Spaces:
Sleeping
Sleeping
File size: 3,983 Bytes
2ee4d32 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 |
import streamlit as st
from crewai import Agent, Task, Crew, Process
import os
from crewai_tools import ScrapeWebsiteTool, SerperDevTool
from dotenv import load_dotenv
from langchain_openai import ChatOpenAI
from docx import Document
from io import BytesIO
import base64
load_dotenv()
# LLM object and API Key
os.environ["OPENAI_API_KEY"] = os.getenv("OPENAI_API_KEY")
os.environ["SERPER_API_KEY"] = os.getenv("SERPER_API_KEY")
def generate_docx(result):
doc = Document()
doc.add_heading('Healthcare Diagnosis and Treatment Recommendations', 0)
doc.add_paragraph(result)
bio = BytesIO()
doc.save(bio)
bio.seek(0)
return bio
def get_download_link(bio, filename):
b64 = base64.b64encode(bio.read()).decode()
return f'<a href="data:application/vnd.openxmlformats-officedocument.wordprocessingml.document;base64,{b64}" download="{filename}">Download Diagnosis and Treatment Plan</a>'
st.set_page_config(
layout="wide"
)
# Title
st.title("AI Agents to Empower Doctors")
# Text Inputs
gender = st.selectbox('Select Gender', ('Male', 'Female', 'Other'))
age = st.number_input('Enter Age', min_value=0, max_value=120, value=25)
symptoms = st.text_area('Enter Symptoms', 'e.g., fever, cough, headache')
medical_history = st.text_area('Enter Medical History', 'e.g., diabetes, hypertension')
# Initialize Tools
search_tool = SerperDevTool()
scrape_tool = ScrapeWebsiteTool()
llm = ChatOpenAI(
model="gpt-3.5-turbo-16k",
temperature=0.1,
max_tokens=8000
)
# Define Agents
diagnostician = Agent(
role="Medical Diagnostician",
goal="Analyze patient symptoms and medical history to provide a preliminary diagnosis.",
backstory="This agent specializes in diagnosing medical conditions based on patient-reported symptoms and medical history. It uses advanced algorithms and medical knowledge to identify potential health issues.",
verbose=True,
allow_delegation=False,
tools=[search_tool, scrape_tool],
llm=llm
)
treatment_advisor = Agent(
role="Treatment Advisor",
goal="Recommend appropriate treatment plans based on the diagnosis provided by the Medical Diagnostician.",
backstory="This agent specializes in creating treatment plans tailored to individual patient needs. It considers the diagnosis, patient history, and current best practices in medicine to recommend effective treatments.",
verbose=True,
allow_delegation=False,
tools=[search_tool, scrape_tool],
llm=llm
)
# Define Tasks
diagnose_task = Task(
description=(
"1. Analyze the patient's symptoms ({symptoms}) and medical history ({medical_history}).\n"
"2. Provide a preliminary diagnosis with possible conditions based on the provided information.\n"
"3. Limit the diagnosis to the most likely conditions."
),
expected_output="A preliminary diagnosis with a list of possible conditions.",
agent=diagnostician
)
treatment_task = Task(
description=(
"1. Based on the diagnosis, recommend appropriate treatment plans step by step.\n"
"2. Consider the patient's medical history ({medical_history}) and current symptoms ({symptoms}).\n"
"3. Provide detailed treatment recommendations, including medications, lifestyle changes, and follow-up care."
),
expected_output="A comprehensive treatment plan tailored to the patient's needs.",
agent=treatment_advisor
)
# Create Crew
crew = Crew(
agents=[diagnostician, treatment_advisor],
tasks=[diagnose_task, treatment_task],
verbose=2
)
# Execution
if st.button("Get Diagnosis and Treatment Plan"):
with st.spinner('Generating recommendations...'):
result = crew.kickoff(inputs={"symptoms": symptoms, "medical_history": medical_history})
st.write(result)
docx_file = generate_docx(result)
download_link = get_download_link(docx_file, "diagnosis_and_treatment_plan.docx")
st.markdown(download_link, unsafe_allow_html=True) |