Hopsworks feature store allows users to track provenance (lineage) between storage connectors, feature groups, feature views, training datasets and models. Tracking lineage allows users to determine where/if a feature group is being used. You can track if feature groups are being used to create additional (derived) feature groups or feature views.
You can interact with the provenance graph using the UI and the APIs.
The relationship between storage connectors and feature groups is captured automatically when you create an external feature group. You can inspect the relationship between storage connectors and feature groups using the APIs.
=== "Python"
```python
# Retrieve the storage connector
snowflake_sc = fs.get_storage_connector("snowflake_sc")
# Create the user profiles feature group
user_profiles_fg = fs.create_external_feature_group(
name="user_profiles",
version=1,
storage_connector=snowflake_sc,
query="SELECT * FROM USER_PROFILES"
)
user_profiles_fg.save()
```
Starting from a feature group metadata object, you can traverse upstream the provenance graph to retrieve the metadata objects of the storage connectors that are part of the feature group. To do so, you can use the [get_storage_connector_provenance](https://docs.hopsworks.ai/feature-store-api/{{{ hopsworks_version }}}/generated/api/feature_group_api/#get_storage_connector_provenance) method.
=== "Python"
```python
# Returns all storage connectors linked to the provided feature group
lineage = user_profiles_fg.get_storage_connector_provenance()
# List all accessible parent storage connectors
lineage.accessible
# List all deleted parent storage connectors
lineage.deleted
# List all the inaccessible parent storage connectors
lineage.inaccessible
```
=== "Python"
```python
# Returns an accessible storage connector linked to the feature group (if it exists)
user_profiles_fg.get_storage_connector()
```
To traverse the provenance graph in the opposite direction (i.e. from the storage connector to the feature group), you can use the [get_feature_groups_provenance](https://docs.hopsworks.ai/feature-store-api/{{{ hopsworks_version }}}/generated/api/storage_connector_api/#get_feature_groups_provenance) method. When navigating the provenance graph downstream, the deleted feature groups are not tracked by provenance, as such, the deleted property will always return an empty list.
=== "Python"
```python
# Returns all feature groups linked to the provided storage connector
lineage = snowflake_sc.get_feature_groups_provenance()
# List all accessible downstream feature groups
lineage.accessible
# List all the inaccessible downstream feature groups
lineage.inaccessible
```
=== "Python"
```python
# Returns all accessible feature groups linked to the storage connector (if any exists)
snowflake_sc.get_feature_groups()
```
When creating a feature group, it is possible to specify a list of feature groups used to create the derived features. For example, you could have an external feature group defined over a Snowflake or Redshift table, which you use to compute the features and save them in a feature group. You can mark the external feature group as parent of the feature group you are creating by using the parents parameter in the [get_or_create_feature_group](https://docs.hopsworks.ai/feature-store-api/{{{ hopsworks_version }}}/generated/api/feature_group_api/#get_or_create_feature_group) or [create_feature_group](https://docs.hopsworks.ai/feature-store-api/{{{ hopsworks_version }}}/generated/api/feature_group_api/#create_feature_group) methods:
=== "Python"
```python
# Retrieve the feature group
profiles_fg = fs.get_external_feature_group("user_profiles", version=1)
# Do feature engineering
age_df = transaction_df.merge(profiles_fg.read(), on="cc_num", how="left")
transaction_df["age_at_transaction"] = (age_df["datetime"] - age_df["birthdate"]) / np.timedelta64(1, "Y")
# Create the transaction feature group
transaction_fg = fs.get_or_create_feature_group(
name="transaction_fraud_batch",
version=1,
description="Transaction features",
primary_key=["cc_num"],
event_time="datetime",
parents=[profiles_fg]
)
transaction_fg.insert(transaction_df)
```
Another example use case for derived feature group is if you have a feature group containing features with daily resolution and you are using the content of that feature group to populate a second feature group with monthly resolution:
=== "Python"
```python
# Retrieve the feature group
daily_transaction_fg = fs.get_feature_group("daily_transaction", version=1)
daily_transaction_df = daily_transaction_fg.read()
# Do feature engineering
cc_group = daily_transaction_df[["cc_num", "amount", "datetime"]] \
.groupby("cc_num") \
.rolling("1M", on="datetime")
monthly_transaction_df = pd.DataFrame(cc_group.mean())
# Create the transaction feature group
monthly_transaction_fg = fs.get_or_create_feature_group(
name="monthly_transaction_fraud_batch",
version=1,
description="Transaction features - monthly aggregates",
primary_key=["cc_num"],
event_time="datetime",
parents=[daily_transaction_fg]
)
monthly_transaction_fg.insert(monthly_transaction_df)
```
You can query the provenance graph of a feature group using the UI and the APIs. From the APIs you can list the parent feature groups by calling the method [get_parent_feature_groups](https://docs.hopsworks.ai/feature-store-api/{{{ hopsworks_version }}}/generated/api/feature_group_api/#get_parent_feature_groups)
=== "Python"
```python
lineage = transaction_fg.get_parent_feature_groups()
# List all accessible parent feature groups
lineage.accessible
# List all deleted parent feature groups
lineage.deleted
# List all the inaccessible parent feature groups
lineage.inaccessible
```
A parent is marked as deleted (and added to the deleted list) if the parent feature group was deleted. inaccessible if you no longer have access to the parent feature group (e.g. the parent feature group belongs to a project you no longer have access to).
