Create initial snowflake notebook (5 sections)

[cyclux]

Main goal is to get a working notebook with the five sections.
No markdown cells or final documentation required.

Implementation Breakdown

This feature will be decomposed into the following notebook sections:

1. Setup and Data Loading - Dependencies, Snowflake connection, load data via Arrow into getml
2. Annotations - Set roles (join_key, time_stamp, target, numerical, categorical)
3. Data Model - Define StarSchema with store as entity, join peripheral tables
4. Training - FastProp feature learning, fit pipeline
5. Feature Export - Transform to Arrow, write to Snowflake, register External FeatureView

Technical Context

Input: Prepared population table (WEEKLY_SALES_BY_STORE_WITH_TARGET) from data infrastructure (#42)
Output: External FeatureView registered in Snowflake Feature Store

Key APIs:

• Snowflake: snowflake.snowpark.Session, snowflake.ml.feature_store
• getml: DataFrame.from_arrow(), pipe.transform(), .to_arrow()
• External FV: FeatureView(refresh_freq=None)

File: getml-demo/integration/snowflake/notebooks/snowflake_feature_store.ipynb