Upload folder using huggingface_hub

Dockerfile

@@ -1,20 +1,23 @@
-FROM python:3.13.5-slim
+# Use a minimal base image with Python 3.9 installed
+FROM python:3.11
 
+# Set the working directory inside the container to /app
 WORKDIR /app
 
-RUN apt-get update && apt-get install -y \
-    build-essential \
-    curl \
-    git \
-    && rm -rf /var/lib/apt/lists/*
-
-COPY requirements.txt ./
-COPY src/ ./src/
+# Copy all files from the current directory on the host to the container's /app directory
+COPY . .
 
+# Install Python dependencies listed in requirements.txt
 RUN pip3 install -r requirements.txt
 
-EXPOSE 8501
+RUN useradd -m -u 1000 user
+USER user
+ENV HOME=/home/user \
+	PATH=/home/user/.local/bin:$PATH
+
+WORKDIR $HOME/app
 
-HEALTHCHECK CMD curl --fail http://localhost:8501/_stcore/health
+COPY --chown=user . $HOME/app
 
-ENTRYPOINT ["streamlit", "run", "src/streamlit_app.py", "--server.port=8501", "--server.address=0.0.0.0"]
+# Define the command to run the Streamlit app on port "8501" and make it accessible externally
+CMD ["streamlit", "run", "app.py", "--server.port=8501", "--server.address=0.0.0.0", "--server.enableXsrfProtection=false"]

