Update app.py

app.py

@@ -1,39 +1,51 @@
-"""app.py — BizIntel AI Ultra (Gemini‑only, v4)
-A production‑grade BI copilot with:
- • CSV / Excel / Parquet and SQL ingestion
- • Smart sampling + memory‑safe loading for large files
- • Schema + missing-data audit with Gemini-generated insights
- • Drill-down EDA (histogram, violin, scatter-matrix, heatmap)
- • Auto‑detected date column, tunable ARIMA forecasting
- • One-click strategy brief download (Markdown)
+"""app.py — BizIntel AI Ultra (Gemini‑only, v4)
+Robust BI copilot featuring:
+• CSV / Excel / Parquet uploads + live SQL ingestion
+• Memory‑safe sampling for large files (≤ 5 M rows sample)
+• Gemini‑generated narrative insights & quick schema audit
+• Interactive EDA (histogram + box, correlation heatmap)
+• Tunable ARIMA forecasting (p,d,q & horizon)
+• One‑click strategy brief download
 """
-
 from __future__ import annotations
-import os, io, tempfile
+import io, os, tempfile
 from pathlib import Path
 from typing import List
 
 import pandas as pd
-import streamlit as st
 import plotly.express as px
-from statsmodels.tsa.arima.model import ARIMA
+import streamlit as st
 from sqlalchemy import create_engine
+from statsmodels.tsa.arima.model import ARIMA
 import google.generativeai as genai
 
-# ───────────────────── 0 · CONFIGURATION ─────────────────────
-API_KEY = st.secrets.get("GEMINI_APIKEY") or os.getenv("GEMINI_APIKEY")
-if not API_KEY:
-    st.error("❌ Missing `GEMINI_APIKEY` — add it in Settings → Secrets or set env variable.")
+# ───────────────────── 0 · PAGE CONFIG ──────────────────────
+st.set_page_config(
+    page_title="BizIntel AI Ultra",
+    page_icon="📊",
+    layout="wide",
+    initial_sidebar_state="expanded",
+)
+
+auth_key = None
+try:
+    auth_key = st.secrets["GEMINI_APIKEY"]  # type: ignore[attr-defined]
+except Exception:
+    auth_key = os.getenv("GEMINI_APIKEY")
+
+if not auth_key:
+    st.error(
+        "❌ **GEMINI_APIKEY** missing. Add it in *Settings → Secrets* or set env var before running."
+    )
     st.stop()
 
-st.set_page_config("BizIntel AI Ultra", "📊", "wide", initial_sidebar_state="expanded")
-genai.configure(api_key=API_KEY)
+genai.configure(api_key=auth_key)
 GEM_MODEL = "gemini-1.5-pro-latest"
 TMP = Path(tempfile.gettempdir())
 
-# ───────────────────── 1 · UTILITY HELPERS ───────────────────
+# ───────────────────── 1 · HELPERS ──────────────────────────
 @st.cache_data(show_spinner=False)
-def read_file(buf: io.BufferedReader, sample: bool = False) -> pd.DataFrame:
+def load_file(buf: io.BufferedReader, sample: bool = False) -> pd.DataFrame:
     suf = Path(buf.name).suffix.lower()
     if suf in {".xls", ".xlsx"}:
         return pd.read_excel(buf, engine="openpyxl")
@@ -42,105 +54,99 @@ def read_file(buf: io.BufferedReader, sample: bool = False) -> pd.DataFrame:
     return pd.read_csv(buf, nrows=5_000_000 if sample else None)
 
 @st.cache_data(show_spinner=False)
-def sql_tables(uri: str) -> List[str]:
+def list_sql_tables(uri: str) -> List[str]:
     return create_engine(uri).table_names()
 
 @st.cache_data(show_spinner=True)
-def read_table(uri: str, tbl: str) -> pd.DataFrame:
+def fetch_sql_table(uri: str, tbl: str) -> pd.DataFrame:
     return pd.read_sql_table(tbl, create_engine(uri))
 
 @st.cache_data(show_spinner=False)
-def ask_gemini(prompt: str) -> str:
+def gemini(prompt: str) -> str:
     return genai.GenerativeModel(GEM_MODEL).generate_content(prompt).text.strip()
 
