import streamlit as st
hide_streamlit_style = """
"""
st.markdown(hide_streamlit_style, unsafe_allow_html=True)
def free_version():
import torch
import os
import argparse
import shutil
from langchain.document_loaders import YoutubeLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.vectorstores import Chroma
from langchain.embeddings import OpenAIEmbeddings
from langchain.chains import RetrievalQA
from langchain.llms import OpenAI
import streamlit as st
from langchain.chat_models import ChatOpenAI
from langchain import HuggingFaceHub
from langchain.embeddings import HuggingFaceInstructEmbeddings
from urllib.parse import urlparse, parse_qs
from langchain.embeddings import HuggingFaceBgeEmbeddings
from transformers import pipeline
import textwrap
import time
from deep_translator import GoogleTranslator
from langdetect import detect
def typewriter(text: str, speed: float):
container = st.empty()
displayed_text = ""
for char in text:
displayed_text += char
container.markdown(displayed_text)
time.sleep(1/speed)
def wrap_text_preserve_newlines(text, width=110):
# Split the input text into lines based on newline characters
lines = text.split('\n')
# Wrap each line individually
wrapped_lines = [textwrap.fill(line, width=width) for line in lines]
# Join the wrapped lines back together using newline characters
wrapped_text = '\n'.join(wrapped_lines)
return wrapped_text
def process_llm_response(llm_originalresponse2):
#result_text = wrap_text_preserve_newlines(llm_originalresponse2["result"])
typewriter(llm_originalresponse2["result"], speed=40)
def extract_video_id(youtube_url):
try:
parsed_url = urlparse(youtube_url)
query_params = parse_qs(parsed_url.query)
video_id = query_params.get('v', [None])[0]
return video_id
except Exception as e:
print(f"Error extracting video ID: {e}")
return None
def set_openAi_api_key(api_key: str):
st.session_state["OPENAI_API_KEY"] = api_key
os.environ['OPENAI_API_KEY'] = api_key
def openai_api_insert_component():
with st.sidebar:
st.markdown(
"""
## Quick Guide π
1. Get started by adding your [OpenAI API key](https://platform.openai.com/account/api-keys) belowπ
2. Easily input the video url
3. Engage with the content - ask questions, seek answersπ¬
"""
)
api_key_input = st.text_input("Input your OpenAI API Key",
type="password",
placeholder="Format: sk-...",
help="You can get your API key from https://platform.openai.com/account/api-keys.")
if api_key_input == "" or api_key_input is None:
st.sidebar.caption("π :red[Please set your OpenAI API Key here]")
st.caption(":green[Your API is not stored anywhere. It is only used to generate answers to your questions.]")
set_openAi_api_key(api_key_input)
def launchfreeversion():
HUGGINGFACE_API_TOKEN = os.environ['access_code']
model_name = "BAAI/bge-base-en"
encode_kwargs = {'normalize_embeddings': True}
st.title('MKG: Your Chat with Youtube Assistant')
videourl = st.text_input("Insert The video URL", placeholder="Format should be like: https://www.youtube.com/watch?v=pSLeYvld8Mk")
query = st.text_input("Ask any question about the video",help="Suggested queries: Summarize the key points of this video - What is this video about - Ask about a specific thing in the video ")
st.warning("β οΈ Please Keep in mind that the accuracy of the response relies on the :red[Video's quality] and the :red[prompt's Quality]. Occasionally, the response may not be entirely accurate. Consider using the response as a reference rather than a definitive answer.")
if st.button("Submit Question", type="primary"):
with st.spinner('Processing the Video...'):
video_id = extract_video_id(videourl)
loader = YoutubeLoader(video_id)
documents = loader.load()
text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
documents = text_splitter.split_documents(documents)
vectordb = Chroma.from_documents(
documents,
#embedding = HuggingFaceInstructEmbeddings(model_name="hkunlp/instructor-xl",
# model_kwargs={"device": "cuda"})
embedding= HuggingFaceBgeEmbeddings( model_name=model_name, model_kwargs={'device': 'cuda' if torch.cuda.is_available() else 'cpu'}, encode_kwargs=encode_kwargs)
)
repo_id = "tiiuae/falcon-7b-instruct"
qa_chain = RetrievalQA.from_chain_type(
llm=HuggingFaceHub(huggingfacehub_api_token=HUGGINGFACE_API_TOKEN,
repo_id=repo_id,
model_kwargs={"temperature":0.1, "max_new_tokens":1000}),
retriever=vectordb.as_retriever(),
return_source_documents=False,
verbose=False
)
with st.spinner('Generating Answer...'):
llm_response = qa_chain(query)
#llm_originalresponse2=llm_response['result']
process_llm_response(llm_response)
launchfreeversion()
def intro():
st.markdown("""
# MKG: Your Chat with Youtube Assistant π¬π€
Welcome to MKG-Assistant, where AI meets Youtube! ππ
## Base Models
Q&A-Assistant is built on OpenAI's GPT 3.5 for the premium version and Falcon 7B instruct Model for the free version to enhance your websites browsing experience. Whether you're a student, researcher, or professional, we're here to simplify your interactions with the web. π‘π
## How to Get Started
1.Enter the Video URL.
2. Enter your API key.(Only if you chose the premium version. Key is not needed in the free version)
3. Ask questions using everyday language.
4. Get detailed, AI-generated answers.
5. Enjoy a smarter way to Interact with Youtube!
## It is Time to Dive in!
""")
page_names_to_funcs = {
"Main Page": intro,
"Open Source Edition (Free version)": free_version
}
#test
demo_name = st.sidebar.selectbox("Choose a version", page_names_to_funcs.keys())
page_names_to_funcs[demo_name]()
st.sidebar.markdown(' Connect on LinkedIn ', unsafe_allow_html=True)
st.sidebar.markdown(' Check out my GitHub ', unsafe_allow_html=True)