app.py

import streamlit as st
import pandas as pd
from transformers import pipeline, AutoConfig, AutoTokenizer, AutoModelForSeq2SeqLM, AutoModelForCausalLM, MistralForCausalLM
from peft import PeftModel, PeftConfig

#Note this should be used always in compliance with applicable laws and regulations if used with real patient data.

# Instantiate the Tokenizer
tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.1", trust_remote_code=True, padding_side="left")
tokenizer.pad_token = tokenizer.eos_token
tokenizer.padding_side = 'left'
# Load the PEFT model
peft_config = PeftConfig.from_pretrained("pseudolab/K23_MiniMed")
peft_model = MistralForCausalLM.from_pretrained("https://huggingface.co/HuggingFaceH4/zephyr-7b-beta", trust_remote_code=True)
peft_model = PeftModel.from_pretrained(peft_model, "pseudolab/K23_MiniMed")

#Upload Patient Data
uploaded_file = st.file_uploader("Choose a CSV file", type="csv")

# Prepare the context
def prepare_context(data):
    # Format the data as a string
    data_str = data.to_string(index=False, header=False)

    # Tokenize the data
    input_ids = tokenizer.encode(data_str, return_tensors="pt")

    # Truncate the input if it's too long for the model
    max_length = tokenizer.model_max_length
    if input_ids.shape[1] > max_length:
        input_ids = input_ids[:, :max_length]

    return input_ids

if uploaded_file is not None:
    data = pd.read_csv(uploaded_file)
    st.write(data)

    # Generate text based on the context
    context = prepare_context(data)
    generated_text = pipeline('text-generation', model=model)(context)[0]['generated_text']
    st.write(generated_text)

    # Internally prompt the model to data analyze the EHR patient data
    prompt = "You are an Electronic Health Records analyst with nursing school training. Please analyze patient data that you are provided here. Give an organized, step-by-step, formatted health records analysis. You will always be truthful and if you do nont know the answer say you do not know."

    if prompt:
        # Tokenize the prompt
        input_ids = tokenizer.encode(prompt, return_tensors="pt")

        # Generate text based on the prompt
        generated_text = pipeline('text-generation', model=model)(input_ids=input_ids)[0]['generated_text']
        st.write(generated_text)
    else:
        st.write("Please enter patient data")
        
else:
    st.write("No file uploaded")