app.py

import requests
import streamlit as st
import os
from huggingface_hub import InferenceClient

API_URL = 'https://qe55p8afio98s0u3.us-east-1.aws.endpoints.huggingface.cloud'
API_KEY = os.getenv('API_KEY')

headers = {
    "Authorization": f"Bearer {API_KEY}",
    "Content-Type": "application/json"
}

endpoint_url = API_URL
hf_token = API_KEY
client = InferenceClient(endpoint_url, token=hf_token)
gen_kwargs = dict(
    max_new_tokens=512,
    top_k=30,
    top_p=0.9,
    temperature=0.2,
    repetition_penalty=1.02,
    stop_sequences=["\nUser:", "<|endoftext|>", "</s>"],
)
prompt = f"Write instructions to teach anyone to write a discharge plan. List the entities, features and relationships to CCDA and FHIR objects in boldface."
stream = client.text_generation(prompt, stream=True, details=True, **gen_kwargs)
report=[]
res_box = st.empty()
collected_chunks=[]
collected_messages=[]
for r in stream:
    if r.token.special:
        continue
    if r.token.text in gen_kwargs["stop_sequences"]:
        break
    collected_chunks.append(r.token.text)
    chunk_message = r.token.text
    collected_messages.append(chunk_message)

    try:
        report.append(content)
        if len(r.token.text) > 0:
            result="".join(report).strip()
            res_box.markdown(f'*{result}*')
    except:
        st.write(' ')
        
    #full_reply = ''.join()
    #st.markdown(r.token.text, end = "")
    #st.write(r.token.text)


def query(payload):
    response = requests.post(API_URL, headers=headers, json=payload)
    st.markdown(response.json())
    return response.json()

def get_output(prompt):
    return query({"inputs": prompt})

def main():
    st.title("Medical Llama Test Bench with Inference Endpoints Llama 7B")
    example_input = st.text_input("Enter your example text:")

    if st.button("Summarize with Variation 1"):
        prompt = f"Write instructions to teach anyone to write a discharge plan. List the entities, features and relationships to CCDA and FHIR objects in boldface. {example_input}"
        output = get_output(prompt)
        st.markdown(f"**Output:** {output}")

    if st.button("Summarize with Variation 2"):
        prompt = f"Provide a summary of the medical transcription. Highlight the important entities, features, and relationships to CCDA and FHIR objects. {example_input}"
        output = get_output(prompt)
        st.markdown(f"**Output:** {output}")

if __name__ == "__main__":
    main()