app.py

import streamlit as st
import openai
import os
import base64
import glob
import json
import mistune
import pytz

from datetime import datetime
from openai import ChatCompletion
from xml.etree import ElementTree as ET
from bs4 import BeautifulSoup

openai.api_key = os.getenv('OPENAI_KEY')
st.set_page_config(
    page_title="GPT Streamlit Document Reasoner",
    layout="wide")
st.title("GPT Chat with Optional File Context - Talk to your data!")

# Output options sidebar menu
st.sidebar.title("Output Options")
menu = ["txt", "htm", "md"]
choice = st.sidebar.selectbox("Choose an output file type to save your prompt and results", menu)
choicePrefix = "Output and download file set to "
if choice == "txt":
     st.sidebar.write(choicePrefix + "Text file.")
elif choice == "htm":
     st.sidebar.write(choicePrefix + "HTML5.")
elif choice == "md":
     st.sidebar.write(choicePrefix + "Markdown.")

# sidebar slider for file input length to include in inference blocks
max_length = st.sidebar.slider("Max document length", min_value=1000, max_value=32000, value=3000, step=1000)

# Truncate document
def truncate_document(document, length):
    return document[:length]

def chat_with_model(prompts):
    model = "gpt-3.5-turbo"
    #model = "gpt-4-32k" # 32k tokens between prompt and inference tokens
    conversation = [{'role': 'system', 'content': 'You are a helpful assistant.'}]
    conversation.extend([{'role': 'user', 'content': prompt} for prompt in prompts])
    response = openai.ChatCompletion.create(model=model, messages=conversation)
    return response['choices'][0]['message']['content']

def generate_filename(prompt):
    central = pytz.timezone('US/Central')
    safe_date_time = datetime.now(central).strftime("%m%d_%I_%M_%p")  
    safe_prompt = "".join(x for x in prompt if x.isalnum())[:30]
    return f"{safe_date_time}_{safe_prompt}.txt"
    
def create_file(filename, prompt, response):
    with open(filename, 'w') as file:
        file.write(f"<h1>Prompt:</h1> <p>{prompt}</p> <h1>Response:</h1> <p>{response}</p>")

def get_table_download_link_old(file_path):
    with open(file_path, 'r') as file:
        data = file.read()
    b64 = base64.b64encode(data.encode()).decode()  
    href = f'<a href="data:file/htm;base64,{b64}" target="_blank" download="{os.path.basename(file_path)}">{os.path.basename(file_path)}</a>'
    return href

def get_table_download_link(file_path):
    import os
    import base64
    with open(file_path, 'r') as file:
        data = file.read()
    b64 = base64.b64encode(data.encode()).decode()  
    file_name = os.path.basename(file_path)
    ext = os.path.splitext(file_name)[1]  # get the file extension

    if ext == '.txt':
        mime_type = 'text/plain'
    elif ext == '.htm':
        mime_type = 'text/html'
    elif ext == '.md':
        mime_type = 'text/markdown'
    else:
        mime_type = 'application/octet-stream'  # general binary data type

    href = f'<a href="data:{mime_type};base64,{b64}" target="_blank" download="{file_name}">{file_name}</a>'
    return href

def CompressXML(xml_text):
    root = ET.fromstring(xml_text)
    for elem in list(root.iter()):
        if isinstance(elem.tag, str) and 'Comment' in elem.tag:
            elem.parent.remove(elem)
    #return ET.tostring(root, encoding='unicode', method="xml")
    return ET.tostring(root, encoding='unicode', method="xml")[:max_length]  
    
def read_file_content(file):
    if file.type == "application/json":
        content = json.load(file)
        return str(content)
    elif file.type == "text/html" or file.type == "text/htm":
        content = BeautifulSoup(file, "html.parser")
        return content.text
    elif file.type == "application/xml" or file.type == "text/xml":
        tree = ET.parse(file)
        root = tree.getroot()
        #return ET.tostring(root, encoding='unicode')
        return CompressXML(ET.tostring(root, encoding='unicode'))
    elif file.type == "text/markdown" or file.type == "text/md":
        md = mistune.create_markdown()
        content = md(file.read().decode())
        return content
    elif file.type == "text/plain":
        return file.getvalue().decode()
    else:
        return ""

def main():
    prompts = ['']
    file_content = ""
    user_prompt = st.text_area("Your question:", '', height=120)
    uploaded_file = st.file_uploader("Choose a file", type=["xml", "json", "html", "htm", "md", "txt"])

    if user_prompt:
        prompts.append(user_prompt)

    if uploaded_file is not None:
        file_content = read_file_content(uploaded_file)
        prompts.append(file_content)

    if st.button('💬 Chat'):
        st.write('Chatting with GPT-3...')
        response = chat_with_model(prompts)
        st.write('Response:')
        st.write(response)
        
        filename = generate_filename(user_prompt)
        create_file(filename, user_prompt, response)
        st.sidebar.markdown(get_table_download_link(filename), unsafe_allow_html=True)
    
    if len(file_content) > 0:
        st.markdown(f"**Content Added to Prompt:**\n{file_content}")
    
    htm_files = glob.glob("*.txt")
    for file in htm_files:
        st.sidebar.markdown(get_table_download_link(file), unsafe_allow_html=True)
        if st.sidebar.button(f"🗑Delete {file}"):
        #if st.sidebar.button("🗑 Delete"):
            os.remove(file)
            st.experimental_rerun()

if __name__ == "__main__":
    main()