Create app.py

app.py

@@ -0,0 +1,160 @@
+import os
+import gradio as gr
+import requests
+from bs4 import BeautifulSoup
+from youtube_transcript_api import YouTubeTranscriptApi
+from pytube import YouTube
+from PyPDF2 import PdfReader
+import docx
+import pptx
+import faiss
+import numpy as np
+
+from groq import Groq
+
+# Initialize Groq client
+client = Groq(api_key=os.environ.get("MY_API"))
+
+# -------------------------------
+# Global storage for embeddings + chunks
+# -------------------------------
+global_index = None
+global_chunks = []
+
+# -------------------------------
+# Utility Functions
+# -------------------------------
+
+def chunk_text(text, chunk_size=500):
+    return [text[i:i+chunk_size] for i in range(0, len(text), chunk_size)]
+
+def embed_text(chunks):
+    return [np.random.rand(768).astype("float32") for _ in chunks]
+
+def store_in_faiss(embeddings):
+    dim = len(embeddings[0])
+    index = faiss.IndexFlatL2(dim)
+    index.add(np.array(embeddings))
+    return index
+
+def summarize_text(text):
+    chat_completion = client.chat.completions.create(
+        messages=[{"role": "user", "content": f"Summarize this text:\n{text}"}],
+        model="llama-3.1-8b-instant",
+    )
+    return chat_completion.choices[0].message.content
+
+def generate_mcqs(text):
+    chat_completion = client.chat.completions.create(
+        messages=[{"role": "user", "content": f"Generate 20 MCQs with answers from:\n{text}"}],
+        model="llama-3.1-8b-instant",
+    )
+    return chat_completion.choices[0].message.content
+
+def answer_question(question):
+    """Answer user question based on stored chunks using Groq."""
+    if not global_chunks:
+        return "No data loaded yet. Please upload a document, website, or YouTube video first."
+    
+    context = " ".join(global_chunks[:10])  # simple retrieval (first 10 chunks)
+    chat_completion = client.chat.completions.create(
+        messages=[{"role": "user", "content": f"Answer the question based on context:\n{context}\n\nQuestion: {question}"}],
+        model="llama-3.1-8b-instant",
+    )
+    return chat_completion.choices[0].message.content
+
+# -------------------------------
+# Input Handlers
+# -------------------------------
+
+def process_document(file):
+    if file.name.endswith(".pdf"):
+        reader = PdfReader(file)
+        text = " ".join([page.extract_text() for page in reader.pages])
+    elif file.name.endswith(".docx"):
+        doc = docx.Document(file)
+        text = " ".join([para.text for para in doc.paragraphs])
+    elif file.name.endswith(".pptx"):
+        pres = pptx.Presentation(file)
+        text = " ".join([shape.text for slide in pres.slides for shape in slide.shapes if hasattr(shape, "text")])
+    else:
+        text = file.read().decode("utf-8")
+    return text
+
+def process_website(url):
+    response = requests.get(url)
+    soup = BeautifulSoup(response.text, "html.parser")
+    return " ".join([p.get_text() for p in soup.find_all("p")])
+
+def process_youtube(url):
+    yt = YouTube(url)
+    video_id = yt.video_id
+    transcript = YouTubeTranscriptApi.get_transcript(video_id)
+    return " ".join([t["text"] for t in transcript])
+
+# -------------------------------
+# Main Pipeline
+# -------------------------------
+
+def rag_pipeline(input_type, input_data):
+    global global_index, global_chunks
+    
+    if input_type == "Document":
+        text = process_document(input_data)
+    elif input_type == "Website":
+        text = process_website(input_data)
+    elif input_type == "YouTube":
+        text = process_youtube(input_data)
+    else:
+        return "Invalid input type", "Invalid input type"
+
+    # Chunk, embed, store
+    chunks = chunk_text(text)
+    embeddings = embed_text(chunks)
+    index = store_in_faiss(embeddings)
+
+    # Save globally for chatbot
+    global_index = index
+    global_chunks = chunks
+
+    # Summarize + MCQs
+    summary = summarize_text(text)
+    mcqs = generate_mcqs(text)
+
+    return summary, mcqs
+
+# -------------------------------
+# Gradio UI
+# -------------------------------
+
+with gr.Blocks() as demo:
+    gr.Markdown("# 📚 RAG-based Knowledge Assistant (Groq + FAISS + Gradio)")
+    
+    with gr.Tab("Upload Document"):
+        doc_input = gr.File(label="Upload PDF/DOCX/PPTX")
+        doc_output_summary = gr.Textbox(label="Summary")
+        doc_output_mcqs = gr.Textbox(label="MCQs")
+        doc_button = gr.Button("Process Document")
+        doc_button.click(rag_pipeline, inputs=["Document", doc_input], outputs=[doc_output_summary, doc_output_mcqs])
+    
+    with gr.Tab("Website"):
+        web_input = gr.Textbox(label="Enter Website URL")
+        web_output_summary = gr.Textbox(label="Summary")
+        web_output_mcqs = gr.Textbox(label="MCQs")
+        web_button = gr.Button("Process Website")
+        web_button.click(rag_pipeline, inputs=["Website", web_input], outputs=[web_output_summary, web_output_mcqs])
+    
+    with gr.Tab("YouTube"):
+        yt_input = gr.Textbox(label="Enter YouTube URL")
+        yt_output_summary = gr.Textbox(label="Summary")
+        yt_output_mcqs = gr.Textbox(label="MCQs")
+        yt_button = gr.Button("Process YouTube")
+        yt_button.click(rag_pipeline, inputs=["YouTube", yt_input], outputs=[yt_output_summary, yt_output_mcqs])
+    
+    with gr.Tab("Chatbot"):
+        chatbot_input = gr.Textbox(label="Ask a question about your uploaded content")
+        chatbot_output = gr.Textbox(label="Answer")
+        chatbot_button = gr.Button("Ask")
+        chatbot_button.click(answer_question, inputs=chatbot_input, outputs=chatbot_output)
+
+demo.launch()