio_file, "rb") as af:
                     b64_data = base64.b64encode(af.read()).decode()
-                # We'll keep the original file's name as the stable download name
                 download_filename = os.path.basename(audio_file)
                 mime_type = "mpeg" if file_format == "mp3" else "wav"
                 paper['download_base64'] = (
@@ -410,13 +345,8 @@ def create_paper_audio_files(papers, input_question):
             paper['download_base64'] = ''
 
 def display_papers(papers, marquee_settings):
-    """
-    Display the papers in the main area with marquee + expanders + audio.
-    """
     st.write("## Research Papers")
-    
     for i, paper in enumerate(papers, start=1):
-        # Show marquee
         marquee_text = f"📄 {paper['title']} | 👤 {paper['authors'][:120]} | 📝 {paper['summary'][:200]}"
         display_marquee(marquee_text, marquee_settings, key_suffix=f"paper_{i}")
         
@@ -424,7 +354,6 @@ def display_papers(papers, marquee_settings):
             st.markdown(f"**{paper['date']} | {paper['title']} |** [Arxiv Link]({paper['url']})")
             st.markdown(f"*Authors:* {paper['authors']}")
             st.markdown(paper['summary'])
-            
             if paper.get('full_audio'):
                 st.write("📚 Paper Audio")
                 st.audio(paper['full_audio'])
@@ -432,10 +361,6 @@ def display_papers(papers, marquee_settings):
                     st.markdown(paper['download_base64'], unsafe_allow_html=True)
 
 def display_papers_in_sidebar(papers):
-    """
-    New approach: in the sidebar, mirror the paper listing
-    with expanders for each paper, link to arxiv, st.audio, etc.
-    """
     st.sidebar.title("🎶 Papers & Audio")
     for i, paper in enumerate(papers, start=1):
         with st.sidebar.expander(f"{i}. {paper['title']}"):
@@ -444,15 +369,11 @@ def display_papers_in_sidebar(papers):
                 st.audio(paper['full_audio'])
                 if paper['download_base64']:
                     st.markdown(paper['download_base64'], unsafe_allow_html=True)
-            # Show minimal text if desired:
             st.markdown(f"**Authors:** {paper['authors']}")
             if paper['summary']:
                 st.markdown(f"**Summary:** {paper['summary'][:300]}...")
 
 def create_zip_of_files(md_files, mp3_files, wav_files, input_question):
-    """
-    Zip up all relevant files, but limit final zip name to 20 chars.
-    """
     md_files = [f for f in md_files if os.path.basename(f).lower() != 'readme.md']
     all_files = md_files + mp3_files + wav_files
     if not all_files:
@@ -464,7 +385,6 @@ def create_zip_of_files(md_files, mp3_files, wav_files, input_question):
             with open(f, 'r', encoding='utf-8') as file:
                 all_content.append(file.read())
         elif f.endswith('.mp3') or f.endswith('.wav'):
-            # Add some text representation
             basename = os.path.splitext(os.path.basename(f))[0]
             words = basename.replace('_', ' ')
             all_content.append(words)
@@ -474,82 +394,61 @@ def create_zip_of_files(md_files, mp3_files, wav_files, input_question):
     info_terms = get_high_info_terms(combined_content, top_n=10)
     
     timestamp = format_timestamp_prefix()
-    name_text = '-'.join(term for term in info_terms[:5])  # shorter
-    # Limit the final name to 20 chars (excluding .zip)
+    name_text = '-'.join(term for term in info_terms[:5])  
     short_zip_name = (timestamp + "_" + name_text)[:20] + ".zip"
-    
+
     with zipfile.ZipFile(short_zip_name, 'w') as z:
         for f in all_files:
             z.write(f)
-    
     return short_zip_name
 
-
-# ---------------------------- 1/11/2025 - add a constitution to my arxiv system templating to build configurable personality
-
 def perform_ai_lookup(q, vocal_summary=True, extended_refs=False, 
                      titles_summary=True, full_audio=False):
     start = time.time()
-
     ai_constitution = """
-    You are a talented AI coder and songwriter with a unique ability to explain scientific concepts through music with code easter eggs.. Your task is to create a song that not only entertains but also educates listeners about a specific science problem and its potential solutions. 
-
-    (Omitted extra instructions for brevity...)
+    You are a talented AI coder and songwriter...
     """
-
     # Claude:
     client = anthropic.Anthropic(api_key=anthropic_key)
     user_input = q
-
     response = client.messages.create(
         model="claude-3-sonnet-20240229",
         max_tokens=1000,
         messages=[
             {"role": "user", "content": user_input}
         ])
-    
     st.write("Claude's reply 🧠:")
     st.markdown(response.content[0].text)
 
