app.py

import gradio as gr
import torch
from transformers import Qwen2VLForConditionalGeneration, AutoProcessor
from qwen_vl_utils import process_vision_info
from PIL import Image
import numpy as np
from datetime import datetime
import os

def array_to_image_path(image_array):
    if image_array is None:
        raise ValueError("No image provided. Please upload an image before submitting.")
    img = Image.fromarray(np.uint8(image_array))
    timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
    filename = f"image_{timestamp}.png"
    img.save(filename)
    return os.path.abspath(filename)

model = Qwen2VLForConditionalGeneration.from_pretrained(
    "Qwen/Qwen2-VL-7B-Instruct", trust_remote_code=True, torch_dtype="auto"
).cuda().eval()

processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-7B-Instruct", trust_remote_code=True)

@gr.GPU
def analyze_iol_report(image):
    image_path = array_to_image_path(image)
    prompt = "Extract all ophthalmic measurements including AL, K1, K2, ACD, LT, WTW, and CCT. Calculate IOL power based on formulas like Barrett Universal II, Cooke K6, EVO, Hill-RBF, Hoffer QST, Kane, and PEARL GDS."
    
    messages = [
        {
            "role": "user",
            "content": [
                {"type": "image", "image": image_path},
                {"type": "text", "text": prompt},
            ],
        }
    ]
    
    text = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
    image_inputs, video_inputs = process_vision_info(messages)
    inputs = processor(
        text=[text],
        images=image_inputs,
        videos=video_inputs,
        padding=True,
        return_tensors="pt",
    )
    inputs = inputs.to("cuda")
    generated_ids = model.generate(**inputs, max_new_tokens=1024)
    generated_ids_trimmed = [
        out_ids[len(in_ids):] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
    ]
    output_text = processor.batch_decode(
        generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
    )
    
    return output_text[0]

css = """
  #output {
    height: 500px; 
    overflow: auto; 
    border: 1px solid #ccc; 
  }
"""

with gr.Blocks(css=css) as demo:
    gr.Markdown("Fixed Model - Qwen2-VL-7B for IOL Report Analysis")
    with gr.Tab(label="IOL Analysis"):
        with gr.Row():
            input_img = gr.Image(label="Upload IOL Report Image")
            submit_btn = gr.Button(value="Analyze Report")
        output = gr.Textbox(label="Analysis Result", elem_id="output")
        submit_btn.click(analyze_iol_report, inputs=[input_img], outputs=[output])

demo.launch()