To traverse the provenance graph in the opposite direction (i.e. from the parent feature group to the child), you can use the [get_generate_feature_groups](https://docs.hopsworks.ai/feature-store-api/{{{ hopsworks_version }}}/generated/api/feature_group_api/#get_generated_feature_groups) method. When navigating the provenance graph downstream, the deleted feature groups are not tracked by provenance, as such, the deleted property will always return an empty list.
=== "Python"
```python
lineage = transaction_fg.get_generated_feature_groups()
# List all accessible child feature groups
lineage.accessible
# List all the inaccessible child feature groups
lineage.inaccessible
```
You can also visualize the relationship between the parent and child feature groups in the UI. In each feature group overview page you can find a provenance section with the graph of parent storage connectors/feature groups and child feature groups/feature views.
The relationship between feature groups and feature views is captured automatically when you create a feature view. You can inspect the relationship between feature groups and feature views using the APIs or the UI.
Starting from a feature view metadata object, you can traverse upstream the provenance graph to retrieve the metadata objects of the feature groups that are part of the feature view. To do so, you can use the [get_parent_feature_groups](https://docs.hopsworks.ai/feature-store-api/{{{ hopsworks_version }}}/generated/api/feature_view_api/#get_parent_feature_groups) method.
=== "Python"
```python
lineage = fraud_fv.get_parent_feature_groups()
# List all accessible parent feature groups
lineage.accessible
# List all deleted parent feature groups
lineage.deleted
# List all the inaccessible parent feature groups
lineage.inaccessible
```
You can also traverse the provenance graph in the opposite direction. Starting from a feature group you can navigate downstream and list all the feature views the feature group is used in. As for the derived feature group example above, when navigating the provenance graph downstream deleted feature views are not tracked. As such, the deleted property will always be empty.
=== "Python"
```python
lineage = transaction_fg.get_generated_feature_views()
# List all accessible downstream feature views
lineage.accessible
# List all the inaccessible downstream feature views
lineage.inaccessible
```
You can also traverse the provenance graph downstream to retrieve the models which use training datasets of this feature view as its parents. === "Python"
```python
models = fraud_fv.get_models_provenance()
# List all accessible models
lineage.accessible
# List all the inaccessible models
lineage.inaccessible
```
You can also retrieve only the models generated from specific training dataset versions: === "Python"
```python
models = fraud_fv.get_models_provenance(training_dataset_version: 1)
```
You can also retrive directly the accessible model objects, without the need to extract them from the provenance links object: === "Python"
```python
#List all accessible models
models = fraud_fv.get_models()
#List accessible models trained from a specific training dataset version
models = fraud_fv.get_models(training_dataset_version: 1)
```
Also we added a utility method to retrieve from the user's accessible models, the last trained one. Last is determined based on timestamp when it was saved into the model registry. === "Python"
```python
#Retrieve newest model from all user's accessible models based on this feature view
model = fraud_fv.get_newest_model()
#Retrieve newest model from all user's accessible models based on this training dataset version
model = fraud_fv.get_newest_model(training_dataset_version: 1)
```
In the feature view overview UI you can explore the provenance graph of the feature view:
The relationship between feature views and models is captured automatically when you create a model. You can inspect the relationship between feature views and models using the APIs or the UI. === "Python"
```python
lineage = model.get_feature_view_provenance()
# List all accessible parent feature views
lineage.accessible
# List all deleted parent feature views
lineage.deleted
# List all the inaccessible parent feature views
lineage.inaccessible
```
You can also retrieve the training dataset provenance object. === "Python"
```python
lineage = model.get_training_dataset_provenance()
# List all accessible parent training datasets
lineage.accessible
# List all deleted parent training datasets
lineage.deleted
# List all the inaccessible parent training datasets
lineage.inaccessible
```
You can also retrieve directly the parent feature view object, without the need to extract them from the provenance links object === "Python"
```python
feature_view = model.get_feature_view()
```
This utility method also has the options to initialize the required components for batch or online retrieval of feature vectors. === "Python"
```python
model.get_feature_view(init: bool = True, online: Optional[bool]: None)
```
By default, the base init for feature vector retrieval is enabled. In case you have a workflow that requires more particular options, you can disable this base init by setting the init to false.
The method detects if it is running within a deployment and will initialize the feature vector retrieval for the serving.
If the online argument is provided and true it will initialize for online feature vector retrieval.
If the online argument is provided and false it will initialize the feature vector retrieval for batch scoring.