Prerequisites

For the complete documentation index see: llms.txt

All documentation pages available in markdown.

This tutorial assumes you completed Parts 1 and 2 in the provided feature_store_tutorial.ipynb notebook.

Continue your notebook

Open feature_store_tutorial.ipynb. Part 3 starts at Cell 14.

Prepare your environment

Run Cell 14 to confirm Spark is active and your saved model path is available.

Cell 14: Prepare Spark and saved model path

import os
model_path = "./models/trip_decline_risk_lr"
model_exists = os.path.exists(model_path)

print(f"Spark version:  {spark.version}")
print(f"Model path:     {model_path}")
print(f"Saved model:    {'found' if model_exists else 'NOT FOUND - run Part 2 first'}")

Expected output

Spark version:  3.5.3
Model path:     ./models/trip_decline_risk_lr
Saved model:    found

If the model path prints NOT FOUND, re-run the save step from Part 2.

Install the Aerospike Python client

Part 3 introduces a new dependency: the Aerospike Python client. You’ll learn why in the next section.

Run Cell 15 to install and import the Aerospike Python client.

Cell 15: Install and import the Aerospike Python client

%pip install aerospike

import aerospike
print("Aerospike Python client import OK")

Expected output

Aerospike Python client import OK

You’re ready to start building the serving pipeline.