Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
Loading...
The list below contains the functionality that contributors are planning to develop for Feast.
We welcome contribution to all items in the roadmap!
Have questions about the roadmap? Go to the Slack channel to ask on #feast-development.
Data Sources
Offline Stores
Online Stores
Feature Engineering
Streaming
Deployments
Feature Serving
Data Quality Management (See RFC)
Feature Discovery and Governance
Feast users should join #feast-general
or #feast-beginners
to ask questions
Feast developers / contributors should join #feast-development
Design proposals in the form of Request for Comments (RFC).
User surveys and meeting minutes.
Slide decks of conferences our contributors have spoken at.
Slack: Need to speak to a human? Come ask a question in our Slack channel (link above).
We have a user and contributor community call every two weeks (US & EU friendly).
Please join the above Feast user groups in order to see calendar invites to the community calls
Tuesday 10:00 am to 10:30 am PST
We also have a #feast-development
community call every two weeks, where we discuss contributions + brainstorm best practices.
Tuesday 8:00 am to 8:30 am PST
Projects provide complete isolation of feature stores at the infrastructure level. This is accomplished through resource namespacing, e.g., prefixing table names with the associated project. Each project should be considered a completely separate universe of entities and features. It is not possible to retrieve features from multiple projects in a single request. We recommend having a single feature store and a single project per environment (dev
, staging
, prod
).
For offline use cases that only rely on batch data, Feast does not need to ingest data and can query your existing data (leveraging a compute engine, whether it be a data warehouse or (experimental) Spark / Trino). Feast can help manage pushing streaming features to a batch source to make features available for training.
Features are registered as code in a version controlled repository, and tie to data sources + model versions via the concepts of entities, feature views, and feature services. We explore these concepts more in the upcoming concept pages. These features are then stored in a registry, which can be accessed across users and services. The features can then be retrieved via SDK API methods or via a deployed feature server which exposes endpoints to query for online features (to power real time models).
Feast supports several patterns of feature retrieval.
Feature values in Feast are modeled as time-series records. Below is an example of a driver feature view with two feature columns (trips_today
, and earnings_today
):
The above table can be registered with Feast through the following feature view:
Feast is able to join features from one or more feature views onto an entity dataframe in a point-in-time correct way. This means Feast is able to reproduce the state of features at a specific point in the past.
Given the following entity dataframe, imagine a user would like to join the above driver_hourly_stats
feature view onto it, while preserving the trip_success
column:
The timestamps within the entity dataframe above are the events at which we want to reproduce the state of the world (i.e., what the feature values were at those specific points in time). In order to do a point-in-time join, a user would load the entity dataframe and run historical retrieval:
For each row within the entity dataframe, Feast will query and join the selected features from the appropriate feature view data source. Feast will scan backward in time from the entity dataframe timestamp up to a maximum of the TTL time specified.
Please note that the TTL time is relative to each timestamp within the entity dataframe. TTL is not relative to the current point in time (when you run the query).
Below is the resulting joined training dataframe. It contains both the original entity rows and joined feature values:
Three feature rows were successfully joined to the entity dataframe rows. The first row in the entity dataframe was older than the earliest feature rows in the feature view and could not be joined. The last row in the entity dataframe was outside of the TTL window (the event happened 11 hours after the feature row) and also couldn't be joined.
: Find the complete Feast codebase on GitHub.
: See the governance model of Feast, including who the maintainers are and how decisions are made.
: Feel free to ask questions or say hello! This is the main place where maintainers and contributors brainstorm and where users ask questions or discuss best practices.
: We have both a user and developer mailing list.
Feast users should join group by clicking .
Feast developers / contributors should join group by clicking .
: Includes community calls and design meetings.
: This folder is used as a central repository for all Feast resources. For example:
: Our LFAI wiki page contains links to resources for contributors and maintainers.
GitHub Issues: Found a bug or need a feature? .
StackOverflow: Need to ask a question on how to use Feast? We also monitor and respond to .
Zoom:
Meeting notes (incl recordings):
Meeting notes (incl recordings):
Zoom:
The top-level namespace within Feast is a project. Users define one or more within a project. Each feature view contains one or more . These features typically relate to one or more . A feature view must always have a , which in turn is used during the generation of training and when materializing feature values into the online store.
For online use cases, Feast supports ingesting features from batch sources to make them available online (through a process called materialization), and pushing streaming features to make them available both offline / online. We explore this more in the next concept page ()
Training data generation
Fetching user and item features for (user, item) pairs when training a production recommendation model
get_historical_features
Offline feature retrieval for batch predictions
Predicting user churn for all users on a daily basis
get_historical_features
Online feature retrieval for real-time model predictions
Fetching pre-computed features to predict whether a real-time credit card transaction is fraudulent
get_online_features
Feast uses a registry to store all applied Feast objects (e.g. Feature views, entities, etc). The registry exposes methods to apply, list, retrieve and delete these objects, and is an abstraction with multiple implementations.
By default, Feast uses a file-based registry implementation, which stores the protobuf representation of the registry as a serialized file. This registry file can be stored in a local file system, or in cloud storage (in, say, S3 or GCS, or Azure).
The quickstart guides that use feast init
will use a registry on a local file system. To allow Feast to configure a remote file registry, you need to create a GCS / S3 bucket that Feast can understand:
However, there are inherent limitations with a file-based registry, since changing a single field in the registry requires re-writing the whole registry file. With multiple concurrent writers, this presents a risk of data loss, or bottlenecks writes to the registry since all changes have to be serialized (e.g. when running materialization for multiple feature views or time ranges concurrently).
Alternatively, a SQL Registry can be used for a more scalable registry.
