Spaces:
Sleeping
Sleeping
File size: 2,661 Bytes
f648189 bd90524 f648189 f6e6c81 f648189 87fa998 dc425e6 f648189 f129d48 f648189 f129d48 f648189 42d0898 f648189 dc425e6 f129d48 dc425e6 f648189 9de5fcf f648189 9de5fcf f648189 9de5fcf 87f3e40 f648189 fb3f467 f648189 9de5fcf d943582 f648189 9de5fcf d943582 9de5fcf d943582 9de5fcf |
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 92 93 94 95 96 97 98 99 100 |
import os
import streamlit as st
import numpy as np
from sentence_transformers import SentenceTransformer
from ingest import build_index
from rag import retrieve
from groq import Groq
from config import EMBEDDING_MODEL
from pdf_utils import answer_to_pdf
# ---------------- CONFIG ----------------
st.set_page_config(page_title="OTT Bot", layout="wide")
# ---------------- LOAD MODELS ----------------
@st.cache_resource
def load_embedder():
return SentenceTransformer(EMBEDDING_MODEL)
embedder = load_embedder()
# ---------------- GROQ CLIENT ----------------
GROQ_API_KEY = os.environ.get("GROQ_API_KEY")
if not GROQ_API_KEY:
st.error("❌ GROQ_API_KEY not set in HF Secrets")
st.stop()
client = Groq(api_key=GROQ_API_KEY)
# ---------------- UI ----------------
st.title("📚 OTT Bot")
st.markdown("Upload PDFs via HF Dataset and query them using AI")
# ---------------- INGESTION ----------------
if st.button("🔨 Build Index"):
with st.spinner("Building FAISS index from dataset..."):
build_index()
st.success("✅ Index built successfully")
st.divider()
# ---------------- QUESTION INPUT ----------------
question = st.text_input(
"Ask a question about the document",
placeholder="e.g. What is the main concept discussed?"
)
# ---------------- ANSWER ----------------
if question:
with st.spinner("Searching document..."):
query_embedding = embedder.encode([question]).astype("float32")
contexts = retrieve(query_embedding)
context_text = "\n\n".join(c["text"] for c in contexts)
prompt = f"""
You are an expert tutor.
Using ONLY the information in the context below,
provide a detailed and well-structured answer.
Guidelines:
- Explain step by step
- Use headings or bullet points where useful
- Give examples
- Do NOT keep the answer short
- If information is missing, say so clearly
Context:
{context_text}
Question:
{question}
Answer clearly and concisely.
"""
with st.spinner("Thinking..."):
response = client.chat.completions.create(
model="llama-3.1-8b-instant",
messages=[
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": prompt},
],
)
answer_text = response.choices[0].message.content
st.subheader("Answer")
st.write(answer_text)
# ---------------- PDF DOWNLOAD ----------------
pdf_buffer = answer_to_pdf(question, answer_text)
st.download_button(
label="📄 Download Answer as PDF",
data=pdf_buffer,
file_name="OTT_Bot_Answer.pdf",
mime="application/pdf"
)
|