Update app.py

app.py

@@ -184,76 +184,122 @@ def main():
     if "chat_history" not in st.session_state:
         st.session_state.chat_history = []
 
-    action = st.selectbox(
-        "Choose Action",
-        ["Ask a Question", "Capture Photo", "Capture Video"]
+    uploaded_files = st.file_uploader(
+        "Upload files (.txt, .docx, .csv, .xlsx, .pptx, .pdf, .html, .htm, .tex, .jpg, .jpeg, .png, video formats):",
+        type=[
+            "txt", "docx", "csv", "xlsx", "pptx", "pdf", "html", "htm", "tex",
+            "jpg", "jpeg", "png",
+            "mp4", "webm", "mov", "avi", "mkv", "flv", "wmv", "mpeg"
+        ],
+        accept_multiple_files=True
     )
 
-    if action == "Capture Photo":
+    if uploaded_files:
+        all_text_content = ""
+        for uploaded_file in uploaded_files:
+            filetype = uploaded_file.name.split(".")[-1].lower()
+            try:
+                if filetype == "txt":
+                    all_text_content += uploaded_file.read().decode("utf-8") + "\n"
+                elif filetype == "docx":
+                    all_text_content += extract_text_from_docx(uploaded_file) + "\n"
+                elif filetype == "csv":
+                    all_text_content += extract_text_from_csv(uploaded_file) + "\n"
+                elif filetype == "xlsx":
+                    all_text_content += extract_text_from_xlsx(uploaded_file) + "\n"
+                elif filetype == "pptx":
+                    all_text_content += extract_text_from_pptx(uploaded_file) + "\n"
+                elif filetype == "pdf":
+                    all_text_content += extract_text_from_pdf(uploaded_file) + "\n"
+                elif filetype in ["html", "htm"]:
+                    all_text_content += extract_text_from_html(uploaded_file) + "\n"
+                elif filetype == "tex":
+                    all_text_content += extract_text_from_tex(uploaded_file) + "\n"
+                elif filetype in ["jpg", "jpeg", "png"]:
+                    st.session_state.image_data = process_image(uploaded_file)
+                    all_text_content += "Image uploaded and processed.\n"
+                    st.image(uploaded_file, caption=uploaded_file.name, use_container_width=True)
+                elif filetype in ["mp4", "webm", "mov", "avi", "mkv", "flv", "wmv", "mpeg"]:
+                    st.session_state.video_data = process_video(uploaded_file)
+                    all_text_content += f"Video file '{uploaded_file.name}' uploaded and processed.\n"
+                    st.video(uploaded_file)
+                else:
+                    st.error(f"Unsupported file format: {uploaded_file.name}")
+            except Exception as e:
+                st.error(f"Failed to extract/process {uploaded_file.name}: {e}")
+
+        st.session_state.documents_text = all_text_content
+
+    # Media Capture: Buttons to Capture Photo or Video
+    st.sidebar.header("Capture Media")
+    capture_option = st.sidebar.radio("Select Media Capture Option", ["None", "Capture Photo", "Capture Video"])
+
+    if capture_option == "Capture Photo":
         capture_photo()
-
-    elif action == "Capture Video":
+    elif capture_option == "Capture Video":
         capture_video()
 