The configuration roughly looks like:
This supports any SQLAlchemy compatible database as a backend. The exact schema can be seen in sql.py
We recommend users store their Feast feature definitions in a version controlled repository, which then via CI/CD automatically stays synced with the registry. Users will often also want multiple registries to correspond to different environments (e.g. dev vs staging vs prod), with staging and production registries with locked down write access since they can impact real user traffic. See Running Feast in Production for details on how to set this up.
Users can specify the registry through a feature_store.yaml
config file, or programmatically. We often see teams preferring the programmatic approach because it makes notebook driven development very easy:
feature_store.yaml
fileInstantiating a FeatureStore
object can then point to this:
Create Batch Features: ELT/ETL systems like Spark and SQL are used to transform data in the batch store.
Create Stream Features: Stream features are created from streaming services such as Kafka or Kinesis, and can be pushed directly into Feast via the Push API.
Feast Apply: The user (or CI) publishes versioned controlled feature definitions using feast apply
. This CLI command updates infrastructure and persists definitions in the object store registry.
Feast Materialize: The user (or scheduler) executes feast materialize
which loads features from the offline store into the online store.
Model Training: A model training pipeline is launched. It uses the Feast Python SDK to retrieve a training dataset that can be used for training models.
Get Historical Features: Feast exports a point-in-time correct training dataset based on the list of features and entity dataframe provided by the model training pipeline.
Deploy Model: The trained model binary (and list of features) are deployed into a model serving system. This step is not executed by Feast.
Prediction: A backend system makes a request for a prediction from the model serving service.
Get Online Features: The model serving service makes a request to the Feast Online Serving service for online features using a Feast SDK.
A complete Feast deployment contains the following components:
Feast Registry: An object store (GCS, S3) based registry used to persist feature definitions that are registered with the feature store. Systems can discover feature data by interacting with the registry through the Feast SDK.
Feast Python SDK/CLI: The primary user facing SDK. Used to:
Manage version controlled feature definitions.
Materialize (load) feature values into the online store.
Build and retrieve training datasets from the offline store.
Retrieve online features.
Stream Processor: The Stream Processor can be used to ingest feature data from streams and write it into the online or offline stores. Currently, there's an experimental Spark processor that's able to consume data from Kafka.
Batch Materialization Engine: The Batch Materialization Engine component launches a process which loads data into the online store from the offline store. By default, Feast uses a local in-process engine implementation to materialize data. However, additional infrastructure can be used for a more scalable materialization process.
Online Store: The online store is a database that stores only the latest feature values for each entity. The online store is either populated through materialization jobs or through stream ingestion.
Offline Store: The offline store persists batch data that has been ingested into Feast. This data is used for producing training datasets. For feature retrieval and materialization, Feast does not manage the offline store directly, but runs queries against it. However, offline stores can be configured to support writes if Feast configures logging functionality of served features.
Java and Go Clients are also available for online feature retrieval.
The Feast feature registry is a central catalog of all the feature definitions and their related metadata. It allows data scientists to search, discover, and collaborate on new features.
Each Feast deployment has a single feature registry. Feast only supports file-based registries today, but supports four different backends.
Local
: Used as a local backend for storing the registry during development
S3
: Used as a centralized backend for storing the registry on AWS
GCS
: Used as a centralized backend for storing the registry on GCP
[Alpha] Azure
: Used as centralized backend for storing the registry on Azure Blob storage.
The feature registry is updated during different operations when using Feast. More specifically, objects within the registry (entities, feature views, feature services) are updated when running apply
from the Feast CLI, but metadata about objects can also be updated during operations like materialization.
Users interact with a feature registry through the Feast SDK. Listing all feature views:
Or retrieving a specific feature view:
Feast uses online stores to serve features at low latency. Feature values are loaded from data sources into the online store through materialization, which can be triggered through the materialize
command.
Here is an example batch data source:
Once the above data source is materialized into Feast (using feast materialize
), the feature values will be stored as follows:
In this tutorial we will
Deploy a local feature store with a Parquet file offline store and Sqlite online store.
Build a training dataset using our time series features from our Parquet files.
Ingest batch features ("materialization") and streaming features (via a Push API) into the online store.
Read the latest features from the offline store for batch scoring
Read the latest features from the online store for real-time inference.
Explore the (experimental) Feast UI
In this tutorial, we'll use Feast to generate training data and power online model inference for a ride-sharing driver satisfaction prediction model. Feast solves several common issues in this flow:
Training-serving skew and complex data joins: Feature values often exist across multiple tables. Joining these datasets can be complicated, slow, and error-prone.
Feast joins these tables with battle-tested logic that ensures point-in-time correctness so future feature values do not leak to models.
Online feature availability: At inference time, models often need access to features that aren't readily available and need to be precomputed from other data sources.
Feast manages deployment to a variety of online stores (e.g. DynamoDB, Redis, Google Cloud Datastore) and ensures necessary features are consistently available and freshly computed at inference time.
Feature and model versioning: Different teams within an organization are often unable to reuse features across projects, resulting in duplicate feature creation logic. Models have data dependencies that need to be versioned, for example when running A/B tests on model versions.
Feast enables discovery of and collaboration on previously used features and enables versioning of sets of features (via feature services).
(Experimental) Feast enables light-weight feature transformations so users can re-use transformation logic across online / offline use cases and across models.
Install the Feast SDK and CLI using pip:
Bootstrap a new feature repository using feast init
from the command line.
Let's take a look at the resulting demo repo itself. It breaks down into
data/
contains raw demo parquet data
example_repo.py
contains demo feature definitions
feature_store.yaml
contains a demo setup configuring where data sources are
test_workflow.py
showcases how to run all key Feast commands, including defining, retrieving, and pushing features. You can run this with python test_workflow.py
.
The feature_store.yaml
file configures the key overall architecture of the feature store.