-# ───────────────────── 2 · DATA INGESTION ────────────────────
-st.title("📊 BizIntel AI Ultra — Gemini 1.5 Pro BI Copilot")
-mode = st.sidebar.radio("Select Data Source", ["Upload File", "SQL Database"], horizontal=True)
-df: pd.DataFrame = pd.DataFrame()
+# ───────────────────── 2 · DATA LOAD UI ─────────────────────
+st.title("📊 BizIntel AI Ultra — Gemini 1.5 Pro BI Copilot")
+mode = st.sidebar.radio("Data source", ["Upload file", "SQL DB"], horizontal=True)
 
-if mode == "Upload File":
-    upl = st.sidebar.file_uploader("Upload CSV / Excel / Parquet", ["csv", "xls", "xlsx", "parquet"], help="≤2 GB")
-    sample = st.sidebar.checkbox("Load sample (≤ 5M rows)")
+df: pd.DataFrame = pd.DataFrame()
+if mode == "Upload file":
+    upl = st.sidebar.file_uploader("CSV / Excel / Parquet", ["csv", "xls", "xlsx", "parquet"], help="≤ 2 GB")
+    sample = st.sidebar.checkbox("Sample first 5 M rows only")
     if upl:
-        df = read_file(upl, sample)
+        df = load_file(upl, sample)
 else:
     uri = st.sidebar.text_input("SQLAlchemy URI")
     if uri:
-        tbl = st.sidebar.selectbox("Choose Table", sql_tables(uri))
+        tbl = st.sidebar.selectbox("Table", list_sql_tables(uri))
         if tbl:
-            df = read_table(uri, tbl)
+            df = fetch_sql_table(uri, tbl)
 
 if df.empty:
-    st.info("⬅️ Load a dataset to get started.")
+    st.info("⬅️ Load data to start analysis.")
     st.stop()
 
-st.success("✅ Data loaded")
+st.success("✅ Data loaded")
 st.dataframe(df.head(), use_container_width=True)
 
-# ───────────────────── 3 · SUMMARY + GEMINI ───────────────────
+# ───────────────────── 3 · OVERVIEW & GEMINI INSIGHT ─────────
 rows, cols = df.shape
-miss_pct = df.isna().sum().sum() / (rows * cols) * 100
-c1, c2, c3 = st.columns(3)
-c1.metric("Rows", f"{rows:,}")
-c2.metric("Columns", cols)
-c3.metric("Missing %", f"{miss_pct:.1f}")
-
-st.subheader("🧠 Gemini Insights")
-with st.spinner("Generating analysis..."):
-    summary = df.describe(include="all", datetime_is_numeric=True).round(2).to_json()
-    st.markdown(ask_gemini(
-        "You are a senior BI analyst. List 5 key insights and 3 action items based on this dataset: " + summary
+missing_pct = df.isna().sum().sum() / (rows * cols) * 100
+m1,m2,m3 = st.columns(3)
+m1.metric("Rows", f"{rows:,}")
+m2.metric("Columns", cols)
+m3.metric("Missing %", f"{missing_pct:.1f}")
+
+st.subheader("🧠 Gemini Insights")
+with st.spinner("Generating narrative…"):
+    summ_json = df.describe(include="all", datetime_is_numeric=True).round(2).to_json()
+    st.markdown(gemini(
+        "You are a senior BI analyst. Provide five bullet insights and three actions for this dataset:\n\n" + summ_json
     ))
 
-# ───────────────────── 4 · TIME SERIES SETUP ─────────────────
-# try datetime coercion
-for c in df.columns:
-    if not pd.api.types.is_datetime64_any_dtype(df[c]):
+# ───────────────────── 4 · TIME‑SERIES PREP ──────────────────
+# attempt datetime coercion
+for col in df.columns:
+    if not pd.api.types.is_datetime64_any_dtype(df[col]):
         try:
-            df[c] = pd.to_datetime(df[c])
+            df[col] = pd.to_datetime(df[col])
         except: pass
 
