Spaces:
Running
Running
from smolagents import CodeAgent, DuckDuckGoSearchTool, HfApiModel, load_tool, tool | |
import datetime | |
import requests | |
import pytz | |
import yaml | |
from tools.final_answer import FinalAnswerTool | |
from Gradio_UI import GradioUI | |
# Custom Tool to fetch datasets related to body parts or imaging types | |
def my_custom_tool(arg1: str, arg2: int) -> str: | |
""" | |
Search and retrieve publicly available medical datasets from Hugging Face based on any medical-related keyword. | |
Args: | |
arg1: A keyword related to medical data (e.g., 'cancer', 'diabetes', 'CT scan', 'radiology', 'dermoscopy'). | |
arg2: The maximum number of datasets to retrieve. | |
Returns: | |
A list of dataset names matching the search query, or a message stating that no datasets were found. | |
""" | |
try: | |
keyword = arg1.strip().lower() | |
limit = int(arg2) | |
# Define a basic list of medically relevant terms | |
medical_terms = [ | |
# Anatomy / Body Parts | |
"skin", "brain", "lung", "chest", "abdomen", "spine", "bone", "heart", "liver", "kidney", | |
"bladder", "stomach", "colon", "rectum", "esophagus", "pancreas", "breast", "ear", "eye", | |
"retina", "tooth", "teeth", "tongue", "jaw", "neck", "wrist", "hand", "leg", "arm", "shoulder", "pelvis", | |
# Diseases / Conditions | |
"cancer", "tumor", "stroke", "diabetes", "pneumonia", "covid", "asthma", "eczema", "melanoma", | |
"hypertension", "alzheimer", "parkinson", "arthritis", "scoliosis", "epilepsy", "glaucoma", | |
"ulcer", "hepatitis", "leukemia", "lymphoma", "tuberculosis", "anemia", "obesity", "depression", | |
"anxiety", "bipolar", "autism", "adhd", "ptsd", "psychosis", "schizophrenia", | |
# Imaging Modalities | |
"mri", "ct", "xray", "x-ray", "ultrasound", "pet", "fmri", "mammo", "angiography", "radiography", | |
"echocardiogram", "spect", "dermoscopy", "colonoscopy", "endoscopy", "biopsy", "histopathology", | |
# Medical Specialties | |
"radiology", "pathology", "oncology", "cardiology", "neurology", "dermatology", "dentistry", | |
"ophthalmology", "urology", "orthopedics", "gastroenterology", "pulmonology", "nephrology", | |
"psychiatry", "pediatrics", "geriatrics", "infectious disease", | |
# Symptoms / Signs | |
"lesion", "infection", "fever", "pain", "inflammation", "rash", "headache", "swelling", | |
"cough", "seizure", "dizziness", "vomiting", "diarrhea", "nausea", "fatigue", "itching", | |
# Common Specific Diseases | |
"breast cancer", "prostate cancer", "lung cancer", "skin cancer", "colon cancer", | |
"brain tumor", "liver cancer", "cervical cancer", "bladder cancer", "thyroid cancer", | |
# Procedures / Interventions | |
"surgery", "chemotherapy", "radiation", "transplant", "dialysis", "intubation", "stenting", | |
"ventilation", "vaccination", "anesthesia", "rehabilitation", "prosthetics", "orthotics", | |
# Lab Tests / Biomarkers | |
"blood test", "cbc", "glucose", "hemoglobin", "cholesterol", "biomarker", "urinalysis", | |
"pcr", "serology", "antibody", "antigen", | |
# Clinical Settings / Roles | |
"icu", "hospital", "emergency", "clinical notes", "nursing", "physician", "patient", | |
"medical record", "electronic health record", "ehr", "vitals", | |
# Age-based Terms | |
"pediatric", "neonatal", "infant", "child", "adolescent", "geriatrics", "elderly", | |
# Epidemiology / Public Health | |
"epidemiology", "prevalence", "incidence", "mortality", "public health", "health disparity", | |
"risk factor", "social determinant", | |
# Pharmacology / Medications | |
"drug", "medication", "pharmacology", "side effect", "adverse event", "dose", "tablet", | |
"vaccine", "clinical trial", "placebo" | |
] | |
# Check if keyword is in known medical terms | |
if not any(term in keyword for term in medical_terms): | |
return f"No medical datasets found for '{arg1}'." | |
# Fetch datasets from Hugging Face | |
response = requests.get( | |
f"https://huggingface.co/api/datasets?search={keyword}&limit={limit}" | |
) | |
response.raise_for_status() | |
datasets = response.json() | |
# Return message if no datasets found | |
if not datasets: | |
return f"No medical datasets found for '{arg1}'." | |
# Collect and return dataset names | |
results = [f"- {ds.get('id', 'Unknown')}" for ds in datasets[:limit]] | |
return f"Medical datasets related to '{arg1}':\n" + "\n".join(results) | |
except Exception as e: | |
return f"Error searching medical datasets for '{arg1}': {str(e)}" | |
def get_current_time_in_timezone(timezone: str) -> str: | |
""" | |
A tool that fetches the current local time in a specified timezone. | |
Args: | |
timezone: A string representing a valid timezone (e.g., 'America/New_York'). | |
Returns: | |
A string showing the current local time in the specified timezone. | |
""" | |
try: | |
tz = pytz.timezone(timezone) | |
local_time = datetime.datetime.now(tz).strftime("%Y-%m-%d %H:%M:%S") | |
return f"The current local time in {timezone} is: {local_time}" | |
except Exception as e: | |
return f"Error fetching time for timezone '{timezone}': {str(e)}" | |
final_answer = FinalAnswerTool() | |
# Model setup | |
model = HfApiModel( | |
max_tokens=2096, | |
temperature=0.5, | |
model_id='Qwen/Qwen2.5-Coder-32B-Instruct', # this model may be overloaded | |
custom_role_conversions=None, | |
) | |
# Load tool from hub | |
image_generation_tool = load_tool("agents-course/text-to-image", trust_remote_code=True) | |
# Load prompt templates | |
with open("prompts.yaml", 'r') as stream: | |
prompt_templates = yaml.safe_load(stream) | |
# Create the agent | |
agent = CodeAgent( | |
model=model, | |
tools=[final_answer, get_current_time_in_timezone, my_custom_tool], # add your tools here | |
max_steps=6, | |
verbosity_level=1, | |
grammar=None, | |
planning_interval=None, | |
name=None, | |
description=None, | |
prompt_templates=prompt_templates | |
) | |
# Launch the UI | |
GradioUI(agent).launch() | |