Update app.py

app.py

@@ -36,88 +36,21 @@ CONFIG = {
     }
 }
 
-DESCRIPTION = '''
-<div>
-<h1 style="text-align: center;">TxAgent: An AI Agent for Therapeutic Reasoning Across a Universe of Tools</h1>
-</div>
-'''
-
-INTRO = """
-Precision therapeutics require multimodal adaptive models that provide personalized treatment recommendations. 
-We introduce TxAgent, an AI agent that leverages multi-step reasoning and real-time biomedical knowledge 
-retrieval across a toolbox of 211 expert-curated tools to navigate complex drug interactions, 
-contraindications, and patient-specific treatment strategies, delivering evidence-grounded therapeutic decisions.
-"""
-
-LICENSE = """
-We welcome your feedback and suggestions to enhance your experience with TxAgent, and if you're interested 
-in collaboration, please email Marinka Zitnik and Shanghua Gao.
-
-### Medical Advice Disclaimer
-DISCLAIMER: THIS WEBSITE DOES NOT PROVIDE MEDICAL ADVICE
-The information, including but not limited to, text, graphics, images and other material contained on this 
-website are for informational purposes only. No material on this site is intended to be a substitute for 
-professional medical advice, diagnosis or treatment.
-"""
-
-PLACEHOLDER = """
-<div style="padding: 30px; text-align: center; display: flex; flex-direction: column; align-items: center;">
-   <h1 style="font-size: 28px; margin-bottom: 2px; opacity: 0.55;">TxAgent</h1>
-   <p style="font-size: 18px; margin-bottom: 2px; opacity: 0.65;">Tips before using TxAgent:</p>
-   <p style="font-size: 18px; margin-bottom: 2px; opacity: 0.55;">Please click clear🗑️ (top-right) to remove previous context before submitting a new question.</p>
-   <p style="font-size: 18px; margin-bottom: 2px; opacity: 0.55;">Click retry🔄 (below message) to get multiple versions of the answer.</p>
-</div>
-"""
-
-css = """
-h1 {
-  text-align: center;
-  display: block;
-}
-
-#duplicate-button {
-  margin: auto;
-  color: white;
-  background: #1565c0;
-  border-radius: 100vh;
-}
-.small-button button {
-    font-size: 12px !important;
-    padding: 4px 8px !important;
-    height: 6px !important;
-    width: 4px !important;
-}
-.gradio-accordion {
-    margin-top: 0px !important;
-    margin-bottom: 0px !important;
-}
-"""
-
-chat_css = """
-.gr-button { font-size: 20px !important; }
-.gr-button svg { width: 32px !important; height: 32px !important; }
-"""
-
 def safe_load_embeddings(filepath: str) -> any:
-    """Safely load embeddings with proper weights_only handling"""
     try:
         return torch.load(filepath, weights_only=True)
     except Exception as e:
         logger.warning(f"Secure load failed, trying with weights_only=False: {str(e)}")
         try:
-            with torch.serialization.safe_globals([torch.serialization._reconstruct]):
-                return torch.load(filepath, weights_only=False)
+            return torch.load(filepath, weights_only=False)
         except Exception as e:
             logger.error(f"Failed to load embeddings even with safe_globals: {str(e)}")
             return None
 
 def patch_embedding_loading():
-    """Monkey-patch the embedding loading functionality"""
     try:
         from txagent.toolrag import ToolRAGModel
 
-        original_load = ToolRAGModel.load_tool_desc_embedding
-
         def patched_load(self, tooluniverse):
             try:
                 if not os.path.exists(CONFIG["embedding_filename"]):
@@ -129,7 +62,6 @@ def patch_embedding_loading():
                     logger.error("Embedding is None, aborting.")
                     return False
 
-                # Ensure tools is a list (in case it's a generator)
                 tools = list(tooluniverse.get_all_tools()) if hasattr(tooluniverse, 'get_all_tools') else []
                 current_count = len(tools)
                 embedding_count = len(self.tool_desc_embedding)
@@ -159,11 +91,37 @@ def patch_embedding_loading():
         logger.error(f"Failed to patch embedding loading: {str(e)}")
         raise
 
