Spaces:
Sleeping
Sleeping
File size: 3,461 Bytes
8498ff5 fe6177f 8498ff5 76c80b3 8498ff5 022f5f2 8498ff5 | 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 | # streamlit_app.py
import os
import time
import streamlit as st
from datetime import datetime
from dotenv import load_dotenv
load_dotenv(dotenv_path=os.path.join(os.path.dirname(__file__), '..', '.env'))
os.environ['HF_HOME'] = '/app/model_cache'
from reddit_client import search_reddit
from embedding import get_embeddings
from llm_analysis import get_llm_pipelines, analyze_post
from vector_search import build_faiss_index, search_similar
st.set_page_config(page_title="BrandSight AI - Reddit Brand Monitor", layout="wide", page_icon="🔍")
st.title("🔍 Proactive Brand Intelligence Monitor")
st.caption(f"🕒 Last checked: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
# Load models once
with st.spinner("Loading AI models... (first time only)"):
classifier, sentiment_pipe, summarizer = get_llm_pipelines()
brand = st.text_input("Enter a brand or keyword to search on Reddit (e.g., 'Nvidia')")
if brand:
if 'results' not in st.session_state or st.session_state.get('brand') != brand:
st.session_state.brand = brand
with st.spinner(f"🔍 Fetching Reddit posts for **{brand}**..."):
t0 = time.time()
posts = search_reddit(brand, limit=10)
st.write(f"⏱️ Reddit fetch time: {round(time.time() - t0, 2)} seconds")
if not posts:
st.warning("❌ No posts found. Try a different keyword.")
st.stop()
texts = [p.title + " " + (p.selftext or "") for p in posts]
results = []
with st.spinner("🤖 Analyzing Reddit posts with AI..."):
t1 = time.time()
for text in texts:
results.append(analyze_post(text, classifier, sentiment_pipe, summarizer))
st.write(f"⏱️ AI Analysis time: {round(time.time() - t1, 2)} seconds")
st.session_state.posts = posts
st.session_state.results = results
st.session_state.embeddings = get_embeddings(texts)
st.session_state.index = build_faiss_index(st.session_state.embeddings)
st.header(f"📟 Results for: {st.session_state.brand}")
categories = ["Bug Report", "Feature Request", "Competitor Mention", "Positive Feedback"]
category_filter = st.multiselect("Filter by category:", categories, default=categories)
for i, post in enumerate(st.session_state.posts):
result = st.session_state.results[i]
if result["category"] in category_filter:
with st.expander(f"{post.title} (r/{post.subreddit.display_name})"):
st.markdown(f"**Category:** {result['category']} | **Sentiment:** {result['sentiment']}")
st.markdown(f"**Summary:** *{result['summary']}*")
st.write(f"🔗 [View on Reddit](https://reddit.com{post.permalink})")
st.header("🔎 Find Similar Posts")
query = st.text_input("Search similar topics (e.g., 'overheating issue')")
if query:
q_emb = get_embeddings([query])
distances, indices = search_similar(st.session_state.index, q_emb)
st.subheader("Top 5 similar Reddit posts:")
for rank, idx in enumerate(indices[0]):
sim_post = st.session_state.posts[idx]
sim_res = st.session_state.results[idx]
st.markdown(f"**{rank + 1}. {sim_post.title}** (r/{sim_post.subreddit.display_name})")
st.markdown(f"> *{sim_res['summary']}*")
st.write("---")
else:
st.info("Type a brand or keyword above to begin.")
|