The provider value sets default offline and online stores.
The offline store provides the compute layer to process historical data (for generating training data & feature values for serving).
The online store is a low latency store of the latest feature values (for powering real-time inference).
Valid values for provider
in feature_store.yaml
are:
local: use a SQL registry or local file registry. By default, use a file / Dask based offline store + SQLite online store
gcp: use a SQL registry or GCS file registry. By default, use BigQuery (offline store) + Google Cloud Datastore (online store)
aws: use a SQL registry or S3 file registry. By default, use Redshift (offline store) + DynamoDB (online store)
The raw feature data we have in this demo is stored in a local parquet file. The dataset captures hourly stats of a driver in a ride-sharing app.
There's an included test_workflow.py
file which runs through a full sample workflow:
Register feature definitions through feast apply
Generate a training dataset (using get_historical_features
)
Generate features for batch scoring (using get_historical_features
)
Ingest batch features into an online store (using materialize_incremental
)
Fetch online features to power real time inference (using get_online_features
)
Ingest streaming features into offline / online stores (using push
)
Verify online features are updated / fresher
We'll walk through some snippets of code below and explain
The apply
command scans python files in the current directory for feature view/entity definitions, registers the objects, and deploys infrastructure. In this example, it reads example_repo.py
and sets up SQLite online store tables. Note that we had specified SQLite as the default online store by configuring online_store
in feature_store.yaml
.
To train a model, we need features and labels. Often, this label data is stored separately (e.g. you have one table storing user survey results and another set of tables with feature values). Feast can help generate the features that map to these labels.
Feast needs a list of entities (e.g. driver ids) and timestamps. Feast will intelligently join relevant tables to create the relevant feature vectors. There are two ways to generate this list:
The user can query that table of labels with timestamps and pass that into Feast as an entity dataframe for training data generation.
Note that we include timestamps because we want the features for the same driver at various timestamps to be used in a model.
To power a batch model, we primarily need to generate features with the get_historical_features
call, but using the current timestamp
We now serialize the latest values of features since the beginning of time to prepare for serving (note: materialize-incremental
serializes all new features since the last materialize
call).
At inference time, we need to quickly read the latest feature values for different drivers (which otherwise might have existed only in batch sources) from the online feature store using get_online_features()
. These feature vectors can then be fed to the model.
The driver_activity_v1
feature service pulls all features from the driver_hourly_stats
feature view:
View all registered features, data sources, entities, and feature services with the Web UI.
One of the ways to view this is with the feast ui
command.
test_workflow.py
Take a look at test_workflow.py
again. It showcases many sample flows on how to interact with Feast. You'll see these show up in the upcoming concepts + architecture + tutorial pages as well.
The feature registry is a of Feast metadata. This Protobuf file can be read programmatically from other programming languages, but no compatibility guarantees are made on the internal structure of the registry.
The storage schema of features within the online store mirrors that of the original data source. One key difference is that for each , only the latest feature values are stored. No historical values are stored.
Features can also be written directly to the online store via .
In this tutorial, we focus on a local deployment. For a more in-depth guide on how to use Feast with Snowflake / GCP / AWS deployments, see
Note that there are many other offline / online stores Feast works with, including Spark, Azure, Hive, Trino, and PostgreSQL via community plugins. See for all supported data sources.
A custom setup can also be made by following .
The user can also query that table with a SQL query which pulls entities. See the documentation on for details
You can also use feature services to manage multiple features, and decouple feature view definitions and the features needed by end applications. The feature store can also be used to fetch either online or historical features using the same API below. More information can be found .
Join the to get new updates on Feast / feature stores.
Read the page to understand the Feast data model.
Read the page.
Check out our section for more examples on how to use Feast.
Follow our guide for a more in-depth tutorial on using Feast.
Join other Feast users and contributors in and become part of the community!
These Feast tutorials showcase how to use Feast to simplify end to end model training / serving.
Credit scoring models are used to approve or reject loan applications. In this tutorial we will build a real-time credit scoring system on AWS.
When individuals apply for loans from banks and other credit providers, the decision to approve a loan application is often made through a statistical model. This model uses information about a customer to determine the likelihood that they will repay or default on a loan, in a process called credit scoring.
In this example, we will demonstrate how a real-time credit scoring system can be built using Feast and Scikit-Learn on AWS, using feature data from S3.
This real-time system accepts a loan request from a customer and responds within 100ms with a decision on whether their loan has been approved or rejected.
This end-to-end tutorial will take you through the following steps:
Deploying S3 with Parquet as your primary data source, containing both loan features and zip code features
Deploying Redshift as the interface Feast uses to build training datasets
Registering your features with Feast and configuring DynamoDB for online serving
Building a training dataset with Feast to train your credit scoring model
Loading feature values from S3 into DynamoDB
Making online predictions with your credit scoring model using features from DynamoDB
Making a prediction using a linear regression model is a common use case in ML. This model predicts if a driver will complete a trip based on features ingested into Feast.
In this example, you'll learn how to use some of the key functionality in Feast. The tutorial runs in both local mode and on the Google Cloud Platform (GCP). For GCP, you must have access to a GCP project already, including read and write permissions to BigQuery.
This tutorial guides you on how to use Feast with Scikit-learn. You will learn how to:
Train a model locally (on your laptop) using data from BigQuery
Test the model for online inference using SQLite (for fast iteration)
Test the model for online inference using Firestore (for production use)
Try it and let us know what you think!
The Feast CLI can be used to deploy a feature store to your infrastructure, spinning up any necessary persistent resources like buckets or tables in data stores. The deployment target and effects depend on the provider
that has been configured in your feature_store.yaml file, as well as the feature definitions found in your feature repository.