 DATE_COL = st.selectbox("Date column", [c for c in df.columns if pd.api.types.is_datetime64_any_dtype(df[c])])
-METRIC_COL = st.selectbox("Numeric metric", [c for c in df.select_dtypes("number").columns])
-
-series = (
-    df[[DATE_COL, METRIC_COL]].dropna()
-      .groupby(DATE_COL)[METRIC_COL].mean().sort_index()
-)
-fig_ts = px.line(series, title=f"{METRIC_COL} Trend", labels={"index": "Date", METRIC_COL: METRIC_COL})
-st.plotly_chart(fig_ts, use_container_width=True)
-
-# ───────────────────── 5 · ARIMA FORECASTING ─────────────────
-st.subheader("🔮 Forecast")
-steps = st.slider("Forecast Horizon", 3, 365, 90)
-p = st.number_input("AR Order (p)", 0, 5, 1)
-d = st.number_input("Diff Order (d)", 0, 2, 1)
-q = st.number_input("MA Order (q)", 0, 5, 1)
-
-with st.spinner("Training ARIMA model..."):
-    model = ARIMA(series, order=(p, d, q)).fit()
-    fut_idx = pd.date_range(series.index[-1], periods=steps + 1, freq=pd.infer_freq(series.index) or "D")[1:]
-    forecast = pd.Series(model.forecast(steps), index=fut_idx)
-fig_fc = px.line(pd.concat([series, forecast.rename("Forecast")], axis=1), title="Actual vs Forecast")
-st.plotly_chart(fig_fc, use_container_width=True)
-
-# ───────────────────── 6 · EDA TOOLS ─────────────────────────
-st.subheader("🔍 Exploratory Data Dashboard")
+NUM_COL = st.selectbox("Numeric metric", [c for c in df.select_dtypes("number").columns])
+
+ts = df[[DATE_COL, NUM_COL]].dropna().groupby(DATE_COL)[NUM_COL].mean().sort_index()
+fig_tr = px.line(ts, title=f"{NUM_COL} Trend", labels={"index":"Date", NUM_COL:NUM_COL})
+st.plotly_chart(fig_tr, use_container_width=True)
+
+# ───────────────────── 5 · ARIMA FORECAST ───────────────────
+st.subheader("🔮 ARIMA Forecast")
+steps = st.slider("Horizon (days)", 3, 365, 90)
+p = st.number_input("p", 0, 5, 1); d = st.number_input("d", 0, 2, 1); q = st.number_input("q", 0, 5, 1)
+try:
+    model = ARIMA(ts, order=(p, d, q)).fit()
+    future_idx = pd.date_range(ts.index[-1], periods=steps + 1, freq=pd.infer_freq(ts.index) or "D")[1:]
+    forecast = pd.Series(model.forecast(steps), index=future_idx, name="Forecast")
+    fig_fc = px.line(pd.concat([ts, forecast], axis=1), title="Actual vs Forecast")
+    st.plotly_chart(fig_fc, use_container_width=True)
+except Exception as e:
+    st.error(f"ARIMA failed: {e}")
+
+# ──────���────────────── 6 · EDA EXPANDERS ────────────────────
+st.subheader("🔍 EDA Dashboard")
 with st.expander("Histogram + Box"):
-    col = st.selectbox("Metric column", METRIC_COL, key="hist")
-    st.plotly_chart(px.histogram(df, x=col, marginal="box", template="plotly_dark"), use_container_width=True)
-
+    st.plotly_chart(px.histogram(df, x=NUM_COL, marginal="box", template="plotly_dark"), use_container_width=True)
 with st.expander("Correlation Heatmap"):
     corr = df.select_dtypes("number").corr()
-    st.plotly_chart(px.imshow(corr, color_continuous_scale="RdBu", aspect="auto", title="Correlation Matrix"), use_container_width=True)
+    st.plotly_chart(px.imshow(corr, color_continuous_scale="RdBu", title="Correlation"), use_container_width=True)
 
-# ───────────────────── 7 · STRATEGY DOWNLOAD ────────────────
+# ───────────────────── 7 · STRATEGY BRIEF ───────────────────
 brief = (
     "# Strategy Brief\n"
-    "* Clean missing date values for better time modeling.\n"
-    "* Investigate top correlations for potential drivers.\n"
-    "* Leverage forecast for inventory and staff planning.\n"
-    "* Watch for outliers >3σ weekly.\n"
-    "* Segment by region and product for precise actions."
+    "• Clean missing date values for stable models.\n"
+    "• Investigate high correlations for driver analysis.\n"
+    "• Use forecast for inventory & workforce planning.\n"
+    "• Monitor outliers (>3σ) weekly.\n"
+    "• Segment by region/product for targeted actions."
 )
-st.download_button("⬇️ Download Strategy (.md)", brief, "bizintel_brief.md", "text/markdown")
+st.download_button("⬇️ Download Strategy (.md)", brief, "bizintel_brief.md", "text/markdown")