Feast feature store basics in MLOps - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to import the Feast client.

MLOps

from feast import [1]

Drag options to blanks, or click blank then click option'

AFeatureStore

BClient

CFeature

DStore

Attempts:

3 left

💡 Hint

Common Mistakes

Importing Client instead of FeatureStore

Using Store which is not a valid class

The FeatureStore class is used to interact with Feast feature stores.

2fill in blank

medium

Complete the code to initialize the Feast feature store client.

MLOps

store = FeatureStore(repo_path=[1])

Drag options to blanks, or click blank then click option'

Arepo_path

BFeatureStore

C'/path/to/repo'

Dclient

Attempts:

3 left

💡 Hint

Common Mistakes

Passing variable names instead of string path

Omitting quotes around the path

The repo_path parameter expects a string path to the feature store repository.

3fill in blank

hard

Fix the error in the code to retrieve features from Feast.

MLOps

features = store.get_online_features(feature_refs=[1], entity_rows=entity_rows).to_dict()

Drag options to blanks, or click blank then click option'

A'driver.feature1,driver.feature2'

B'driver:feature1,driver:feature2'

Cdriver:feature1,driver:feature2

D['driver:feature1', 'driver:feature2']

Attempts:

3 left

💡 Hint

Common Mistakes

Passing a list instead of a string

Using dot instead of colon in feature references

The feature_refs parameter expects a single string with comma-separated feature references.

4fill in blank

hard

Fill both blanks to define an entity row and retrieve features.

MLOps

entity_rows = [{ '[1]': 1001 }]
features = store.get_online_features(feature_refs='driver:[2]', entity_rows=entity_rows).to_dict()

Drag options to blanks, or click blank then click option'

Adriver_id

Brating

Ctrip_count

Ddriver

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'driver' instead of 'driver_id' as key

Using wrong feature name in feature_refs

The entity row dictionary key should be the entity name driver_id. The feature reference is driver:rating.

5fill in blank

hard

Fill all three blanks to create a feature view and register it.

MLOps

from feast import Feature, [1], ValueType

features = [Feature(name='[2]', dtype=ValueType.FLOAT)]
feature_view = [3](
    name='driver_stats',
    entities=['driver_id'],
    features=features,
    ttl=None
)
store.apply([feature_view])

Drag options to blanks, or click blank then click option'

AFeatureView

Btrip_duration

CDriverStatsView

DFeatureStore

Attempts:

3 left

💡 Hint

Common Mistakes

Using a wrong class name for feature view

Using incorrect feature names

Confusing FeatureStore with FeatureView

The class to create a feature view is FeatureView. The feature name is trip_duration. We use FeatureView to create the feature view instance.