Update app.py

app.py

@@ -106,6 +106,16 @@ 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"
 
+# To track user checkboxes for Claude / Arxiv / Autorun / AutoSave
+if 'run_claude' not in st.session_state:
+    st.session_state['run_claude'] = True  # enabled by default
+if 'run_arxiv' not in st.session_state:
+    st.session_state['run_arxiv'] = False  # disabled by default
+if 'autorun_searches' not in st.session_state:
+    st.session_state['autorun_searches'] = False
+if 'autosave_output' not in st.session_state:
+    st.session_state['autosave_output'] = False
+
 # API Keys
 openai_api_key = os.getenv('OPENAI_API_KEY', "")
 anthropic_key = os.getenv('ANTHROPIC_API_KEY_3', "")
@@ -295,18 +305,23 @@ def play_and_download_audio(file_path, file_type="mp3"):
         st.markdown(dl_link, unsafe_allow_html=True)
 
 def save_qa_with_audio(question, answer, voice=None):
-    """Save Q&A to markdown and also generate audio."""
+    """Save Q&A to markdown and also generate audio, returning file paths."""
     if not voice:
         voice = st.session_state['tts_voice']
     
     combined_text = f"# Question\n{question}\n\n# Answer\n{answer}"
-    md_file = create_file(question, answer, "md")
-    audio_text = f"{question}\n\nAnswer: {answer}"
-    audio_file = speak_with_edge_tts(
-        audio_text,
-        voice=voice,
-        file_format=st.session_state['audio_format']
-    )
+    md_file = None
+    audio_file = None
+
+    # Only create the files if autosave is enabled
+    if st.session_state['autosave_output']:
+        md_file = create_file(question, answer, "md")
+        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
 
 # ─────────────────────────────────────────────────────────
@@ -377,9 +392,12 @@ def create_paper_links_md(papers):
 
 def create_paper_audio_files(papers, input_question):
     """
-    For each paper, generate TTS audio summary, store the path in `paper['full_audio']`,
-    and also store a base64 link for stable downloading.
+    For each paper, generate TTS audio summary if autosave is on,
+    store the path in `paper['full_audio']`,
+    and store a base64 link for stable downloading.
     """
+    if not st.session_state['autosave_output']:
+        return
     for paper in papers:
         try:
             audio_text = f"{paper['title']} by {paper['authors']}. {paper['summary']}"
@@ -479,97 +497,95 @@ def create_zip_of_files(md_files, mp3_files, wav_files, input_question):
 # 5. MAIN LOGIC: AI LOOKUP & VOICE INPUT
 # ─────────────────────────────────────────────────────────
 
-def perform_ai_lookup(q, vocal_summary=True, extended_refs=False, 
-                     titles_summary=True, full_audio=False):
-    """Main routine that uses Anthropic (Claude) + Gradio ArXiv RAG pipeline."""
-    start = time.time()
-    ai_constitution = """
-    You are a talented AI coder and songwriter...
-    """
-
-    # --- 1) Claude API
+def run_claude_search(q):
+    """Call Anthropic (Claude) for the user's query."""
     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}
+            {"role": "user", "content": q}
         ])
-    st.write("Claude's reply 🧠:")
-    st.markdown(response.content[0].text)
-
-    # Save & produce audio and text result
     result = response.content[0].text
-    create_file(q, result) 
-    md_file, audio_file = save_qa_with_audio(q, result)
 
+    st.write("### Claude's reply 🧠:")
+    st.markdown(result)
+    return result
+
+def run_arxiv_search(q):
+    """Call the Arxiv RAG pipeline for the user's query."""
+    st.write("### Arxiv's AI: Mixtral 8x7B RAG")
+    client = Client("awacke1/Arxiv-Paper-Search-And-QA-RAG-Pattern")
+    refs = client.predict(
+        q,
+        10,  # topK
+        "Semantic Search", 
+        "mistralai/Mixtral-8x7B-Instruct-v0.1",
+        api_name="/update_with_rag_md"
+    )[0]
     
