app.py

from dotenv import load_dotenv
load_dotenv()

import streamlit as st
import os
from PIL import Image
import google.generativeai as genai
genai.configure(api_key=os.getenv("GOOGLE_API_KEY"))


model = genai.GenerativeModel('gemini-pro-vision')


def get_gemini_response(input, image, prompt):
    response = model.generate_content([input, image[0], prompt])
    return response.text

def input_image_details(uploaded_file):
    if uploaded_file is not None:
        bytes_data = uploaded_file.getvalue()

        image_parts = [
            {
                "mime_type": uploaded_file.type,
                "data": bytes_data
            }
        ]
        return image_parts
    else:
        raise FileNotFoundError("No file uploaded")



####streamlit initialization

st.set_page_config(page_title="Multilanguage Invoice Extractor")
st.header("Multilanguage Invoice Extractor")
input = st.text_input("Input prompt: ", key="input")
uploaded_file = st.file_uploader("choose an image", type=["jpg", "jpeg", "png"])
image = ""

if uploaded_file is not None:
    image = Image.open(uploaded_file)
    st.image(image, caption="uploaded image", use_column_width=True)

submit = st.button("Talk to me")
input_prompt = """You are an expert accountant who is excellent at understanding invoices. We will upload an image as invoices and you will have to answer any questions based on the uploaded invoice image."""

### if submit is pressed
if submit:
    image_data = input_image_details(uploaded_file)
    response = get_gemini_response(input_prompt, image_data, input)
    st.subheader("The response is:")
    st.write(response)