app.py

import streamlit as st
import requests
import html



st.set_page_config(
    page_title="VectorVault – Semantic Search Engine",
    layout="wide",
    page_icon="🔍"
)


st.markdown("""
<h1 style="text-align:center; color:#4A90E2;">
🔍 VectorVault – Semantic Document Search
</h1>
<p style="text-align:center; font-size:18px;">
Search across your documents using embeddings, FAISS, and semantic similarity.
</p>
""", unsafe_allow_html=True)


st.sidebar.header("⚙️ Settings")

top_k = st.sidebar.slider(
    "Number of results",
    min_value=1, 
    max_value=20, 
    value=5
)

show_preview = st.sidebar.checkbox("Show preview text", value=True)
show_keywords = st.sidebar.checkbox("Show matched keywords", value=True)
show_similarity_bar = st.sidebar.checkbox("Show similarity bars", value=True)

st.sidebar.markdown("---")
st.sidebar.caption("Built by Ash Nicolus")


query = st.text_input(
    "Enter your search query",
    placeholder="e.g., machine learning basics, quantum physics, climate change..."
)

search_button = st.button("Search 🔎")



def highlight_keywords(text, keywords):
    for kw in keywords:
        text = text.replace(kw, f"<mark style='background-color:#FFF59D'>{kw}</mark>")
    return text

if search_button:
    if not query.strip():
        st.warning("⚠️ Please enter a query to search.")
    else:
        payload = {"query": query, "top_k": top_k}

        try:
            response = requests.post("http://localhost:8000/search", json=payload)
            data = response.json()

            results = data.get("results", [])

            if len(results) == 0:
                st.error("No results found!")
            else:
                st.success(f"Found {len(results)} matching documents.")

            # ----------------------------
            # DISPLAY RESULTS BEAUTIFULLY
            # ----------------------------

            for item in results:
                doc_id = item["doc_id"]
                score = round(item["score"], 4)
                preview = html.escape(item["preview"])
                explanation = item["explanation"]

                overlap_words = explanation["overlap_keywords"]
                overlap_ratio = round(explanation["overlap_ratio"], 3)

                # CARD LAYOUT
                with st.container():
                    st.markdown(f"""
                    <div style="padding:20px; margin-bottom:20px; border-radius:10px; background-color:#F9FAFB; border:1px solid #E5E7EB;">
                        <h3 style="margin-bottom:5px;">📄 {doc_id}</h3>
                        <p style="color:#6B7280; margin-top:-5px;">
                            Similarity Score: <b>{score}</b>
                        </p>
                    </div>
                    """, unsafe_allow_html=True)

                    # SIMILARITY BAR
                    if show_similarity_bar:
                        st.progress(min(max(score, 0.0), 1.0))

                    # PREVIEW TEXT
                    if show_preview:
                        st.markdown("**Preview**")
                        if show_keywords:
                            highlighted = highlight_keywords(preview, overlap_words)
                            st.markdown(f"<p>{highlighted}</p>", unsafe_allow_html=True)
                        else:
                            st.write(preview)

                    # EXPLANATION BLOCK
                    with st.expander("🧠 Why this document was matched?"):
                        st.write(f"**Matched Keywords:** {overlap_words}")
                        st.write(f"**Overlap Ratio:** {overlap_ratio}")

                st.markdown("---")

        except Exception as e:
            st.error(f"❌ Error communicating with API: {str(e)}")