+def prepare_tool_files():
+    os.makedirs(os.path.join(current_dir, 'data'), exist_ok=True)
+    if not os.path.exists(CONFIG["tool_files"]["new_tool"]):
+        logger.info("Generating tool list using ToolUniverse...")
+        tu = ToolUniverse()
+        tools = list(tu.get_all_tools()) if hasattr(tu, 'get_all_tools') else []
+        with open(CONFIG["tool_files"]["new_tool"], "w") as f:
+            json.dump(tools, f, indent=2)
+        logger.info(f"Saved {len(tools)} tools to {CONFIG['tool_files']['new_tool']}")
+
+def create_agent():
+    patch_embedding_loading()
+    prepare_tool_files()
+
+    agent = TxAgent(
+        CONFIG["model_name"],
+        CONFIG["rag_model_name"],
+        tool_files_dict=CONFIG["tool_files"],
+        force_finish=True,
+        enable_checker=True,
+        step_rag_num=10,
+        seed=100,
+        additional_default_tools=['DirectResponse', 'RequireClarification']
+    )
+    agent.init_model()
+    return agent
+
 def update_model_parameters(agent, enable_finish, enable_rag, enable_summary,
                           init_rag_num, step_rag_num, skip_last_k,
                           summary_mode, summary_skip_last_k, summary_context_length, 
                           force_finish, seed):
-    """Update model parameters"""
     updated_params = agent.update_parameters(
         enable_finish=enable_finish,
         enable_rag=enable_rag,
@@ -180,13 +138,11 @@ def update_model_parameters(agent, enable_finish, enable_rag, enable_summary,
     return updated_params
 
 def update_seed(agent):
-    """Update random seed"""
     seed = random.randint(0, 10000)
     updated_params = agent.update_parameters(seed=seed)
     return updated_params
 
 def handle_retry(agent, history, retry_data: gr.RetryData, temperature, max_new_tokens, max_tokens, multi_agent, conversation, max_round):
-    """Handle retry functionality"""
     print("Updated seed:", update_seed(agent))
     new_history = history[:retry_data.index]
     previous_prompt = history[retry_data.index]['content']
@@ -197,160 +153,42 @@ def handle_retry(agent, history, retry_data: gr.RetryData, temperature, max_new_
 PASSWORD = "mypassword"
 
 def check_password(input_password):
-    """Check password for protected settings"""
     if input_password == PASSWORD:
         return gr.update(visible=True), ""
     else:
         return gr.update(visible=False), "Incorrect password, try again!"
 
 def create_demo(agent):
-    """Create the Gradio interface"""
-    default_temperature = 0.3
-    default_max_new_tokens = 1024
-    default_max_tokens = 81920
-    default_max_round = 30
-    
-    question_examples = [
-        ['Given a 50-year-old patient experiencing severe acute pain and considering the use of the newly approved medication, Journavx, how should the dosage be adjusted considering the presence of moderate hepatic impairment?'],
-        ['Given a 50-year-old patient experiencing severe acute pain and considering the use of the newly approved medication, Journavx, how should the dosage be adjusted considering the presence of severe hepatic impairment?'],
-        ['A 30-year-old patient is taking Prozac to treat their depression. They were recently diagnosed with WHIM syndrome and require a treatment for that condition as well. Is Xolremdi suitable for this patient, considering contraindications?'],
-    ]
-
-    chatbot = gr.Chatbot(height=800, placeholder=PLACEHOLDER,
-                        label='TxAgent', type="messages", show_copy_button=True)
-    
-    with gr.Blocks(css=css) as demo:
-        gr.Markdown(DESCRIPTION)
-        gr.Markdown(INTRO)
-        
-        temperature_state = gr.State(value=default_temperature)
-        max_new_tokens_state = gr.State(value=default_max_new_tokens)
-        max_tokens_state = gr.State(value=default_max_tokens)
-        max_round_state = gr.State(value=default_max_round)
-        
-        chatbot.retry(
-            lambda *args: handle_retry(agent, *args), 
-            inputs=[chatbot, chatbot, temperature_state, max_new_tokens_state,
-                   max_tokens_state, gr.Checkbox(value=False, render=False), 
-                   gr.State([]), max_round_state]
-        )
-
-        gr.ChatInterface(
-            fn=lambda *args: agent.run_gradio_chat(*args),
-            chatbot=chatbot,
-            fill_height=True, 
-            fill_width=True, 
-            stop_btn=True,
-            additional_inputs_accordion=gr.Accordion(
-                label="⚙️ Inference Parameters", open=False, render=False),
-            additional_inputs=[
-                temperature_state, max_new_tokens_state, max_tokens_state,
-                gr.Checkbox(
-                    label="Activate multi-agent reasoning mode", 
-                    value=False, 
-                    render=False),
-                gr.State([]),
-                max_round_state,
-                gr.Number(label="Seed", value=100, render=False)
-            ],
-            examples=question_examples,
-            cache_examples=False,
-            css=chat_css,
-        )
-
-        with gr.Accordion("Settings", open=False):
-            temperature_slider = gr.Slider(
-                minimum=0,
-                maximum=1,
-                step=0.1,
-                value=default_temperature,
-                label="Temperature"
-            )
-            max_new_tokens_slider = gr.Slider(
-                minimum=128,
-                maximum=4096,
-                step=1,
-                value=default_max_new_tokens,
-                label="Max new tokens"
-            )
-            max_tokens_slider = gr.Slider(
-                minimum=128,
-                maximum=32000,
-                step=1,
-                value=default_max_tokens,
-                label="Max tokens"
-            )
-            max_round_slider = gr.Slider(
-                minimum=0,
-                maximum=50,
-                step=1,
-                value=default_max_round,
-                label="Max round")
-
-            temperature_slider.change(
-                lambda x: x, inputs=temperature_slider, outputs=temperature_state)
-            max_new_tokens_slider.change(
-                lambda x: x, inputs=max_new_tokens_slider, outputs=max_new_tokens_state)
-            max_tokens_slider.change(
-                lambda x: x, inputs=max_tokens_slider, outputs=max_tokens_state)
-            max_round_slider.change(
-                lambda x: x, inputs=max_round_slider, outputs=max_round_state)
+    chatbot = gr.Chatbot()
+    with gr.Blocks(css=chat_css) as demo:
+        with gr.Row():
+            gr.Markdown("""
+            # TxAgent Interface
+            Ask biomedical questions and get reasoning-based answers.
+            """)
+
+        user_input = gr.Textbox(label="Your question")
+        temperature = gr.Slider(0, 1, value=0.3, step=0.1, label="Temperature")
+        max_new_tokens = gr.Slider(128, 4096, value=1024, step=1, label="Max New Tokens")
+        max_tokens = gr.Slider(128, 81920, value=81920, step=1, label="Max Tokens")
+        max_round = gr.Slider(1, 30, value=30, step=1, label="Max Rounds")
+        multi_agent = gr.Checkbox(label="Multi-Agent Mode")
+
+        submit = gr.Button("Run TxAgent")
+
+        def run_agent(message, history, temperature, max_new_tokens, max_tokens, multi_agent, max_round):
+            return agent.run_gradio_chat(history + [{"role": "user", "content": message}],
+                                         temperature, max_new_tokens, max_tokens,
+                                         multi_agent, [], max_round)
+
+        submit.click(run_agent, inputs=[user_input, chatbot, temperature, max_new_tokens, max_tokens, multi_agent, max_round], outputs=chatbot)
 