app.py

@@ -0,0 +1,166 @@
+
+import os
+import streamlit as st
+import pandas as pd
+import joblib
+from huggingface_hub import hf_hub_download
+
+# Config
+REPO_ID = "surnellas/Visit-With-Us"
+MODEL_FILENAME = "best_tourism_model_v1.joblib"
+DATA_FILENAME = "tourism.csv"
+CLASSIFICATION_THRESHOLD = 0.45
+
+st.title("Visit-With-Us — Wellness Package Purchase Prediction")
+st.write(
+    "Enter customer details below. The model predicts the probability that the customer "
+    "will purchase the Wellness Tourism Package."
+)
+
+# Feature lists (used by the model)
+numeric_features = [
+    "Age",
+    "CityTier",
+    "DurationOfPitch",
+    "NumberOfPersonVisiting",
+    "NumberOfFollowups",
+    "PreferredPropertyStar",
+    "NumberOfTrips",
+    "Passport",
+    "PitchSatisfactionScore",
+    "OwnCar",
+    "NumberOfChildrenVisiting",
+    "MonthlyIncome",
+]
+
+categorical_features = [
+    "TypeofContact",
+    "Occupation",
+    "Gender",
+    "ProductPitched",
+    "MaritalStatus",
+    "Designation",
+]
+
+# Try to download dataset from HF to extract sensible options and ranges
+defaults = {}
+options = {}
+try:
+    local_data = hf_hub_download(repo_id=REPO_ID, repo_type="dataset", filename=DATA_FILENAME, token=os.environ.get("HF_TOKEN"))
+    template_df = pd.read_csv(local_data)
+    # Convert object columns to category for safer unique values
+    for c in categorical_features:
+        if c in template_df.columns:
+            options[c] = sorted(template_df[c].astype(str).unique().tolist())
+    for n in numeric_features:
+        if n in template_df.columns:
+            defaults[n] = {
+                "min": int(template_df[n].min()),
+                "max": int(template_df[n].max()),
+                "mean": float(template_df[n].median()),
+            }
+except Exception:
+    # Fallback defaults if we cannot download dataset
+    options = {
+        "TypeofContact": ["Company Invited", "Self Enquiry"],
+        "Occupation": ["Salaried", "Small Business", "Free Lancer", "Other"],
+        "Gender": ["Male", "Female"],
+        "ProductPitched": ["Basic", "Standard", "Deluxe", "Super Deluxe", "King"],
+        "MaritalStatus": ["Single", "Married", "Divorced", "Unmarried"],
+        "Designation": ["Executive", "Manager", "Senior Manager", "AVP", "VP"],
+    }
+    defaults = {
+        "Age": {"min": 18, "max": 80, "mean": 35},
+        "CityTier": {"min": 1, "max": 3, "mean": 2},
+        "DurationOfPitch": {"min": 1, "max": 60, "mean": 10},
+        "NumberOfPersonVisiting": {"min": 1, "max": 10, "mean": 3},
+        "NumberOfFollowups": {"min": 0, "max": 12, "mean": 3},
+        "PreferredPropertyStar": {"min": 1, "max": 5, "mean": 3},
+        "NumberOfTrips": {"min": 0, "max": 20, "mean": 2},
+        "Passport": {"min": 0, "max": 1, "mean": 1},
+        "PitchSatisfactionScore": {"min": 1, "max": 5, "mean": 3},
+        "OwnCar": {"min": 0, "max": 1, "mean": 1},
+        "NumberOfChildrenVisiting": {"min": 0, "max": 5, "mean": 0},
+        "MonthlyIncome": {"min": 0, "max": 200000, "mean": 30000},
+    }
+
+# UI inputs for numeric features
+st.sidebar.header("Numeric inputs")
+user_inputs = {}
+for n in numeric_features:
+    conf = defaults.get(n, {"min": 0, "max": 1000, "mean": 0})
+    step = 1 if isinstance(conf["mean"], int) or n != "MonthlyIncome" else 1
+    if n in ["Passport", "OwnCar"]:
+        # Use selectbox for binary features
+        user_inputs[n] = st.sidebar.selectbox(n, options=[0, 1], index=int(conf["mean"]))
+    else:
+        # number_input with reasonable range
+        if n == "MonthlyIncome":
+            user_inputs[n] = st.sidebar.number_input(
+                n,
+                min_value=int(conf["min"]),
+                max_value=int(conf["max"]) if conf["max"] > 0 else 1_000_000,
+                value=int(conf["mean"]),
+                step=step
+            )
+        else:
+            user_inputs[n] = st.sidebar.number_input(
+                n,
+                min_value=int(conf["min"]),
+                max_value=int(conf["max"]) if conf["max"] > 0 else 10000,
+                value=int(conf["mean"]),
+                step=1,
+            )
+
+# UI inputs for categorical features
+st.sidebar.header("Categorical inputs")
+for c in categorical_features:
+    vals = options.get(c)
+    if vals:
+        user_inputs[c] = st.sidebar.selectbox(c, vals)
+    else:
+        # If we don't know categories, allow free text
+        user_inputs[c] = st.sidebar.text_input(c, "")
+
+# Assemble input as DataFrame (matching training columns)
+input_df = pd.DataFrame([user_inputs])
+
+# Ensure categorical dtype for relevant cols
+for c in categorical_features:
+    if c in input_df.columns:
+        input_df[c] = input_df[c].astype("category")
+
+st.subheader("Input preview")
+st.write(input_df.T)
+
+# Load model (download from HF hub)
+model = None
+load_error = None
+try:
+    model_path = hf_hub_download(repo_id=REPO_ID, repo_type="model", filename=MODEL_FILENAME, token=os.environ.get("HF_TOKEN"))
+    model = joblib.load(model_path)
+except Exception as e:
+    load_error = str(e)
+
+if load_error:
+    st.error("Failed to load model from Hugging Face Hub. Check HF_TOKEN and network.\n\n" + load_error)
+else:
+    if st.button("Predict purchase probability"):
+        # Ensure ordering of columns matches model's expected features
+        ordered_cols = numeric_features + categorical_features
+        # Some environments may store y columns as dataframes; ensure all columns present
+        missing = [c for c in ordered_cols if c not in input_df.columns]
+        if missing:
+            st.error(f"Missing features required by model: {missing}")
+        else:
+            X_input = input_df[ordered_cols].copy()
+            proba = model.predict_proba(X_input)[:, 1][0]
+            pred = int(proba >= CLASSIFICATION_THRESHOLD)
+
+            st.metric("Purchase Probability", f"{proba:.3f}")
+            st.metric("Predicted Purchase", "Yes" if pred == 1 else "No")
+            st.write(
+                "Notes: probability threshold = "
+                + str(CLASSIFICATION_THRESHOLD)
+                + ". Adjust threshold for sensitivity/precision tradeoff."
+            )

requirements.txt

@@ -1,3 +1,7 @@
-altair
-pandas
-streamlit
+pandas==2.2.2
+huggingface_hub==0.32.6
+streamlit==1.43.2
+joblib==1.5.1
+scikit-learn==1.6.0
+xgboost==2.1.4
+mlflow==3.0.1