-    st.subheader("📝 Claude Response Audio")
-    play_and_download_audio(audio_file, st.session_state['audio_format'])
-
-    if st.checkbox("Run Arxiv Research AI"):
-        # --- 2) Arxiv RAG
-        st.write("Arxiv's AI this Evening is Mixtral 8x7B...")
-        client = Client("awacke1/Arxiv-Paper-Search-And-QA-RAG-Pattern")
-        refs = client.predict(
-            q, 
-            #20, 
-            10, 
-            "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}"
-        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'])
+    r2 = client.predict(
+        q, 
+        "mistralai/Mixtral-8x7B-Instruct-v0.1", 
+        True, 
+        api_name="/ask_llm"
+    )
     
-        # --- 3) Parse + handle papers
-        papers = parse_arxiv_refs(refs)
-        if papers:
-            # Create minimal links page first
+    # Combine the final result
+    result = f"### 🔎 {q}\n\n{r2}\n\n{refs}"
+    st.markdown(result)
+
+    # Parse + handle papers
+    papers = parse_arxiv_refs(refs)
+    if papers:
+        # Create minimal links page if autosave is on
+        if st.session_state['autosave_output']:
             paper_links = create_paper_links_md(papers)
-            links_file = create_file(q, paper_links, "md")
-            st.markdown(paper_links)
+            create_file(q, paper_links, "md")
+
+        create_paper_audio_files(papers, input_question=q)
+        display_papers(papers, get_marquee_settings())
+        display_papers_in_sidebar(papers)
+    else:
+        st.warning("No papers found in the response.")
+    return result
+
+def perform_selections(q):
+    """
+    Checks which search(s) are enabled and runs them in sequence,
+    returning a combined response string.
+    """
+    combined_response = ""
     
-            # Then create audio for each paper
-            create_paper_audio_files(papers, input_question=q)
-            display_papers(papers, get_marquee_settings())
-            display_papers_in_sidebar(papers)
-        else:
-            st.warning("No papers found in the response.")
+    if st.session_state['run_claude']:
+        claude_response = run_claude_search(q)
+        combined_response += "\n\nCLAUDE:\n" + claude_response
     
-        elapsed = time.time() - start
-        st.write(f"**Total Elapsed:** {elapsed:.2f} s")
-        
-    return result
+    if st.session_state['run_arxiv']:
+        arxiv_response = run_arxiv_search(q)
+        combined_response += "\n\nARXIV:\n" + arxiv_response
+    
+    return combined_response
 
 def process_voice_input(text):
-    """When user sends voice query, we run the AI lookup + Q&A with audio."""
+    """When user sends voice query, we run whichever searches are enabled."""
     if not text:
         return
     st.subheader("🔍 Search Results")
-    result = perform_ai_lookup(
-        text, 
-        vocal_summary=True,
-        extended_refs=False,
-        titles_summary=True,
-        full_audio=True
-    )
+    # 1) Run the searches user has checked
+    result = perform_selections(text)
+
+    # 2) If autosave is turned on, store Q&A output
     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'])
+
+    # 3) If we saved TTS audio, play it
+    if audio_file:
+        st.subheader("📝 Generated Audio")
+        play_and_download_audio(audio_file, st.session_state['audio_format'])
+
+    # 4) Refresh the sidebar file listing if autosave was used
+    if st.session_state['autosave_output']:
+        display_file_history_in_sidebar()
 