Here we'll be using the example repository we created in the previous guide, Create a feature store. You can re-create it by running feast init
in a new directory.
To have Feast deploy your infrastructure, run feast apply
from your command line while inside a feature repository:
Depending on whether the feature repository is configured to use a local
provider or one of the cloud providers like GCP
or AWS
, it may take from a couple of seconds to a minute to run to completion.
At this point, no data has been materialized to your online store. Feast apply simply registers the feature definitions with Feast and spins up any necessary infrastructure such as tables. To load data into the online store, run feast materialize
. See Load data into the online store for more details.
If you need to clean up the infrastructure created by feast apply
, use the teardown
command.
Warning: teardown
is an irreversible command and will remove all feature store infrastructure. Proceed with caution!
****
Install Feast using pip:
Install Feast with Snowflake dependencies (required when using Snowflake):
Install Feast with GCP dependencies (required when using BigQuery or Firestore):
Install Feast with AWS dependencies (required when using Redshift or DynamoDB):
Install Feast with Redis dependencies (required when using Redis, either through AWS Elasticache or independently):
A feature repository is a directory that contains the configuration of the feature store and individual features. This configuration is written as code (Python/YAML) and it's highly recommended that teams track it centrally using git. See Feature Repository for a detailed explanation of feature repositories.
The easiest way to create a new feature repository to use feast init
command:
The init
command creates a Python file with feature definitions, sample data, and a Feast configuration file for local development:
Enter the directory:
You can now use this feature repository for development. You can try the following:
Run feast apply
to apply these definitions to Feast.
Edit the example feature definitions in example.py
and run feast apply
again to change feature definitions.
Initialize a git repository in the same directory and checking the feature repository into version control.
A common use case in machine learning, this tutorial is an end-to-end, production-ready fraud prediction system. It predicts in real-time whether a transaction made by a user is fraudulent.
Throughout this tutorial, we’ll walk through the creation of a production-ready fraud prediction system. A prediction is made in real-time as the user makes the transaction, so we need to be able to generate a prediction at low latency.
Our end-to-end example will perform the following workflows:
Computing and backfilling feature data from raw data
Building point-in-time correct training datasets from feature data and training a model
Making online predictions from feature data
Here's a high-level picture of our system architecture on Google Cloud Platform (GCP):
Feast is designed to be easy to use and understand out of the box, with as few infrastructure dependencies as possible. However, there are components used by default that may not scale well. Since Feast is designed to be modular, it's possible to swap such components with more performant components, at the cost of Feast depending on additional infrastructure.
The default Feast registry is a file-based registry. Any changes to the feature repo, or materializing data into the online store, results in a mutation to the registry.
However, there are inherent limitations with a file-based registry, since changing a single field in the registry requires re-writing the whole registry file. With multiple concurrent writers, this presents a risk of data loss, or bottlenecks writes to the registry since all changes have to be serialized (e.g. when running materialization for multiple feature views or time ranges concurrently).
The recommended solution in this case is to use the SQL based registry, which allows concurrent, transactional, and fine-grained updates to the registry. This registry implementation requires access to an existing database (such as MySQL, Postgres, etc).
The default Feast materialization process is an in-memory process, which pulls data from the offline store before writing it to the online store. However, this process does not scale for large data sets, since it's executed on a single-process.
Feast supports pluggable Materialization Engines, that allow the materialization process to be scaled up. Aside from the local process, Feast supports a Lambda-based materialization engine, and a Bytewax-based materialization engine.
Users may also be able to build an engine to scale up materialization using existing infrastructure in their organizations.
The Feast Python SDK allows users to retrieve feature values from an online store. This API is used to look up feature values at low latency during model serving in order to make online predictions.
Online stores only maintain the current state of features, i.e latest feature values. No historical data is stored or served.
Please ensure that you have materialized (loaded) your feature values into the online store before starting
Create a list of features that you would like to retrieve. This list typically comes from the model training step and should accompany the model binary.
Next, we will create a feature store object and call get_online_features()
which reads the relevant feature values directly from the online store.
Feast allows users to load their feature data into an online store in order to serve the latest features to models for online prediction.
Before proceeding, please ensure that you have applied (registered) the feature views that should be materialized.
The materialize command allows users to materialize features over a specific historical time range into the online store.
The above command will query the batch sources for all feature views over the provided time range, and load the latest feature values into the configured online store.
It is also possible to materialize for specific feature views by using the -v / --views
argument.
The materialize command is completely stateless. It requires the user to provide the time ranges that will be loaded into the online store. This command is best used from a scheduler that tracks state, like Airflow.
For simplicity, Feast also provides a materialize command that will only ingest new data that has arrived in the offline store. Unlike materialize
, materialize-incremental
will track the state of previous ingestion runs inside of the feature registry.
The example command below will load only new data that has arrived for each feature view up to the end date and time (2021-04-08T00:00:00
).
The materialize-incremental
command functions similarly to materialize
in that it loads data over a specific time range for all feature views (or the selected feature views) into the online store.
Unlike materialize
, materialize-incremental
automatically determines the start time from which to load features from batch sources of each feature view. The first time materialize-incremental
is executed it will set the start time to the oldest timestamp of each data source, and the end time as the one provided by the user. For each run of materialize-incremental
, the end timestamp will be tracked.
Subsequent runs of materialize-incremental
will then set the start time to the end time of the previous run, thus only loading new data that has arrived into the online store. Note that the end time that is tracked for each run is at the feature view level, not globally for all feature views, i.e, different feature views may have different periods that have been materialized into the online store.
Feast allows users to build a training dataset from time-series feature data that already exists in an offline store. Users are expected to provide a list of features to retrieve (which may span multiple feature views), and a dataframe to join the resulting features onto. Feast will then execute a point-in-time join of multiple feature views onto the provided dataframe, and return the full resulting dataframe.