-    # Save and produce audio for Claude response
     result = response.content[0].text
-    create_file(q, result)  # MD file
+    create_file(q, result) 
     md_file, audio_file = save_qa_with_audio(q, result)
     st.subheader("📝 Main Response Audio")
     play_and_download_audio(audio_file, st.session_state['audio_format'])
 
-
     # Arxiv:
-    st.write("Arxiv's AI this Evening is Mixtral 8x7B MoE Instruct with 9 English Voices 🧠:")
-
+    st.write("Arxiv's AI this Evening...")
+    from gradio_client import Client
     client = Client("awacke1/Arxiv-Paper-Search-And-QA-RAG-Pattern")
     refs = client.predict(q, 20, "Semantic Search", 
                          "mistralai/Mixtral-8x7B-Instruct-v0.1",
                          api_name="/update_with_rag_md")[0]
-
     r2 = client.predict(q, "mistralai/Mixtral-8x7B-Instruct-v0.1", 
                        True, api_name="/ask_llm")
 
     result = f"### 🔎 {q}\n\n{r2}\n\n{refs}"
-    
-    # Save and produce audio for second response
     md_file, audio_file = save_qa_with_audio(q, result)
-    
     st.subheader("📝 Main Response Audio")
     play_and_download_audio(audio_file, st.session_state['audio_format'])
 
     papers = parse_arxiv_refs(refs)
     if papers:
-        # 4) Create & show a minimal markdown links page before generating audio
         paper_links = create_paper_links_md(papers)
         links_file = create_file(q, paper_links, "md")
         st.markdown(paper_links)
 
-        # Now produce audio for each paper
         create_paper_audio_files(papers, input_question=q)
         display_papers(papers, get_marquee_settings())
-        
-        # Also display in the sidebar as requested
         display_papers_in_sidebar(papers)
     else:
         st.warning("No papers found in the response.")
@@ -561,7 +460,6 @@ def perform_ai_lookup(q, vocal_summary=True, extended_refs=False,
 def process_voice_input(text):
     if not text:
         return
-        
     st.subheader("🔍 Search Results")
     result = perform_ai_lookup(
         text, 
@@ -570,38 +468,77 @@ def process_voice_input(text):
         titles_summary=True,
         full_audio=True
     )
-    
-    # Save final Q&A with audio
     md_file, audio_file = save_qa_with_audio(text, result)
-    
     st.subheader("📝 Generated Files")
     st.write(f"Markdown: {md_file}")
     st.write(f"Audio: {audio_file}")
     play_and_download_audio(audio_file, st.session_state['audio_format'])
 
+# ----------------------------------------------------------------------------
+# ADD HERE — FILE HISTORY SIDEBAR
+def display_file_history_in_sidebar():
+    """
+    Shows a history of each local .md, .mp3, .wav file in descending
+    order of modification time, with quick icons and optional download links.
+    """
+    st.sidebar.markdown("---")
+    st.sidebar.markdown("### 📂 File History")
+
+    # Gather all files of interest
+    md_files = glob.glob("*.md")
+    mp3_files = glob.glob("*.mp3")
+    wav_files = glob.glob("*.wav")
+    all_files = md_files + mp3_files + wav_files
+
+    if not all_files:
+        st.sidebar.write("No files found.")
+        return
+
+    # Sort by newest first
+    all_files = sorted(all_files, key=os.path.getmtime, reverse=True)
+
+    for f in all_files:
+        fname = os.path.basename(f)
+        ext = os.path.splitext(fname)[1].lower().strip('.')
+        emoji = FILE_EMOJIS.get(ext, '📦')
+        time_str = datetime.fromtimestamp(os.path.getmtime(f)).strftime("%Y-%m-%d %H:%M:%S")
+
+        with st.sidebar.expander(f"{emoji} {fname}"):
+            st.write(f"**Modified:** {time_str}")
+            # Optionally show a snippet for .md:
+            if ext == "md":
+                with open(f, "r", encoding="utf-8") as file_in:
+                    snippet = file_in.read(200).replace("\n", " ")
+                if len(snippet) == 200:
+                    snippet += "..."
+                st.write(snippet)
+                st.markdown(get_download_link(f, file_type="md"), unsafe_allow_html=True)
+            # If it's audio, let user play it
+            elif ext in ["mp3","wav"]:
+                st.audio(f)
+                st.markdown(get_download_link(f, file_type=ext), unsafe_allow_html=True)
+            else:
+                st.markdown(get_download_link(f), unsafe_allow_html=True)
+
+# ----------------------------------------------------------------------------
+
 def main():
-    # Update marquee settings UI
     update_marquee_settings_ui()
     marquee_settings = get_marquee_settings()
-
-    # Initial welcome marquee
     display_marquee(st.session_state['marquee_content'], 
                     {**marquee_settings, "font-size": "28px", "lineHeight": "50px"},
                     key_suffix="welcome")
 