 # ─────────────────────────────────────────────────────────
 # 6. FILE HISTORY SIDEBAR
@@ -588,7 +604,7 @@ def display_file_history_in_sidebar():
     with col1:
         if st.button("🗑️ Delete All"):
             # Delete all files except README.md
-            for pattern in ["*.md", "*.mp3", "*.mp4"]:
+            for pattern in ["*.md", "*.mp3", "*.wav", "*.mp4"]:
                 for f in glob.glob(pattern):
                     if f.lower() != "readme.md":
                         try:
@@ -600,16 +616,20 @@ def display_file_history_in_sidebar():
     with col2:
         # Get all files for potential zip
         md_files = [f for f in glob.glob("*.md") if f.lower() != "readme.md"]
-        mp4_files = glob.glob("*.mp4")
-        if md_files or mp4_files:
-            # Use last query if available, otherwise generic name
+        mp3_files = glob.glob("*.mp3")
+        wav_files = glob.glob("*.wav")
+
+        if md_files or mp3_files or wav_files:
             zip_name = "Download.zip"
             if 'last_query' in st.session_state and st.session_state['last_query']:
-                zip_name = f"{clean_text_for_filename(st.session_state['last_query'])[:30]}.zip"
-            
+                # Use last_query in the zip name
+                clean_q = clean_text_for_filename(st.session_state['last_query'])[:30]
+                if clean_q:
+                    zip_name = f"{clean_q}.zip"
+
             if st.button("📦 Download All"):
                 with zipfile.ZipFile(zip_name, 'w') as z:
-                    for f in md_files + mp4_files:
+                    for f in md_files + mp3_files + wav_files:
                         z.write(f)
                 st.sidebar.markdown(get_download_link(zip_name), unsafe_allow_html=True)
 
@@ -623,8 +643,23 @@ def display_file_history_in_sidebar():
         st.sidebar.write("No files found.")
         return
 
-
-
+    # Sort by modification time descending
+    all_files_sorted = sorted(all_files, key=os.path.getmtime, reverse=True)
+
+    for f in all_files_sorted:
+        ext = os.path.splitext(f)[1].lower().replace('.', '')
+        emoji = FILE_EMOJIS.get(ext, "📁")
+        mod_time = datetime.fromtimestamp(os.path.getmtime(f)).strftime("%m-%d %H:%M")
+        # Download link
+        dl_link = get_download_link(f, file_type=ext)
+        with st.sidebar.expander(f"{emoji} {os.path.basename(f)} [{mod_time}]"):
+            if ext in ["mp3", "wav"]:
+                st.audio(f)
+            if ext == "md":
+                with open(f, 'r', encoding='utf-8') as file:
+                    content = file.read()
+                st.markdown(f"```\n{content[:500]}\n...```")
+            st.markdown(dl_link, unsafe_allow_html=True)
 
 # ─────────────────────────────────────────────────────────
 # 7. MAIN APP
@@ -635,16 +670,25 @@ def main():
     update_marquee_settings_ui()
     marquee_settings = get_marquee_settings()
 
-    # 2) Display the marquee welcome
+    # 2) Place the radio for main tabs AND the row of checkboxes
+    colA, colB, colC, colD, colE = st.columns([2,1,1,1,1])
+    with colA:
+        tab_main = st.radio("Action:", ["🎤 Voice", "📸 Media", "🔍 ArXiv", "📝 Editor"], horizontal=True)
+    with colB:
+        st.session_state['run_claude'] = st.checkbox("Claude", value=st.session_state['run_claude'])
+    with colC:
+        st.session_state['run_arxiv'] = st.checkbox("Arxiv", value=st.session_state['run_arxiv'])
+    with colD:
+        st.session_state['autorun_searches'] = st.checkbox("Autorun", value=st.session_state['autorun_searches'])
+    with colE:
+        st.session_state['autosave_output'] = st.checkbox("AutoSave", value=st.session_state['autosave_output'])
+
+    # 3) Display the marquee welcome
     display_marquee(st.session_state['marquee_content'], 
                     {**marquee_settings, "font-size": "28px", "lineHeight": "50px"},
                     key_suffix="welcome")
 
-    # 3) Main action tabs
-    tab_main = st.radio("Action:", ["🎤 Voice", "📸 Media", "🔍 ArXiv", "📝 Editor"], 
-                        horizontal=True)
-
-    # Example custom component usage
+    # Example custom component usage:
     mycomponent = components.declare_component("mycomponent", path="mycomponent")
     val = mycomponent(my_input_value="Hello")
 
