import streamlit as st

from langchain_openai import ChatOpenAI

import base64

from PIL import Image

from langchain.schema.messages import HumanMessage,AIMessage

llm=ChatOpenAI(model="gpt-4o",max_tokens=2048)

def encode_image(upload_file):

image_bytes=upload_file.getvalue()

base64_image=base64.b64encode(image_bytes).decode("utf-8")

return base64_image

def get_response(b64image):

msg=llm.invoke(

[

AIMessage(

content="""

you are intelligent assistant who can solve any mathematical problems on derivatives,

you will be shared with a image where each line contains a problem ,

your task will be to solve all of them with possible explanation and step by step solutions

and formulas , then create a complete answer book

"""

),

HumanMessage(

content=[

{"type":"image_url",

"image_url":{

"url":"data:image/jpg;base64, " + b64image,

"detail":"auto"

}

}

]

)

]

)

content=msg.content

content=content.replace("\\","")

return content

def main():

st.title="Student Assignment Solver"

upload_file=st.file_uploader("upload your assignment:",type=["jpg","png"])

if upload_file is not None:

image=Image.open(upload_file)

st.image(image,caption="your assignment",use_column_width=True)

st.text("your assignment uploaded successfully")

base64_image=encode_image(upload_file)

btn=st.button("submit")

if btn:

response=get_response(base64_image)

print(response)

st.markdown(response,unsafe_allow_html=True)

if __name__=="__main__":

main()