app.py

import os
import tempfile
import pandas as pd
import streamlit as st
import google.generativeai as genai
import plotly.graph_objects as go

from tools.csv_parser import parse_csv_tool
from tools.plot_generator import plot_sales_tool
from tools.forecaster import forecast_tool
from tools.visuals import (
    histogram_tool,
    scatter_matrix_tool,
    corr_heatmap_tool,
)

# ── Gemini 1.5‑Pro configuration ─────────────────────────────
genai.configure(api_key=os.getenv("GEMINI_APIKEY"))
gemini = genai.GenerativeModel(
    "gemini-1.5-pro-latest",
    generation_config={
        "temperature": 0.7,
        "top_p": 0.9,
        "response_mime_type": "text/plain",
    },
)

# ── Streamlit page setup ─────────────────────────────────────
st.set_page_config(page_title="BizIntel AI Ultra – Gemini 1.5 Pro", layout="wide")
st.title("📊 BizIntel AI Ultra – Advanced Analytics")

TEMP_DIR = tempfile.gettempdir()

# ── CSV upload ───────────────────────────────────────────────
csv_file = st.file_uploader("Upload CSV (≤ 200 MB)", type=["csv"])
if csv_file is None:
    st.info("⬆️ Upload a CSV to begin.")
    st.stop()

csv_path = os.path.join(TEMP_DIR, csv_file.name)
with open(csv_path, "wb") as f:
    f.write(csv_file.read())
st.success("CSV saved ✅")

# Preview + date column selection
df_preview = pd.read_csv(csv_path, nrows=5)
st.dataframe(df_preview)
date_col = st.selectbox("Select date/time column for forecasting", df_preview.columns)

# ── Local tools: summary, sales trend, forecast ──────────────
with st.spinner("Parsing CSV…"):
    summary_text = parse_csv_tool(csv_path)

with st.spinner("Generating sales trend chart…"):
    sales_fig = plot_sales_tool(csv_path, date_col=date_col)
    st.plotly_chart(sales_fig, use_container_width=True)

with st.spinner("Forecasting future metrics…"):
    forecast_text = forecast_tool(csv_path, date_col=date_col)
    if os.path.exists("forecast_plot.png"):
        forecast_img = "forecast_plot.png"
    else:
        forecast_img = None

# ── Gemini strategy insights ─────────────────────────────────
prompt = (
    f"You are **BizIntel Strategist AI**.\n\n"
    f"### CSV Summary\n```\n{summary_text}\n```\n\n"
    f"### Forecast Output\n```\n{forecast_text}\n```\n\n"
    "Return Markdown with:\n"
    "1. **Five key insights** (bullet list)\n"
    "2. **Three actionable strategies** (with expected impact)\n"
    "3. **Risk factors or anomalies**\n"
    "4. **Suggested additional visuals**\n"
)

st.subheader("🚀 Strategy Recommendations (Gemini 1.5 Pro)")
with st.spinner("Gemini 1.5 Pro is generating insights…"):
    strategy_md = gemini.generate_content(prompt).text

st.markdown(strategy_md)

# Display forecast image if exists
if forecast_img:
    st.image(forecast_img, caption="Sales Forecast", use_column_width=True)

# ── Optional exploratory visuals ─────────────────────────────
st.markdown("---")
st.subheader("🔍 Optional Exploratory Visuals")

num_cols = df_preview.select_dtypes("number").columns

# Histogram
if st.checkbox("Show histogram"):
    hist_col = st.selectbox("Histogram variable", num_cols, key="hist")
    fig_hist = histogram_tool(csv_path, hist_col)
    st.plotly_chart(fig_hist, use_container_width=True)

# Scatter‑matrix
if st.checkbox("Show scatter‑matrix"):
    multi_cols = st.multiselect("Choose up to 5 columns", num_cols, default=num_cols[:3])
    if multi_cols:
        fig_scatter = scatter_matrix_tool(csv_path, multi_cols)
        st.plotly_chart(fig_scatter, use_container_width=True)

# Correlation heat‑map
if st.checkbox("Show correlation heat‑map"):
    fig_corr = corr_heatmap_tool(csv_path)
    st.plotly_chart(fig_corr, use_container_width=True)

# ── CSV summary text at bottom ───────────────────────────────
st.markdown("---")
st.subheader("📑 CSV Summary (full Stats)")
st.text(summary_text)