-            password_input = gr.Textbox(
-                label="Enter Password for More Settings", type="password")
-            incorrect_message = gr.Textbox(visible=False, interactive=False)
-            
-            with gr.Accordion("⚙️ Settings", open=False, visible=False) as protected_accordion:
-                with gr.Row():
-                    with gr.Column(scale=1):
-                        with gr.Accordion("⚙️ Model Loading", open=False):
-                            model_name_input = gr.Textbox(
-                                label="Enter model path", value=CONFIG["model_name"])
-                            load_model_btn = gr.Button(value="Load Model")
-                            load_model_btn.click(
-                                agent.load_models, 
-                                inputs=model_name_input, 
-                                outputs=gr.Textbox(label="Status"))
-                    with gr.Column(scale=1):
-                        with gr.Accordion("⚙️ Functional Parameters", open=False):
-                            enable_finish = gr.Checkbox(label="Enable Finish", value=True)
-                            enable_rag = gr.Checkbox(label="Enable RAG", value=True)
-                            enable_summary = gr.Checkbox(label="Enable Summary", value=False)
-                            init_rag_num = gr.Number(label="Initial RAG Num", value=0)
-                            step_rag_num = gr.Number(label="Step RAG Num", value=10)
-                            skip_last_k = gr.Number(label="Skip Last K", value=0)
-                            summary_mode = gr.Textbox(label="Summary Mode", value='step')
-                            summary_skip_last_k = gr.Number(label="Summary Skip Last K", value=0)
-                            summary_context_length = gr.Number(label="Summary Context Length", value=None)
-                            force_finish = gr.Checkbox(label="Force FinalAnswer", value=True)
-                            seed = gr.Number(label="Seed", value=100)
-                            submit_btn = gr.Button("Update Parameters")
-                            updated_parameters_output = gr.JSON()
-                            submit_btn.click(
-                                lambda *args: update_model_parameters(agent, *args),
-                                inputs=[enable_finish, enable_rag, enable_summary, 
-                                      init_rag_num, step_rag_num, skip_last_k,
-                                      summary_mode, summary_skip_last_k, 
-                                      summary_context_length, force_finish, seed],
-                                outputs=updated_parameters_output
-                            )
-            
-            submit_button = gr.Button("Submit")
-            submit_button.click(
-                check_password,
-                inputs=password_input,
-                outputs=[protected_accordion, incorrect_message]
-            )
-        
-        gr.Markdown(LICENSE)
-    
     return demo
 
 def main():
-    """Main function to run the application"""
     agent = create_agent()
     demo = create_demo(agent)
     demo.launch(share=True)
 
 if __name__ == "__main__":
-    main()
+    main()