app.py

import streamlit as st
import requests
import pandas as pd
import plotly.express as px
from app.data_handling import save_results_to_csv
import os


BACKEND_URL = os.getenv("BACKEND_URL", "http://localhost:8000")  # Local backend URL if running locally.  HF Spaces need the correct backend URL

st.title("GenAI Model Evaluator By ARAVINDSELVAN")

# Input fields
prompt = st.text_area("Enter your prompt:")
llm_names = st.multiselect("Select LLMs to compare:", ["openai", "gemini"], default=["openai", "gemini"])
dataset_file = st.file_uploader("Upload a CSV dataset (optional)", type=["csv"])

# Button to run the benchmark
if st.button("Run Benchmark"):
    results = []

    if dataset_file:
        # Process dataset file
        df = pd.read_csv(dataset_file)
        if "prompt" not in df.columns:
            st.error("CSV file must have a 'prompt' column.")
            st.stop()
        prompts = df["prompt"].tolist()
    else:
        prompts = [prompt]

    for llm_name in llm_names:
        for p in prompts:
            try:
                response = requests.post(f"{BACKEND_URL}/generate", json={"prompt": p, "llm_name": llm_name})
                response.raise_for_status()  # Raise HTTPError for bad responses (4xx or 5xx)
                data = response.json()
                results.append({"llm": llm_name, "prompt": p, **data})

            except requests.exceptions.RequestException as e:
                st.error(f"Error calling backend for {llm_name}: {e}")
                continue

    if results:
        # Display results
        results_df = pd.DataFrame(results)
        st.dataframe(results_df)

        # Visualization - Response Time
        fig_time = px.bar(results_df, x="llm", y="response_time", title="Response Time Comparison")
        st.plotly_chart(fig_time)

        # Visualization - Cost Comparison
        fig_cost = px.bar(results_df, x="llm", y="cost", title="Cost Comparison")
        st.plotly_chart(fig_cost)

        if st.button("Save Results to CSV"):
            save_results_to_csv(results, "benchmark_results.csv")
            st.success("Results saved to benchmark_results.csv")
    else:
         st.warning("No results to display.")