-    # Main action tabs
+    # -- Insert your main app tabs, logic, etc. --
     tab_main = st.radio("Action:", ["🎤 Voice", "📸 Media", "🔍 ArXiv", "📝 Editor"], 
                         horizontal=True)
 
-    # Simple example usage of a Streamlit component (placeholder)
     mycomponent = components.declare_component("mycomponent", path="mycomponent")
     val = mycomponent(my_input_value="Hello")
 
-    # Quick example - if the component returns text:
     if val:
         val_stripped = val.replace('\\n', ' ')
         edited_input = st.text_area("✏️ Edit Input:", value=val_stripped, height=100)
-        
         run_option = st.selectbox("Model:", ["Arxiv"])
         col1, col2 = st.columns(2)
         with col1:
@@ -629,11 +566,9 @@ def main():
                                   titles_summary=True, 
                                   full_audio=full_audio)
     
-    # --- Tab: ArXiv
     if tab_main == "🔍 ArXiv":
         st.subheader("🔍 Query ArXiv")
         q = st.text_input("🔍 Query:", key="arxiv_query")
-    
         st.markdown("### 🎛 Options")
         vocal_summary = st.checkbox("🎙ShortAudio", value=True, key="option_vocal_summary")
         extended_refs = st.checkbox("📜LongRefs", value=False, key="option_extended_refs")
@@ -648,18 +583,14 @@ def main():
             if full_transcript:
                 create_file(q, result, "md")
 
-    # --- Tab: Voice  
     elif tab_main == "🎤 Voice":
         st.subheader("🎤 Voice Input")
-        
-        # Voice and format settings
         st.markdown("### 🎤 Voice Settings")
         selected_voice = st.selectbox(
             "Select TTS Voice:",
             options=EDGE_TTS_VOICES,
             index=EDGE_TTS_VOICES.index(st.session_state['tts_voice'])
         )
-        
         st.markdown("### 🔊 Audio Format")
         selected_format = st.radio(
             "Choose Audio Format:",
@@ -674,7 +605,6 @@ def main():
             st.session_state['audio_format'] = selected_format.lower()
             st.rerun()
 
-        # User text
         user_text = st.text_area("💬 Message:", height=100)
         user_text = user_text.strip().replace('\n', ' ')
 
@@ -686,11 +616,9 @@ def main():
             st.write("**You:**", c["user"])
             st.write("**Response:**", c["claude"])
 
-    # --- Tab: Media
     elif tab_main == "📸 Media":
         st.header("📸 Media Gallery")
-        tabs = st.tabs(["🎵 Audio", "🖼 Images", "🎥 Video"])  # audio first = default
-        # --- Audio Tab
+        tabs = st.tabs(["🎵 Audio", "🖼 Images", "🎥 Video"])  # audio first
         with tabs[0]:
             st.subheader("🎵 Audio Files")
             audio_files = glob.glob("*.mp3") + glob.glob("*.wav")
@@ -703,8 +631,6 @@ def main():
                         st.markdown(dl_link, unsafe_allow_html=True)
             else:
                 st.write("No audio files found.")
-                
-        # --- Images Tab
         with tabs[1]:
             st.subheader("🖼 Image Files")
             imgs = glob.glob("*.png") + glob.glob("*.jpg") + glob.glob("*.jpeg")
@@ -716,8 +642,6 @@ def main():
                         st.image(Image.open(f), use_container_width=True)
             else:
                 st.write("No images found.")
-
-        # --- Video Tab
         with tabs[2]:
             st.subheader("🎥 Video Files")
             vids = glob.glob("*.mp4") + glob.glob("*.mov") + glob.glob("*.avi")
@@ -728,9 +652,11 @@ def main():
             else:
                 st.write("No videos found.")
 
-    # --- Tab: Editor
     elif tab_main == "📝 Editor":
-        st.write("Select or create a file to edit. (Currently minimal demo)")
+        st.write("Select or create a file to edit. (Currently minimal)")
+
+    # --- IMPORTANT: Display the file-history in the sidebar
+    display_file_history_in_sidebar()
 
     st.markdown("""
     <style>
@@ -744,6 +670,5 @@ def main():
         st.session_state.should_rerun = False
         st.rerun()
 
-
 if __name__ == "__main__":
     main()