@@ -652,52 +696,33 @@ def main():
         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:
-            autorun = st.checkbox("⚙ AutoRun", value=True)
-        with col2:
-            full_audio = st.checkbox("📚FullAudio", value=False)
-
+        full_audio = st.checkbox("📚FullAudio", value=False)
         input_changed = (val != st.session_state.old_val)
 
-        if autorun and input_changed:
+        # We'll define: if autorun is on, run immediately after input changes
+        if st.session_state['autorun_searches'] and input_changed:
             st.session_state.old_val = val
             st.session_state.last_query = edited_input
-            perform_ai_lookup(edited_input, 
-                              vocal_summary=True, 
-                              extended_refs=False, 
-                              titles_summary=True, 
-                              full_audio=full_audio)
+            process_voice_input(edited_input)
         else:
             if st.button("▶ Run"):
                 st.session_state.old_val = val
                 st.session_state.last_query = edited_input
-                perform_ai_lookup(edited_input, 
-                                  vocal_summary=True, 
-                                  extended_refs=False, 
-                                  titles_summary=True, 
-                                  full_audio=full_audio)
-    
+                process_voice_input(edited_input)
+
     # ─────────────────────────────────────────────────────────
     # TAB: ArXiv
+    # (kept for demonstration if user chooses to do Arxiv only)
     # ─────────────────────────────────────────────────────────
     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")
-        titles_summary = st.checkbox("🔖TitlesOnly", value=True, key="option_titles_summary")
-        full_audio = st.checkbox("📚FullAudio", value=False, key="option_full_audio")
-        full_transcript = st.checkbox("🧾FullTranscript", value=False, key="option_full_transcript")
-        
         if q and st.button("🔍Run"):
             st.session_state.last_query = q
-            result = perform_ai_lookup(q, vocal_summary=vocal_summary, extended_refs=extended_refs, 
-                                       titles_summary=titles_summary, full_audio=full_audio)
-            if full_transcript:
-                create_file(q, result, "md")
+            # Even if the tab is "ArXiv," we can just call our standard function 
+            # that uses whichever checkboxes are selected (or you can do Arxiv only).
+            process_voice_input(q)
 
     # ─────────────────────────────────────────────────────────
     # TAB: Voice
@@ -710,13 +735,15 @@ def main():
             "Select TTS Voice:",
             options=EDGE_TTS_VOICES,
             index=EDGE_TTS_VOICES.index(st.session_state['tts_voice'])
+            if st.session_state['tts_voice'] in EDGE_TTS_VOICES
+            else 0
         )
         
         st.markdown("### 🔊 Audio Format")
         selected_format = st.radio(
             "Choose Audio Format:",
             options=["MP3", "WAV"],
-            index=0
+            index=0 if st.session_state['audio_format'] == "mp3" else 1
         )
 
         # Update session state if voice/format changes
@@ -727,11 +754,10 @@ def main():
             st.session_state['audio_format'] = selected_format.lower()
             st.rerun()
 
-        # Input text
-        user_text = st.text_area("💬 Message:", height=100)
-        user_text = user_text.strip().replace('\n', ' ')
-
+        user_text = st.text_area("💬 Message:", height=100).strip().replace('\n', ' ')
         if st.button("📨 Send"):
+            st.session_state.last_query = user_text
+            # If autorun is off, we explicitly run
             process_voice_input(user_text)
 
         st.subheader("📜 Chat History")
@@ -744,8 +770,6 @@ def main():
     # ─────────────────────────────────────────────────────────
     elif tab_main == "📸 Media":
         st.header("📸 Media Gallery")
-
-        # By default, show audio first
         tabs = st.tabs(["🎵 Audio", "🖼 Images", "🎥 Video"])
         
         # AUDIO sub-tab