Please ensure that you have created a feature repository and that you have registered (applied) your feature views with Feast.
Start by defining the feature references (e.g., driver_trips:average_daily_rides
) for the features that you would like to retrieve from the offline store. These features can come from multiple feature tables. The only requirement is that the feature tables that make up the feature references have the same entity (or composite entity), and that they aren't located in the same offline store.
3. Create an entity dataframe
An entity dataframe is the target dataframe on which you would like to join feature values. The entity dataframe must contain a timestamp column called event_timestamp
and all entities (primary keys) necessary to join feature tables onto. All entities found in feature views that are being joined onto the entity dataframe must be found as column on the entity dataframe.
It is possible to provide entity dataframes as either a Pandas dataframe or a SQL query.
Pandas:
In the example below we create a Pandas based entity dataframe that has a single row with an event_timestamp
column and a driver_id
entity column. Pandas based entity dataframes may need to be uploaded into an offline store, which may result in longer wait times compared to a SQL based entity dataframe.
SQL (Alternative):
Below is an example of an entity dataframe built from a BigQuery SQL query. It is only possible to use this query when all feature views being queried are available in the same offline store (BigQuery).
4. Launch historical retrieval
Once the feature references and an entity dataframe are defined, it is possible to call get_historical_features()
. This method launches a job that executes a point-in-time join of features from the offline store onto the entity dataframe. Once completed, a job reference will be returned. This job reference can then be converted to a Pandas dataframe by calling to_df()
.
Feast is highly pluggable and configurable:
One can use existing plugins (offline store, online store, batch materialization engine, providers) and configure those using the built in options. See reference documentation for details.
The other way to customize Feast is to build your own custom components, and then point Feast to delegate to them.
Below are some guides on how to add new custom components:
All Feast operations execute through a provider
. Operations like materializing data from the offline to the online store, updating infrastructure like databases, launching streaming ingestion jobs, building training datasets, and reading features from the online store.
Custom providers allow Feast users to extend Feast to execute any custom logic. Examples include:
Launching custom streaming ingestion jobs (Spark, Beam)
Launching custom batch ingestion (materialization) jobs (Spark, Beam)
Adding custom validation to feature repositories during feast apply
Adding custom infrastructure setup logic which runs during feast apply
Extending Feast commands with in-house metrics, logging, or tracing
Feast comes with built-in providers, e.g, LocalProvider
, GcpProvider
, and AwsProvider
. However, users can develop their own providers by creating a class that implements the contract in the Provider class.
This guide also comes with a fully functional custom provider demo repository. Please have a look at the repository for a representative example of what a custom provider looks like, or fork the repository when creating your own provider.
The fastest way to add custom logic to Feast is to extend an existing provider. The most generic provider is the LocalProvider
which contains no cloud-specific logic. The guide that follows will extend the LocalProvider
with operations that print text to the console. It is up to you as a developer to add your custom code to the provider methods, but the guide below will provide the necessary scaffolding to get you started.
The first step is to define a custom provider class. We've created the MyCustomProvider
below.
Notice how in the above provider we have only overwritten two of the methods on the LocalProvider
, namely update_infra
and materialize_single_feature_view
. These two methods are convenient to replace if you are planning to launch custom batch or streaming jobs. update_infra
can be used for launching idempotent streaming jobs, and materialize_single_feature_view
can be used for launching batch ingestion jobs.
It is possible to overwrite all the methods on the provider class. In fact, it isn't even necessary to subclass an existing provider like LocalProvider
. The only requirement for the provider class is that it follows the Provider contract.
Configure your feature_store.yaml file to point to your new provider class:
Notice how the provider
field above points to the module and class where your provider can be found.
Now you should be able to use your provider by running a Feast command:
It may also be necessary to add the module root path to your PYTHONPATH
as follows:
That's it. You should now have a fully functional custom provider!
Have a look at the custom provider demo repository for a fully functional example of a custom provider. Feel free to fork it when creating your own custom provider!
Tutorial on how to use the SQL registry for scalable registry updates
By default, the registry Feast uses a file-based registry implementation, which stores the protobuf representation of the registry as a serialized file. This registry file can be stored in a local file system, or in cloud storage (in, say, S3 or GCS).
However, there's inherent limitations with a file-based registry, since changing a single field in the registry requires re-writing the whole registry file. With multiple concurrent writers, this presents a risk of data loss, or bottlenecks writes to the registry since all changes have to be serialized (e.g. when running materialization for multiple feature views or time ranges concurrently).
An alternative to the file-based registry is the SQLRegistry which ships with Feast. This implementation stores the registry in a relational database, and allows for changes to individual objects atomically. Under the hood, the SQL Registry implementation uses SQLAlchemy to abstract over the different databases. Consequently, any database supported by SQLAlchemy can be used by the SQL Registry. The following databases are supported and tested out of the box:
PostgreSQL
MySQL
Sqlite
Feast can use the SQL Registry via a config change in the feature_store.yaml file. An example of how to configure this would be:
Specifically, the registry_type needs to be set to sql in the registry config block. On doing so, the path should refer to the Database URL for the database to be used, as expected by SQLAlchemy. No other additional commands are currently needed to configure this registry.
Should you choose to use a database technology that is compatible with one of Feast's supported registry backends, but which speaks a different dialect (e.g. cockroachdb
, which is compatible with postgres
) then some further intervention may be required on your part.
SQLAlchemy
, used by the registry, may not be able to detect your database version without first updating your DSN scheme to the appropriate DBAPI/dialect combination. When this happens, your database is likely using what is referred to as an external dialect in SQLAlchemy
terminology. See your database's documentation for examples on how to set its scheme in the Database URL.
