app.py

import gradio as gr
import openai
from openai import OpenAI
import google.generativeai as genai
import os
import io
import base64

# Set API key
#base_url = os.environ.get("OPENAI_API_BASE")
api_key = ""

endpoints = ''
    
# Define the model to be used
MODEL = os.environ.get("MODEL")

DESCRIPTION = '''
<div>
<h1 style="text-align: center;">Medster - Medical Diagnostic Assistant</h1>
<p>An AI tool that helps you analyze symptoms and test reports.  </p>
<p>🔎 Select the department you need to consult, and enter the symptom description or physical examination information in the input box; you can also upload the test report image in the picture box.  </p>
<p>🦕 Please note that the generated information may be inaccurate and does not have any actual reference value. Please contact a professional doctor if necessary.  </p>
</div>
'''


css = """
h1 {
    text-align: center;
    display: block;
}
footer {
    display:none !important
}
"""


LICENSE = 'MODEL: ' + MODEL + ' LOADED'


def endpoints(api_key):
    if api_key not None:
        if api_key[:3] == "sk-":
            return 'OPENAI'
        else:
            return 'GOOGLE'
    return

endpoints = endpoints(api_key)

def read(filename):
    with open(filename) as f:
        data = f.read()
    return data

SYS_PROMPT = read('system_prompt.txt')

def process_text(text_input, unit):
    if text_input and endpoints == 'OPENAI':
        client = OpenAI(api_key=api_key)
        completion = client.chat.completions.create(
            model=MODEL,
            messages=[
                {"role": "system", "content": f" You are a experienced {unit} doctor." + SYS_PROMPT},
                {"role": "user", "content": f"Hello! Could you solve {text_input}?"}
            ]
        )
        return completion.choices[0].message.content
    elif text_input and endpoints == "GOOGLE":
        genai.configure(api_key=GOOGLE_API_KEY)
        model = genai.GenerativeModel(MODEL),
        return response = model.generate_content(SYS_PROMPT + text_input).text
    return ""

def encode_image_to_base64(image_input):
    buffered = io.BytesIO()
    image_input.save(buffered, format="JPEG")
    img_str = base64.b64encode(buffered.getvalue()).decode("utf-8")
    return img_str

def process_image(image_input, unit):
    if image_input is not None and endpoints == 'OPENAI':
        client = OpenAI(api_key=api_key)
        #with open(image_input.name, "rb") as f:
        #    base64_image = base64.b64encode(f.read()).decode("utf-8")
        base64_image = encode_image_to_base64(image_input)
        response = client.chat.completions.create(
            model=MODEL,
            messages=[
                {"role": "system", "content": f" You are a experienced {unit} doctor." + SYS_PROMPT},
                {"role": "user", "content": [
                    {"type": "text", "text": "Help me understand what is in this picture and analysis."},
                    {"type": "image_url", 
                     "image_url": {
                        "url": f"data:image/jpeg;base64,{base64_image}",
                        "detail":"low"}
                    }
                ]}
            ],
            temperature=0.0,
            max_tokens=1024,
        )
        return response.choices[0].message.content
    elif image_input is not None and ENDPOINTS == "GOOGLE":
        genai.configure(api_key=GOOGLE_API_KEY)
        model = genai.GenerativeModel(MODEL),
        return response = model.generate_content(image_input).text


def main(text_input="", image_input=None, unit=""):
    if text_input and image_input is None:
        return process_text(text_input,unit)
    elif image_input is not None:
        return process_image(image_input,unit)

with gr.Blocks(theme='shivi/calm_seafoam', css=css, title="Medster - Medical Diagnostic Assistant") as iface:
    with gr.Accordion(""):
        gr.Markdown(DESCRIPTION)
        unit = gr.Dropdown(label="🩺Department", value='Traditional Medicine', elem_id="units",
                             choices=["Traditional Medicine", "Internal Medicine", "Surgery", "Obstetrics and Gynecology", "Pediatrics", \
                                      "Orthodontics", "Andrology", "Dermatology and Venereology", "Infectious Diseases", "Psychiatry", \
                                         "Plastic Surgery Department", "Nutrition Department", "Reproductive Center", "Anesthesiology Department", "Medical Imaging Department", \
                                             "Orthopedics", "Oncology", "Emergency Department", "Laboratory Department"])
    with gr.Row():
        output_box = gr.Markdown(label="Diagnosis")  # Create an output textbox
    with gr.Row():
        api_key = gr.Textbox(label="API Key") # Input API key
    with gr.Row():
        image_input = gr.Image(type="pil", label="Upload Image")  # Create an image upload button
        text_input = gr.Textbox(label="Submit")  # Create a text input box
    with gr.Row():
        submit_btn = gr.Button("🚀 Send")  # Create a submit button
        clear_btn = gr.ClearButton(output_box, value="🗑️ Clear") # Create a clear button

    # Set up the event listeners
    submit_btn.click(main, inputs=[api_key, text_input, image_input, unit], outputs=output_box)

    gr.Markdown(LICENSE)
    
#gr.close_all()

iface.queue().launch(show_api=False)  # Launch the Gradio interface