Add user token input - users provide their own HF token

app.py

@@ -14,40 +14,43 @@ from langchain_core.messages import SystemMessage, HumanMessage
 # Load environment variables
 load_dotenv()
 
-# Global state - Initialize on startup
+# Global state - Initialize RAG only (not agent)
 rag = None
-agent = None
-initialized = False
-
-def initialize_on_startup():
-    """Initialize system automatically on startup."""
-    global rag, agent, initialized
 
+def initialize_rag():
+    """Initialize RAG system (embeddings only, no LLM needed)."""
+    global rag
     data_dir = Path("./data")
-
     if not data_dir.exists():
-        return
-
+        return False
     try:
         rag = ProjectRAG(data_dir)
         rag.load_and_index()
-
-        if rag.meetings:
-            agent = ProjectAgent(rag)
-            initialized = True
+        return True
     except Exception as e:
-        print(f"Initialization error: {e}")
+        print(f"RAG initialization error: {e}")
+        return False
 
-# Initialize on module load
-initialize_on_startup()
+# Initialize RAG on module load
+initialize_rag()
 
-def chat(message, history, project_filter):
+def chat(message, history, project_filter, hf_token):
     """Process chat message."""
-    if not initialized or not agent:
-        yield "⚠️ System initializing... Please wait and try again."
+    if not hf_token or hf_token.strip() == "":
+        yield "⚠️ Please enter your HuggingFace token first (get one at https://huggingface.co/settings/tokens)"
+        return
+
+    if not rag:
+        yield "⚠️ System not initialized. Please check the data directory."
         return
 
     try:
+        # Set token in environment for this request
+        os.environ["HF_TOKEN"] = hf_token.strip()
+
+        # Create agent with user's token
+        agent = ProjectAgent(rag)
+
         # Add project context if specified
         if project_filter and project_filter != "All Projects":
             enhanced_prompt = f"[Project: {project_filter}] {message}"
@@ -57,30 +60,31 @@ def chat(message, history, project_filter):
         response = agent.query(enhanced_prompt)
         yield response
     except Exception as e:
-        yield f"❌ Error: {str(e)}"
+        yield f"❌ Error: {str(e)}\n\nMake sure your HuggingFace token is valid."
 
 def get_projects():
     """Get list of projects."""
-    if not initialized or not rag:
+    if not rag:
         return ["All Projects"]
 
     projects = rag.get_all_projects()
     return ["All Projects"] + projects
 
-def structure_meeting(project_name, meeting_title, meeting_date, participants, meeting_text):
+def structure_meeting(project_name, meeting_title, meeting_date, participants, meeting_text, hf_token):
     """Structure meeting notes using AI."""
+    if not hf_token or hf_token.strip() == "":
+        return "❌ Please enter your HuggingFace token first"
+
     if not project_name or not meeting_text:
         return "❌ Please provide both project name and meeting notes"
 
     try:
-        # Use HF Inference API
-        # HF Spaces provides token as HF_TOKEN or HUGGING_FACE_HUB_TOKEN
-        hf_token = os.getenv("HF_TOKEN") or os.getenv("HUGGING_FACE_HUB_TOKEN")
+        # Use HF Inference API with user's token
         endpoint = HuggingFaceEndpoint(
             repo_id="meta-llama/Llama-3.2-3B-Instruct",
             temperature=0.3,
             max_new_tokens=1024,
-            huggingfacehub_api_token=hf_token
+            huggingfacehub_api_token=hf_token.strip()
         )
         llm = ChatHuggingFace(llm=endpoint)
 
@@ -172,6 +176,21 @@ with gr.Blocks(title="Sherlock: AI Project Assistant", theme=gr.themes.Soft(), c
         with gr.Tab("💬 Chat"):
             gr.Markdown("### Ask questions about your projects")
 
+            # HuggingFace Token input
+            with gr.Row():
+                hf_token_chat = gr.Textbox(
+                    label="🔑 HuggingFace Token (Required)",
+                    placeholder="Enter your HF token from https://huggingface.co/settings/tokens",
+                    type="password",
+                    scale=3
+                )
+                gr.Markdown("""
+                **Get a free token:**
+                1. Go to [huggingface.co/settings/tokens](https://huggingface.co/settings/tokens)
+                2. Create a new token with "Read" permissions
+                3. Paste it here
+                """, scale=1)
+
             # Project selection dropdown
             project_dropdown = gr.Dropdown(
                 label="Select Project",
@@ -217,13 +236,13 @@ with gr.Blocks(title="Sherlock: AI Project Assistant", theme=gr.themes.Soft(), c
                         submit_btn = gr.Button("Send", variant="primary", scale=1)
                         clear_btn = gr.Button("Clear", scale=1)
 
-            def respond(message, chat_history, project):
+            def respond(message, chat_history, project, token):
                 if not message:
                     return chat_history, ""
 
                 # Get bot response
                 bot_message = ""
-                for response_chunk in chat(message, chat_history, project):
+                for response_chunk in chat(message, chat_history, project, token):
                     bot_message = response_chunk
 
                 # Add to history as tuple
@@ -233,13 +252,13 @@ with gr.Blocks(title="Sherlock: AI Project Assistant", theme=gr.themes.Soft(), c
 
             submit_btn.click(
                 fn=respond,
-                inputs=[msg, chatbot, project_dropdown],
+                inputs=[msg, chatbot, project_dropdown, hf_token_chat],
                 outputs=[chatbot, msg]
             )
 
             msg.submit(
                 fn=respond,
-                inputs=[msg, chatbot, project_dropdown],
+                inputs=[msg, chatbot, project_dropdown, hf_token_chat],
                 outputs=[chatbot, msg]
             )
 
@@ -249,6 +268,13 @@ with gr.Blocks(title="Sherlock: AI Project Assistant", theme=gr.themes.Soft(), c
         with gr.Tab("📤 Upload Meeting"):
             gr.Markdown("### Upload plain text meeting notes and let AI structure them")
 
+            # HuggingFace Token input
+            hf_token_upload = gr.Textbox(
+                label="🔑 HuggingFace Token (Required)",
+                placeholder="Enter your HF token from https://huggingface.co/settings/tokens",
+                type="password"
+            )
+
             # Project selection with toggle
             with gr.Row():
                 with gr.Column():
@@ -312,15 +338,15 @@ Charlie is blocked waiting for API credentials.""",
             structure_btn = gr.Button("🤖 Structure Meeting with AI", variant="primary")
             structure_output = gr.Markdown(label="Structured Output")
 
-            def structure_meeting_wrapper(mode, existing_proj, new_proj, title, date, participants, text):
+            def structure_meeting_wrapper(mode, existing_proj, new_proj, title, date, participants, text, token):
                 """Wrapper to handle both project modes."""
                 # Determine which project name to use
                 project_name = existing_proj if mode == "Use Existing Project" else new_proj
-                return structure_meeting(project_name, title, date, participants, text)
+                return structure_meeting(project_name, title, date, participants, text, token)
 
             structure_btn.click(
                 fn=structure_meeting_wrapper,
-                inputs=[project_mode, existing_project, new_project, upload_title, upload_date, upload_participants, upload_text],
+                inputs=[project_mode, existing_project, new_project, upload_title, upload_date, upload_participants, upload_text, hf_token_upload],
                 outputs=structure_output
             )