Psycopg2
, which is the database library leveraged by the online and offline stores, is not impacted by the need to speak a particular dialect, and so the following only applies to the registry.
If you are not running Feast in a container, to accomodate SQLAlchemy
's need to speak an external dialect, install additional Python modules like we do as follows using cockroachdb
for example:
If you are running Feast in a container, you will need to create a custom image like we do as follows, again using cockroachdb
as an example:
If you are running Feast in Kubernetes, set the image.repository
and imagePullSecrets
Helm values accordingly to utilize your custom image.
There are some things to note about how the SQL registry works:
Once instantiated, the Registry ensures the tables needed to store data exist, and creates them if they do not.
Upon tearing down the feast project, the registry ensures that the tables are dropped from the database.
The schema for how data is laid out in tables can be found . It is intentionally simple, storing the serialized protobuf versions of each Feast object keyed by its name.
The SQL Registry should be used when materializing feature views concurrently to ensure correctness of data in the registry. This can be achieved by simply running feast materialize or feature_store.materialize multiple times using a correctly configured feature_store.yaml. This will make each materialization process talk to the registry database concurrently, and ensure the metadata updates are serialized.
Feast uses an internal type system to provide guarantees on training and serving data. Feast currently supports eight primitive types - INT32
, INT64
, FLOAT32
, FLOAT64
, STRING
, BYTES
, BOOL
, and UNIX_TIMESTAMP
- and the corresponding array types. Null types are not supported, although the UNIX_TIMESTAMP
type is nullable. The type system is controlled by Value.proto
in protobuf and by types.py
in Python. Type conversion logic can be found in type_map.py
.
During feast apply
, Feast runs schema inference on the data sources underlying feature views. For example, if the schema
parameter is not specified for a feature view, Feast will examine the schema of the underlying data source to determine the event timestamp column, feature columns, and entity columns. Each of these columns must be associated with a Feast type, which requires conversion from the data source type system to the Feast type system.
The feature inference logic calls _infer_features_and_entities
.
_infer_features_and_entities
calls source_datatype_to_feast_value_type
.
source_datatype_to_feast_value_type
cals the appropriate method in type_map.py
. For example, if a SnowflakeSource
is being examined, snowflake_python_type_to_feast_value_type
from type_map.py
will be called.
Feast serves feature values as Value
proto objects, which have a type corresponding to Feast types. Thus Feast must materialize feature values into the online store as Value
proto objects.
The local materialization engine first pulls the latest historical features and converts it to pyarrow.
Then it calls _convert_arrow_to_proto
to convert the pyarrow table to proto format.
This calls python_values_to_proto_values
in type_map.py
to perform the type conversion.
The Feast type system is typically not necessary when retrieving historical features. A call to get_historical_features
will return a RetrievalJob
object, which allows the user to export the results to one of several possible locations: a Pandas dataframe, a pyarrow table, a data lake (e.g. S3 or GCS), or the offline store (e.g. a Snowflake table). In all of these cases, the type conversion is handled natively by the offline store. For example, a BigQuery query exposes a to_dataframe
method that will automatically convert the result to a dataframe, without requiring any conversions within Feast.
As mentioned above in the section on materialization, Feast persists feature values into the online store as Value
proto objects. A call to get_online_features
will return an OnlineResponse
object, which essentially wraps a bunch of Value
protos with some metadata. The OnlineResponse
object can then be converted into a Python dictionary, which calls feast_value_type_to_python_type
from type_map.py
, a utility that converts the Feast internal types to Python native types.
In Feast, each batch data source is associated with a corresponding offline store. For example, a SnowflakeSource
can only be processed by the Snowflake offline store. Otherwise, the primary difference between batch data sources is the set of supported types. Feast has an internal type system, and aims to support eight primitive types (bytes
, string
, int32
, int64
, float32
, float64
, bool
, and timestamp
) along with the corresponding array types. However, not every batch data source supports all of these types.
Below is a matrix indicating which data sources support which types.
For more details on the Feast type system, see .
There are currently four core batch data source implementations: FileSource
, BigQuerySource
, SnowflakeSource
, and RedshiftSource
. There are several additional implementations contributed by the Feast community (PostgreSQLSource
, SparkSource
, and TrinoSource
), which are not guaranteed to be stable or to match the functionality of the core implementations. Details for each specific data source can be found .
bytes
yes
yes
yes
yes
yes
yes
yes
string
yes
yes
yes
yes
yes
yes
yes
int32
yes
yes
yes
yes
yes
yes
yes
int64
yes
yes
yes
yes
yes
yes
yes
float32
yes
yes
yes
yes
yes
yes
yes
float64
yes
yes
yes
yes
yes
yes
yes
bool
yes
yes
yes
yes
yes
yes
yes
timestamp
yes
yes
yes
yes
yes
yes
yes
array types
yes
yes
no
no
yes
yes
no
Note: Feature views do not work with non-timestamped data. A workaround is to insert dummy timestamps.
A feature view is an object that represents a logical group of time-series feature data as it is found in a data source. Depending on the kind of feature view, it may contain some lightweight (experimental) feature transformations (see [Alpha] On demand feature views).
Feature views consist of:
zero or more entities
If the features are not related to a specific object, the feature view might not have entities; see feature views without entities below.
a name to uniquely identify this feature view in the project.
(optional, but recommended) a schema specifying one or more features (without this, Feast will infer the schema by reading from the data source)
(optional, but recommended) metadata (for example, description, or other free-form metadata via tags
)
(optional) a TTL, which limits how far back Feast will look when generating historical datasets
Feature views allow Feast to model your existing feature data in a consistent way in both an offline (training) and online (serving) environment. Feature views generally contain features that are properties of a specific object, in which case that object is defined as an entity and included in the feature view.