-    elif action == "Ask a Question":
-        if st.session_state.documents_text:
-            st.markdown("### \U0001F4AC Conversation")
-            for user_q, gemini_a in st.session_state.conversation:
-                st.markdown(f"**You:** {user_q}")
-                st.markdown(f"**Gemini:**\n\n{gemini_a}")
-
-            if st.session_state.chat_active:
-                with st.form(key="chat_form", clear_on_submit=True):
-                    user_input = st.text_input("Ask a question about the documents (type 'exit' to stop):")
-                    submit = st.form_submit_button("Send")
-
-                    if submit and user_input:
-                        if user_input.strip().lower() == "exit":
-                            st.session_state.chat_active = False
-                            st.success("Chat ended. Reload to start again.")
-                        else:
-                            with st.spinner("Let me think..."):
-                                content_blocks = []
-
-                                # System prompt for context
-                                content_blocks.append({
-                                    "text": (
-                                        "Your role is to analyze documents, images, and videos to help identify and classify driver states..."
-                                    )
-                                })
-
-                                if st.session_state.conversation:
-                                    history_text = "\n\n".join(
-                                        f"Q: {entry[0]}\nA: {entry[1]}"
-                                        for entry in st.session_state.conversation
-                                    )
-                                    content_blocks.append({"text": f"Previous conversation:\n{history_text}"})
-
-                                if st.session_state.documents_text.strip():
-                                    content_blocks.append({"text": f"Context:\n{st.session_state.documents_text}"})
-
-                                if "image_data" in st.session_state:
-                                    content_blocks.append(st.session_state.image_data)
-
-                                if "video_data" in st.session_state:
-                                    content_blocks.append(st.session_state.video_data)
-
-                                content_blocks.append({"text": f"Question: {user_input}"})
-
-                                response = client.models.generate_content(
-                                    model="gemini-2.0-flash",
-                                    contents=content_blocks,
-                                    config={"tools": [{"google_search": {}}]}
+    # Display conversation and chat
+    if st.session_state.documents_text:
+        st.markdown("### \U0001F4AC Conversation")
+        for user_q, gemini_a in st.session_state.conversation:
+            st.markdown(f"**You:** {user_q}")
+            st.markdown(f"**Gemini:**\n\n{gemini_a}")
+
+        if st.session_state.chat_active:
+            with st.form(key="chat_form", clear_on_submit=True):
+                user_input = st.text_input("Ask a question about the documents (type 'exit' to stop):")
+                submit = st.form_submit_button("Send")
+
+                if submit and user_input:
+                    if user_input.strip().lower() == "exit":
+                        st.session_state.chat_active = False
+                        st.success("Chat ended. Reload to start again.")
+                    else:
+                        with st.spinner("Let me think..."):
+                            content_blocks = []
+
+                            content_blocks.append({
+                                "text": (
+                                    "Your role is to analyze documents, images, and videos to help identify and classify driver states, "
+                                    "including: drowsy, alert, yawning, microsleep, eyes/lips open or closed, and other distractions. "
+                                    "The user may ask for timestamps of events in video, technical explanations from code/docs, "
+                                    "or structured summaries. Prioritize your answers for this use case and act as a helpful assistant "
+                                    "within this domain."
+                                )
+                            })
+
+                            if st.session_state.conversation:
+                                history_text = "\n\n".join(
+                                    f"Q: {entry[0]}\nA: {entry[1]}" for entry in st.session_state.conversation
                                 )
+                                content_blocks.append({"text": f"Previous conversation:\n{history_text}"})
+
+                            if st.session_state.documents_text.strip():
+                                content_blocks.append({"text": f"Context:\n{st.session_state.documents_text}"})
+
+                            if "image_data" in st.session_state:
+                                content_blocks.append(st.session_state.image_data)
+
+                            if "video_data" in st.session_state:
+                                content_blocks.append(st.session_state.video_data)
+
+                            content_blocks.append({"text": f"Question: {user_input}"})
+
+                            response = client.models.generate_content(
+                                model="gemini-2.0-flash",
+                                contents=content_blocks,
+                                config={"tools": [{"google_search": {}}]}
+                            )
 
-                                st.session_state.conversation.append((user_input, response.text))
-                                st.success("\U0001F4A1 Answer:")
-                                st.write(response.text)
+                            st.session_state.conversation.append((user_input, response.text))
+                            st.success("\U0001F4A1 Answer:")
+                            st.write(response.text)
 
-                                st.session_state.chat_history.append({
-                                    "question": user_input,
-                                    "answer": response.text
-                                })
+                            st.session_state.chat_history.append({
+                                "question": user_input,
+                                "answer": response.text
+                            })
 
         if st.session_state.conversation:
             pdf_path = export_conversation_to_pdf(st.session_state.conversation)