app.py

import streamlit as st
import pandas as pd
from rdflib import Graph, Namespace
import os

# Page Config
st.set_page_config(page_title="Cars Knowledge Graph Demo", layout="wide")

# Load Graph
@st.cache_resource
def load_graph():
    g = Graph()
    graph_path = "cars_knowledge_graph.ttl"
    if os.path.exists(graph_path):
        g.parse(graph_path, format="turtle")
    return g

try:
    g = load_graph()
except Exception as e:
    st.error(f"Failed to load graph: {e}")
    st.stop()

if len(g) == 0:
    st.warning("Graph is empty or not found. Please run 'src/convert_data.py' first.")
    st.stop()

# Namespaces
EX = Namespace("http://example.org/cars/")

# Sidebar Filters
st.sidebar.header("Filter Cars")

# 1. Manufacturer Filter
manu_query = """
    PREFIX ex: <http://example.org/cars/>
    PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
    SELECT DISTINCT ?name WHERE { 
        ?m a ex:Manufacturer ; rdfs:label ?name . 
    } ORDER BY ?name
"""
manufacturers = ["All"] + [str(row.name) for row in g.query(manu_query)]
selected_manu = st.sidebar.selectbox("Manufacturer", manufacturers)

# 2. Price Range
price_query = """
    PREFIX ex: <http://example.org/cars/>
    SELECT (MIN(?p) as ?min) (MAX(?p) as ?max) WHERE { ?s ex:hasPriceValue ?p }
"""
price_res = list(g.query(price_query))[0]
min_price, max_price = float(price_res.min), float(price_res.max)
selected_price = st.sidebar.slider("Max Price (USD)", min_price, max_price, max_price)

# 3. Min Horsepower
hp_query = """
    PREFIX ex: <http://example.org/cars/>
    SELECT (MIN(?hp) as ?min) (MAX(?hp) as ?max) WHERE { ?s ex:hasHorsePowerValue ?hp }
"""
hp_res = list(g.query(hp_query))[0]
min_hp, max_hp = int(hp_res.min), int(hp_res.max)
selected_hp = st.sidebar.slider("Min Horsepower", min_hp, max_hp, min_hp)

# Main Area
st.title("🚗 Cars Knowledge Graph Explorer")
st.markdown("This application queries the RDF Knowledge Graph directly using **SPARQL**.")

# Construct Query based on filters
sparql_query = f"""
    PREFIX ex: <http://example.org/cars/>
    PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
    
    SELECT ?carName ?manuName ?price ?hp ?topSpeed ?seats
    WHERE {{
        ?car a ex:Car ;
             rdfs:label ?carName ;
             ex:hasManufacturer ?manu ;
             ex:hasPriceValue ?price ;
             ex:hasHorsePowerValue ?hp ;
             ex:hasTopSpeedKMH ?topSpeed ;
             ex:hasSeatCount ?seats .
        
        ?manu rdfs:label ?manuName .
        
        FILTER (?price <= {selected_price})
        FILTER (?hp >= {selected_hp})
        {f'FILTER (?manuName = "{selected_manu}")' if selected_manu != "All" else ""}
    }}
    ORDER BY DESC(?price)
    LIMIT 100
"""

# Run Query
results = g.query(sparql_query)

# Display Results
data = []
for row in results:
    data.append({
        "Car Model": str(row.carName),
        "Manufacturer": str(row.manuName),
        "Price ($)": f"${float(row.price):,.2f}",
        "Horsepower": int(row.hp),
        "Top Speed (km/h)": int(row.topSpeed),
        "Seats": int(row.seats)
    })

df = pd.DataFrame(data)

col1, col2, col3 = st.columns(3)
col1.metric("Total Cars Found", len(df))
col2.metric("Graph Triples", len(g))
col3.metric("Selected Manufacturer", selected_manu)

if not df.empty:
    st.dataframe(df, use_container_width=True)
else:
    st.info("No cars match your filters.")

# Advanced: Raw SPARQL
with st.expander("Run Custom SPARQL Query"):
    custom_query = st.text_area("SPARQL Query", """
        PREFIX ex: <http://example.org/cars/>
        SELECT ?name ?price WHERE {
            ?c ex:hasPriceValue ?price ;
               rdfs:label ?name .
        } LIMIT 5
    """)
    if st.button("Run Query"):
        try:
            raw_res = g.query(custom_query)
            st.write(list(raw_res))
        except Exception as e:
            st.error(f"Error: {e}")