Feature views are used during
The generation of training datasets by querying the data source of feature views in order to find historical feature values. A single training dataset may consist of features from multiple feature views.
Loading of feature values into an online store. Feature views determine the storage schema in the online store. Feature values can be loaded from batch sources or from stream sources.
Retrieval of features from the online store. Feature views provide the schema definition to Feast in order to look up features from the online store.
If a feature view contains features that are not related to a specific entity, the feature view can be defined without entities (only timestamps are needed for this feature view).
If the schema
parameter is not specified in the creation of the feature view, Feast will infer the features during feast apply
by creating a Field
for each column in the underlying data source except the columns corresponding to the entities of the feature view or the columns corresponding to the timestamp columns of the feature view's data source. The names and value types of the inferred features will use the names and data types of the columns from which the features were inferred.
"Entity aliases" can be specified to join entity_dataframe
columns that do not match the column names in the source table of a FeatureView.
This could be used if a user has no control over these column names or if there are multiple entities are a subclass of a more general entity. For example, "spammer" and "reporter" could be aliases of a "user" entity, and "origin" and "destination" could be aliases of a "location" entity as shown below.
It is suggested that you dynamically specify the new FeatureView name using .with_name
and join_key_map
override using .with_join_key_map
instead of needing to register each new copy.
A field or feature is an individual measurable property. It is typically a property observed on a specific entity, but does not have to be associated with an entity. For example, a feature of a customer
entity could be the number of transactions they have made on an average month, while a feature that is not observed on a specific entity could be the total number of posts made by all users in the last month. Supported types for fields in Feast can be found in sdk/python/feast/types.py
.
Fields are defined as part of feature views. Since Feast does not transform data, a field is essentially a schema that only contains a name and a type:
Together with data sources, they indicate to Feast where to find your feature values, e.g., in a specific parquet file or BigQuery table. Feature definitions are also used when reading features from the feature store, using feature references.
Feature names must be unique within a feature view.
Each field can have additional metadata associated with it, specified as key-value tags.
On demand feature views allows data scientists to use existing features and request time data (features only available at request time) to transform and create new features. Users define python transformation logic which is executed in both the historical retrieval and online retrieval paths.
Currently, these transformations are executed locally. This is fine for online serving, but does not scale well to offline retrieval.
This enables data scientists to easily impact the online feature retrieval path. For example, a data scientist could
Call get_historical_features
to generate a training dataframe
Iterate in notebook on feature engineering in Pandas
Copy transformation logic into on demand feature views and commit to a dev branch of the feature repository
Verify with get_historical_features
(on a small dataset) that the transformation gives expected output over historical data
Verify with get_online_features
on dev branch that the transformation correctly outputs online features
Submit a pull request to the staging / prod branches which impact production traffic
A stream feature view is an extension of a normal feature view. The primary difference is that stream feature views have both stream and batch data sources, whereas a normal feature view only has a batch data source.
Stream feature views should be used instead of normal feature views when there are stream data sources (e.g. Kafka and Kinesis) available to provide fresh features in an online setting. Here is an example definition of a stream feature view with an attached transformation:
See here for a example of how to use stream feature views to register your own streaming data pipelines in Feast.
Generally, Feast supports several patterns of feature retrieval:
Training data generation (via feature_store.get_historical_features(...)
)
Offline feature retrieval for batch scoring (via feature_store.get_historical_features(...)
)
Online feature retrieval for real-time model predictions
via the SDK: feature_store.get_online_features(...)
via deployed feature server endpoints: requests.post('http://localhost:6566/get-online-features', data=json.dumps(online_request))
Each of these retrieval mechanisms accept:
some way of specifying entities (to fetch features for)
some way to specify the features to fetch (either via feature services, which group features needed for a model version, or feature references)
Before beginning, you need to instantiate a local FeatureStore
object that knows how to parse the registry (see more details)
For code examples of how the below work, inspect the generated repository from feast init -t [YOUR TEMPLATE]
(gcp
, snowflake
, and aws
are the most fully fleshed).
Before diving into how to retrieve features, we need to understand some high level concepts in Feast.
A feature service is an object that represents a logical group of features from one or more feature views. Feature Services allows features from within a feature view to be used as needed by an ML model. Users can expect to create one feature service per model version, allowing for tracking of the features used by models.
Feature services are used during
The generation of training datasets when querying feature views in order to find historical feature values. A single training dataset may consist of features from multiple feature views.
Retrieval of features for batch scoring from the offline store (e.g. with an entity dataframe where all timestamps are now()
)
Retrieval of features from the online store for online inference (with smaller batch sizes). The features retrieved from the online store may also belong to multiple feature views.
Applying a feature service does not result in an actual service being deployed.
Feature services enable referencing all or some features from a feature view.
Retrieving from the online store with a feature service
Retrieving from the offline store with a feature service
This mechanism of retrieving features is only recommended as you're experimenting. Once you want to launch experiments or serve models, feature services are recommended.
Feature references uniquely identify feature values in Feast. The structure of a feature reference in string form is as follows: <feature_view>:<feature>
Feature references are used for the retrieval of features from Feast:
It is possible to retrieve features from multiple feature views with a single request, and Feast is able to join features from multiple tables in order to build a training dataset. However, it is not possible to reference (or retrieve) features from multiple projects at the same time.
Note, if you're using Feature views without entities, then those features can be added here without additional entity values in the entity_rows
parameter.
The timestamp on which an event occurred, as found in a feature view's data source. The event timestamp describes the event time at which a feature was observed or generated.
Event timestamps are used during point-in-time joins to ensure that the latest feature values are joined from feature views onto entity rows. Event timestamps are also used to ensure that old feature values aren't served to models during online serving.
