# ✅ Fully optimized app.py for Hugging Face Space with persistent 150GB storage from concurrent.futures import ThreadPoolExecutor, as_completed from threading import Thread # Use /data for persistent HF storage base_dir = "/data" model_cache_dir = os.path.join(base_dir, "txagent_models") tool_cache_dir = os.path.join(base_dir, "tool_cache") file_cache_dir = os.path.join(base_dir, "cache") report_dir = os.path.join(base_dir, "reports") vllm_cache_dir = os.path.join(base_dir, "vllm_cache") for d in [model_cache_dir, tool_cache_dir, file_cache_dir, report_dir, vllm_cache_dir]: os.makedirs(d, exist_ok=True) # Set persistent HF + VLLM cache os.environ.update({ "HF_HOME": model_cache_dir, "TRANSFORMERS_CACHE": model_cache_dir, "VLLM_CACHE_DIR": vllm_cache_dir, "TOKENIZERS_PARALLELISM": "false", "CUDA_LAUNCH_BLOCKING": "1" }) # Force local loading only LOCAL_TXAGENT_PATH = os.path.join(model_cache_dir, "mims-harvard", "TxAgent-T1-Llama-3.1-8B") LOCAL_RAG_PATH = os.path.join(model_cache_dir, "mims-harvard", "ToolRAG-T1-GTE-Qwen2-1.5B") sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), "src"))) from txagent.txagent import TxAgent def file_hash(path): return hashlib.md5(open(path, "rb").read()).hexdigest() def sanitize_utf8(text): return text.encode("utf-8", "ignore").decode("utf-8") MEDICAL_KEYWORDS = {"diagnosis", "assessment", "plan", "results", "medications", "summary", "findings"} def extract_priority_pages(file_path, max_pages=20): try: with pdfplumber.open(file_path) as pdf: pages = [] for i, page in enumerate(pdf.pages[:3]): pages.append(f"=== Page {i+1} ===\n{(page.extract_text() or '').strip()}") for i, page in enumerate(pdf.pages[3:max_pages], start=4): text = page.extract_text() or "" if any(re.search(rf'\\b{kw}\\b', text.lower()) for kw in MEDICAL_KEYWORDS): pages.append(f"=== Page {i} ===\n{text.strip()}") return "\n\n".join(pages) except Exception as e: return f"PDF processing error: {str(e)}" def convert_file_to_json(file_path, file_type): try: h = file_hash(file_path) cache_path = os.path.join(file_cache_dir, f"{h}.json") if os.path.exists(cache_path): return open(cache_path, "r", encoding="utf-8").read() if file_type == "pdf": text = extract_priority_pages(file_path) result = json.dumps({"filename": os.path.basename(file_path), "content": text, "status": "initial"}) Thread(target=full_pdf_processing, args=(file_path, h)).start() elif file_type == "csv": df = pd.read_csv(file_path, encoding_errors="replace", header=None, dtype=str) result = json.dumps({"filename": os.path.basename(file_path), "rows": df.fillna('').astype(str).values.tolist()}) elif file_type in ["xls", "xlsx"]: df = pd.read_excel(file_path, engine="openpyxl", header=None, dtype=str) result = json.dumps({"filename": os.path.basename(file_path), "rows": df.fillna('').astype(str).values.tolist()}) else: return json.dumps({"error": f"Unsupported file type: {file_type}"}) with open(cache_path, "w", encoding="utf-8") as f: f.write(result) return result except Exception as e: return json.dumps({"error": str(e)}) def full_pdf_processing(file_path, h): try: cache_path = os.path.join(file_cache_dir, f"{h}_full.json") if os.path.exists(cache_path): return with pdfplumber.open(file_path) as pdf: full_text = "\n".join([f"=== Page {i+1} ===\n{(p.extract_text() or '').strip()}" for i, p in enumerate(pdf.pages)]) with open(cache_path, "w", encoding="utf-8") as f: f.write(json.dumps({"content": full_text})) except: pass def init_agent(): target_tool_path = os.path.join(tool_cache_dir, "new_tool.json") if not os.path.exists(target_tool_path): shutil.copy(os.path.abspath("data/new_tool.json"), target_tool_path) agent = TxAgent( model_name=LOCAL_TXAGENT_PATH, rag_model_name=LOCAL_RAG_PATH, tool_files_dict={"new_tool": target_tool_path}, force_finish=True, enable_checker=True, step_rag_num=8, seed=100 ) agent.init_model() return agent agent_container = {"agent": None} def get_agent(): if agent_container["agent"] is None: agent_container["agent"] = init_agent() return agent_container["agent"] def create_ui(): with gr.Blocks(theme=gr.themes.Soft()) as demo: gr.Markdown("""

🩺 Clinical Oversight Assistant

""") chatbot = gr.Chatbot(label="Analysis", height=600) msg_input = gr.Textbox(placeholder="Ask a question about the patient...") file_upload = gr.File(file_types=[".pdf", ".csv", ".xls", ".xlsx"], file_count="multiple") send_btn = gr.Button("Analyze", variant="primary") state = gr.State([]) def analyze(message, history, conversation, files): try: extracted, hval = "", "" if files: with ThreadPoolExecutor(max_workers=3) as pool: futures = [pool.submit(convert_file_to_json, f.name, f.name.split(".")[-1].lower()) for f in files] extracted = "\n".join([sanitize_utf8(f.result()) for f in as_completed(futures)]) hval = file_hash(files[0].name) prompt = f"""Review these medical records and identify exactly what might have been missed: 1. Missed diagnoses 2. Medication conflicts 3. Incomplete assessments 4. Abnormal results needing follow-up Medical Records:\n{extracted[:15000]} """ final_response = "" for chunk in get_agent().run_gradio_chat(prompt, history=[], temperature=0.2, max_new_tokens=1024, max_token=4096, call_agent=False, conversation=conversation): if isinstance(chunk, str): final_response += chunk elif isinstance(chunk, list): final_response += "".join([c.content for c in chunk if hasattr(c, 'content')]) cleaned = final_response.replace("[TOOL_CALLS]", "").strip() updated_history = history + [[message, cleaned]] return updated_history, None except Exception as e: return history + [[message, f"❌ Error: {str(e)}"]], None send_btn.click(analyze, inputs=[msg_input, chatbot, state, file_upload], outputs=[chatbot, gr.File()]) msg_input.submit(analyze, inputs=[msg_input, chatbot, state, file_upload], outputs=[chatbot, gr.File()]) return demo if __name__ == "__main__": ui = create_ui() ui.queue(api_open=False).launch( server_name="0.0.0.0", server_port=7860, show_error=True, allowed_paths=["/data/reports"], share=False )