Update app file

streamlit_app/app_streamlit.py

@@ -5,159 +5,23 @@ import pandas as pd
 import streamlit as st
 import requests
 
-# ----------------------------
-# Configuration
-# ----------------------------
-# Set this via Hugging Face Space secrets / variables or Docker env
+# -----------------------------------------------------
+# MUST COME FIRST BEFORE ANY OTHER STREAMLIT COMMAND
+# -----------------------------------------------------
+st.set_page_config(
+    page_title="ExtraaLearn – Lead Conversion",
+    page_icon="📈",
+    layout="wide"
+)
+
+# -----------------------------------------------------
+# Backend API URL (from HF Environment Variable)
+# -----------------------------------------------------
 BACKEND_URL = os.getenv("BACKEND_URL", "").rstrip("/")
-if BACKEND_URL == "":
-    st.warning(
-        "⚠️ BACKEND_URL is not set. "
-        "Set an environment variable `BACKEND_URL` to your backend Space base URL, "
-        "e.g. https://<username>-<backendspace>.hf.space"
-    )
-
-st.set_page_config(page_title="ExtraaLearn – Lead Conversion", page_icon="📈", layout="wide")
-st.title("📈 ExtraaLearn – Lead Conversion (Frontend)")
-
-with st.expander("ℹ️ Instructions", expanded=False):
-    st.markdown(
-        """
-- Set environment variable **`BACKEND_URL`** to your backend Space URL (without trailing slash).
-- Use the **Single Prediction** tab for an individual lead.
-- Use the **Batch Scoring** tab to upload a CSV and get predictions back.
-        """
-    )
-
-# ----------------------------
-# Utilities
-# ----------------------------
-def ping_backend():
-    if not BACKEND_URL:
-        return False, {"error": "BACKEND_URL not set"}
-    try:
-        r = requests.get(f"{BACKEND_URL}/", timeout=10)
-        return r.ok, r.json()
-    except Exception as e:
-        return False, {"error": str(e)}
-
-def post_json(endpoint: str, payload):
-    url = f"{BACKEND_URL}{endpoint}"
-    r = requests.post(url, json=payload, timeout=60)
-    r.raise_for_status()
-    return r
-
-# ----------------------------
-# Health check
-# ----------------------------
-colA, colB = st.columns([1, 2])
-with colA:
-    if st.button("🔁 Check Backend Health"):
-        ok, info = ping_backend()
-        if ok:
-            st.success(f"Backend OK · model_loaded={info.get('model_loaded')}")
-        else:
-            st.error(f"Backend unreachable: {info}")
-
-# ----------------------------
-# Tabs
-# ----------------------------
-tab1, tab2 = st.tabs(["Single Prediction", "Batch Scoring"])
-
-# ---- Tab 1: Single ----
-with tab1:
-    st.subheader("Single Prediction")
-    c1, c2, c3 = st.columns(3)
 
-    with c1:
-        age = st.number_input("Age", min_value=10, max_value=100, value=30)
-        website_visits = st.number_input("Website Visits", min_value=0, max_value=500, value=3)
-        time_spent_on_website = st.number_input("Time Spent on Website (sec)", min_value=0, max_value=200000, value=300)
-        page_views_per_visit = st.number_input("Page Views per Visit", min_value=0.0, max_value=100.0, value=3.5, step=0.1)
-
-    with c2:
-        current_occupation = st.selectbox("Current Occupation", ["Professional", "Student", "Unemployed"])
-        first_interaction = st.selectbox("First Interaction", ["Website", "Mobile App"])
-        last_activity = st.selectbox("Last Activity", ["Email Activity", "Phone Activity", "Website Activity"])
-        profile_completed = st.selectbox(
-            "Profile Completed",
-            ["Low (0-50%)", "Medium (50-75%)", "High (75-100%)"]
-        )
-
-    with c3:
-        print_media_type1 = st.selectbox("Saw Newspaper Ad (print_media_type1)", ["No", "Yes"])
-        print_media_type2 = st.selectbox("Saw Magazine Ad (print_media_type2)", ["No", "Yes"])
-        digital_media = st.selectbox("Saw Digital Ad", ["No", "Yes"])
-        educational_channels = st.selectbox("Heard via Educational Channels", ["No", "Yes"])
-        referral = st.selectbox("Referred by Someone", ["No", "Yes"])
-
-    if st.button("🚀 Predict"):
-        if not BACKEND_URL:
-            st.error("Please set BACKEND_URL to your backend Space URL.")
-        else:
-            payload = {
-                "age": age,
-                "website_visits": website_visits,
-                "time_spent_on_website": time_spent_on_website,
-                "page_views_per_visit": page_views_per_visit,
-                "print_media_type1": print_media_type1,
-                "print_media_type2": print_media_type2,
-                "digital_media": digital_media,
-                "educational_channels": educational_channels,
-                "referral": referral,
-                "current_occupation": current_occupation,
-                "first_interaction": first_interaction,
-                "last_activity": last_activity,
-                "profile_completed": profile_completed
-            }
-            try:
-                start = time.time()
-                r = post_json("/predict", payload)
-                elapsed = time.time() - start
-                result = r.json()
-                st.success(f"✅ Inference OK · {elapsed:.2f}s")
-                st.metric("Conversion Probability", f"{result.get('probability', 0.0):.3f}")
-                pred = result.get("prediction", 0)
-                st.write("Prediction:", "✅ Will Convert" if pred == 1 else "❌ Unlikely to Convert")
-                st.code(json.dumps(payload, indent=2), language="json")
-            except requests.HTTPError as http_err:
-                st.error(f"HTTP error: {http_err} · {r.text}")
-            except Exception as e:
-                st.error(f"Request failed: {e}")
-
-# ---- Tab 2: Batch ----
-with tab2:
-    st.subheader("Batch Scoring (CSV)")
-    st.caption("The CSV must include the same columns used during model training.")
-
-    sample_schema = [
-        "age","website_visits","time_spent_on_website","page_views_per_visit",
-        "print_media_type1","print_media_type2","digital_media","educational_channels","referral",
-        "current_occupation","first_interaction","last_activity","profile_completed"
-    ]
-    st.text_area("Expected columns", ", ".join(sample_schema), height=70, disabled=True)
-
-    up = st.file_uploader("Upload CSV", type=["csv"])
-    if up is not None:
-        try:
-            df = pd.read_csv(up)
-            st.write("Preview:", df.head())
-            if st.button("📤 Send to Backend"):
-                if not BACKEND_URL:
-                    st.error("Please set BACKEND_URL to your backend Space URL.")
-                else:
-                    records = df.to_dict(orient="records")
-                    start = time.time()
-                    r = post_json("/predict-batch", {"records": records})
-                    elapsed = time.time() - start
-                    out = pd.DataFrame(r.json())
-                    st.success(f"✅ Batch inference OK · {elapsed:.2f}s")
-                    st.dataframe(out.head(20), use_container_width=True)
-                    st.download_button(
-                        "⬇️ Download Results",
-                        out.to_csv(index=False).encode("utf-8"),
-                        file_name="predictions.csv",
-                        mime="text/csv"
-                    )
-        except Exception as e:
-            st.error(f"Failed to process CSV: {e}")
+if not BACKEND_URL:
+    st.warning(
+        "⚠️ BACKEND_URL is not set. Set an environment variable BACKEND_URL "
+        "to your backend Space URL (without trailing slash). "
+        "Example: https://<username>-<backendspace>.hf.space"
+    )