A dataset is a collection of rows that is produced by a historical retrieval from Feast in order to train a model. A dataset is produced by a join from one or more feature views onto an entity dataframe. Therefore, a dataset may consist of features from multiple feature views.
Dataset vs Feature View: Feature views contain the schema of data and a reference to where data can be found (through its data source). Datasets are the actual data manifestation of querying those data sources.
Dataset vs Data Source: Datasets are the output of historical retrieval, whereas data sources are the inputs. One or more data sources can be used in the creation of a dataset.
Feast abstracts away point-in-time join complexities with the get_historical_features
API.
We go through the major steps, and also show example code. Note that the quickstart templates generally have end-to-end working examples for all these cases.
Feast accepts either:
feature services, which group features needed for a model version
Feast accepts either a Pandas dataframe as the entity dataframe (including entity keys and timestamps) or a SQL query to generate the entities.
Both approaches must specify the full entity key needed as well as the timestamps. Feast then joins features onto this dataframe.
You can also pass a SQL string to generate the above dataframe. This is useful for getting all entities in a timeframe from some data source.
Feast will ensure the latest feature values for registered features are available. At retrieval time, you need to supply a list of entities and the corresponding features to be retrieved. Similar to get_historical_features
, we recommend using feature services as a mechanism for grouping features in a model version.
Note: unlike get_historical_features
, the entity_rows
do not need timestamps since you only want one feature value per entity key.
There are several options for retrieving online features: through the SDK, or through a feature server
Snowflake data sources are Snowflake tables or views. These can be specified either by a table reference or a SQL query.
Using a table reference:
Using a query:
Be careful about how Snowflake handles table and column name conventions. In particular, you can read more about quote identifiers here.
The full set of configuration options is available here.
Snowflake data sources support all eight primitive types, but currently do not support array types. For a comparison against other batch data sources, please see here.
Feast (Feature Store) is a customizable operational data system that re-uses existing infrastructure to manage and serve machine learning features to realtime models.
Feast allows ML platform teams to:
Make features consistently available for training and serving by managing an offline store (to process historical data for scale-out batch scoring or model training), a low-latency online store (to power real-time prediction), and a battle-tested feature server (to serve pre-computed features online).
Avoid data leakage by generating point-in-time correct feature sets so data scientists can focus on feature engineering rather than debugging error-prone dataset joining logic. This ensure that future feature values do not leak to models during training.
Decouple ML from data infrastructure by providing a single data access layer that abstracts feature storage from feature retrieval, ensuring models remain portable as you move from training models to serving models, from batch models to realtime models, and from one data infra system to another.
Note: Feast today primarily addresses timestamped structured data.
Feast helps ML platform teams with DevOps experience productionize real-time models. Feast can also help these teams build towards a feature platform that improves collaboration between engineers and data scientists.
Feast is likely not the right tool if you
are in an organization that’s just getting started with ML and is not yet sure what the business impact of ML is
rely primarily on unstructured data
need very low latency feature retrieval (e.g. p99 feature retrieval << 10ms)
have a small team to support a large number of use cases
a data orchestration tool: Feast does not manage or orchestrate complex workflow DAGs. It relies on upstream data pipelines to produce feature values and integrations with tools like Airflow to make features consistently available.
a data warehouse: Feast is not a replacement for your data warehouse or the source of truth for all transformed data in your organization. Rather, Feast is a light-weight downstream layer that can serve data from an existing data warehouse (or other data sources) to models in production.
a database: Feast is not a database, but helps manage data stored in other systems (e.g. BigQuery, Snowflake, DynamoDB, Redis) to make features consistently available at training / serving time
batch + streaming feature engineering: Feast primarily processes already transformed feature values (though it offers experimental light-weight transformations). Users usually integrate Feast with upstream systems (e.g. existing ETL/ELT pipelines). Tecton is a more fully featured feature platform which addresses these needs.
native streaming feature integration: Feast enables users to push streaming features, but does not pull from streaming sources or manage streaming pipelines. Tecton is a more fully featured feature platform which orchestrates end to end streaming pipelines.
feature sharing: Feast has experimental functionality to enable discovery and cataloguing of feature metadata with a Feast web UI (alpha). Feast also has community contributed plugins with DataHub and Amundsen. Tecton also more robustly addresses these needs.
lineage: Feast helps tie feature values to model versions, but is not a complete solution for capturing end-to-end lineage from raw data sources to model versions. Feast also has community contributed plugins with DataHub and Amundsen. Tecton captures more end-to-end lineage by also managing feature transformations.
data quality / drift detection: Feast has experimental integrations with Great Expectations, but is not purpose built to solve data drift / data quality issues. This requires more sophisticated monitoring across data pipelines, served feature values, labels, and model versions.
Many companies have used Feast to power real-world ML use cases such as:
Personalizing online recommendations by leveraging pre-computed historical user or item features.
Online fraud detection, using features that compare against (pre-computed) historical transaction patterns
Churn prediction (an offline model), generating feature values for all users at a fixed cadence in batch
Credit scoring, using pre-computed historical features to compute probability of default
The best way to learn Feast is to use it. Join our Slack channel and head over to our Quickstart and try it out!
Explore the following resources to get started with Feast:
Quickstart is the fastest way to get started with Feast
Concepts describes all important Feast API concepts
Architecture describes Feast's overall architecture.
Tutorials shows full examples of using Feast in machine learning applications.
Running Feast with Snowflake/GCP/AWS provides a more in-depth guide to using Feast.
Reference contains detailed API and design documents.
Contributing contains resources for anyone who wants to contribute to Feast.
An entity is a collection of semantically related features. Users define entities to map to the domain of their use case. For example, a ride-hailing service could have customers and drivers as their entities, which group related features that correspond to these customers and drivers.