app.py

# refer to repo https://github.com/gradio-app/gradio/blob/main/demo/chatbot_multimodal/run.ipynb for enhancement

import PIL.Image
import gradio as gr
import base64
import time
import os
import google.generativeai as genai

import pathlib

txt_model = genai.GenerativeModel('gemini-pro')
vis_model = genai.GenerativeModel('gemini-pro-vision')

txt_prompt_1 = """The image contains the contents of a letter. I'd like to follow the request mentioned in the letter. Please provide 3 actionable items to assist me. When responding, use the following format:

# Sender and Subject #
1- Action 1 (no more than 20 words)
2- Action 2 (no more than 20 words)
3- Action 3 (no more than 20 words)

For example:
# From Richard regarding 'Shipping to Customer ABC' #
1- Pack Product A
2- Ship before 3:00 PM today
3- Notify Richard after shipment
"""

txt_display_1 = 'content of the letter: '
import os
GOOGLE_API_KEY=os.getenv('GOOGLE_API_KEY')
genai.configure(api_key=GOOGLE_API_KEY)

# Image to Base 64 Converter
def image_to_base64(image_path):
    with open(image_path, 'rb') as img:
        encoded_string = base64.b64encode(img.read())
    return encoded_string.decode('utf-8')

def app1_query(img):
    if not img:
        return txt_prompt_1
    base64 = image_to_base64(img)
    data_url = f"data:image/jpeg;base64,{base64}"
    outputText = [(f"{txt_display_1} ![]({data_url})", None)]
    return outputText

# Function that takes User Inputs, generates Response and displays on Chat UI
def app1_response(img):
    if not img:
        response = txt_model.generate_content(txt_prompt_1)
        return response

    else:
        img = PIL.Image.open(img)
        response = vis_model.generate_content([txt_prompt_1,img])
        return response.text

        
# gradio block
    
with gr.Blocks(theme='snehilsanyal/scikit-learn') as app1:
    with gr.Column():    
        image_box = gr.Image(type="filepath")
        outputbox = gr.Textbox(label="here are the plans...")
        
    btn = gr.Button("Make a Plan")
    clicked = btn.click(app1_response,
                                [image_box],
                                outputbox
                        )
    gr.Markdown("""
    # Make a Plan #
    
    	- screen capture (Win + shift + S)
        - click **Make a Plan** to upload
        - await LLM Bot (Gemini, in this case) response
        - receive THREE actionable items
    
    [demo](https://youtu.be/lJ4jIAEVRNY)

    """)


app1.queue()
app1.launch()