File size: 3,506 Bytes
3bc66db 5d14259 84cc5d0 aa67515 3bc66db afb439e 3bc66db |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 |
import gradio as gr
import os
from langchain_google_genai import GoogleGenerativeAIEmbeddings, ChatGoogleGenerativeAI
from langchain.document_loaders import YoutubeLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.vectorstores import FAISS
from langchain.chains import LLMChain
from langchain.prompts.chat import (
ChatPromptTemplate,
SystemMessagePromptTemplate,
HumanMessagePromptTemplate,
)
def create_db_from_video_url(video_url, api_key):
"""
Creates an Embedding of the Video and performs
"""
embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001", google_api_key=api_key)
loader = YoutubeLoader.from_youtube_url(video_url)
transcripts = loader.load()
# cannot provide this directly to the model so we are splitting the transcripts into small chunks
text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
docs = text_splitter.split_documents(transcripts)
db = FAISS.from_documents(docs, embedding=embeddings)
return db
def get_response(video, request):
"""
Usind Gemini Pro to get the response. It can handle upto 32k tokens.
"""
API_KEY = os.environ.get("API_Key")
db = create_db_from_video_url(video, API_KEY)
docs = db.similarity_search(query=request, k=5)
docs_content = " ".join([doc.page_content for doc in docs])
chat = ChatGoogleGenerativeAI(model="gemini-pro", google_api_key=API_KEY, convert_system_message_to_human=True)
# creating a template for request
template = """
You are an assistant that can answer questions about youtube videos based on
video transcripts: {docs}
Only use factual information from the transcript to answer the question.
If you don't have enough information to answer the question, say "I don't know".
Your Answers should be detailed.
"""
system_msg_prompt = SystemMessagePromptTemplate.from_template(template)
# human prompt
human_template = "Answer the following questions: {question}"
human_msg_prompt = HumanMessagePromptTemplate.from_template(human_template)
chat_prompt = ChatPromptTemplate.from_messages(
[system_msg_prompt, human_msg_prompt]
)
chain = LLMChain(llm=chat, prompt=chat_prompt)
response = chain.run(question=request, docs=docs_content)
return response
# creating title, description for the web app
title = "YouTube Video Assistant 🧑💻"
description = "Answers to the Questions asked by the user on the specified YouTube video. (English Only)\nClick here to view [demo](https://cdn-uploads.huggingface.co/production/uploads/641aa7814577db917f70f8aa/vSEGALDIYsqdRM7t_49rp.mp4)."
article = "Other Projects: \t"\
"💰 [Health Insurance Predictor](http://health-insurance-cost-predictor-k19.streamlit.app/) "\
"📰 [Fake News Detector](https://fake-news-detector-k19.streamlit.app/) "\
"🪶 [Birds Classifier](https://huggingface.co/spaces/Kathir0011/Birds_Classification)"
# building the app
youtube_video_assistant = gr.Interface(
fn=get_response,
inputs=[gr.Text(label="Enter the Youtube Video URL:", placeholder="Example: https://www.youtube.com/watch?v=MnDudvCyWpc"),
gr.Text(label="Enter your Question", placeholder="Example: What's the video is about?")],
outputs=gr.TextArea(label="Answers using Gemini Pro:"),
title=title,
description=description,
article=article
)
# launching the web app
youtube_video_assistant.launch()
|