import logging

import os

import warnings

from pathlib import Path

from typing import Any, Dict, Optional

import yaml

from pydantic import (

BaseModel,

ConfigDict,

Field,

StrictBool,

StrictInt,

StrictStr,

ValidationError,

ValidationInfo,

field_validator,

model_validator,

)

from feast.errors import (

FeastFeatureServerTypeInvalidError,

FeastInvalidAuthConfigClass,

FeastOfflineStoreInvalidName,

FeastOnlineStoreInvalidName,

FeastRegistryNotSetError,

FeastRegistryTypeInvalidError,

)

from feast.importer import import_class

from feast.permissions.auth.auth_type import AuthType

warnings.simplefilter("once", RuntimeWarning)

_logger = logging.getLogger(__name__)

# These dict exists so that:

# - existing values for the online store type in featurestore.yaml files continue to work in a backwards compatible way

# - first party and third party implementations can use the same class loading code path.

REGISTRY_CLASS_FOR_TYPE = {

"file": "feast.infra.registry.registry.Registry",

"sql": "feast.infra.registry.sql.SqlRegistry",

"snowflake.registry": "feast.infra.registry.snowflake.SnowflakeRegistry",

"remote": "feast.infra.registry.remote.RemoteRegistry",

}

BATCH_ENGINE_CLASS_FOR_TYPE = {

"local": "feast.infra.compute_engines.local.compute.LocalComputeEngine",

"snowflake.engine": "feast.infra.compute_engines.snowflake.snowflake_engine.SnowflakeComputeEngine",

"lambda": "feast.infra.compute_engines.aws_lambda.lambda_engine.LambdaComputeEngine",

"k8s": "feast.infra.compute_engines.kubernetes.k8s_engine.KubernetesComputeEngine",

"spark.engine": "feast.infra.compute_engines.spark.compute.SparkComputeEngine",

"ray.engine": "feast.infra.compute_engines.ray.compute.RayComputeEngine",

"flink.engine": "feast.infra.compute_engines.flink.compute.FlinkComputeEngine",

"spark_application": "feast.infra.compute_engines.spark_application.compute.SparkApplicationComputeEngine",

"trino.engine": "feast.infra.compute_engines.trino.compute.TrinoComputeEngine",

"trino": "feast.infra.compute_engines.trino.compute.TrinoComputeEngine",

}

LEGACY_ONLINE_STORE_CLASS_FOR_TYPE = {

"feast.infra.online_stores.contrib.postgres.PostgreSQLOnlineStore": "feast.infra.online_stores.postgres_online_store.PostgreSQLOnlineStore",

"feast.infra.online_stores.contrib.hbase_online_store.hbase.HbaseOnlineStore": "feast.infra.online_stores.hbase_online_store.hbase.HbaseOnlineStore",

"feast.infra.online_stores.contrib.cassandra_online_store.cassandra_online_store.CassandraOnlineStore": "feast.infra.online_stores.cassandra_online_store.cassandra_online_store.CassandraOnlineStore",

"feast.infra.online_stores.contrib.mysql_online_store.mysql.MySQLOnlineStore": "feast.infra.online_stores.mysql_online_store.mysql.MySQLOnlineStore",

"feast.infra.online_stores.contrib.hazelcast_online_store.hazelcast_online_store.HazelcastOnlineStore": "feast.infra.online_stores.hazelcast_online_store.hazelcast_online_store.HazelcastOnlineStore",

"feast.infra.online_stores.contrib.elasticsearch.ElasticSearchOnlineStore": "feast.infra.online_stores.elasticsearch_online_store.elasticsearch.ElasticSearchOnlineStore",

"feast.infra.online_stores.contrib.singlestore_online_store.singlestore.SingleStoreOnlineStore": "feast.infra.online_stores.singlestore_online_store.singlestore.SingleStoreOnlineStore",

"feast.infra.online_stores.contrib.qdrant.QdrantOnlineStore": "feast.infra.online_stores.qdrant_online_store.qdrant.QdrantOnlineStore",

"feast.infra.online_stores.contrib.milvus.MilvusOnlineStore": "feast.infra.online_stores.milvus.MilvusOnlineStore",

}

ONLINE_STORE_CLASS_FOR_TYPE = {

"aerospike": "feast.infra.online_stores.aerospike_online_store.AerospikeOnlineStore",

"sqlite": "feast.infra.online_stores.sqlite.SqliteOnlineStore",

"datastore": "feast.infra.online_stores.datastore.DatastoreOnlineStore",

"redis": "feast.infra.online_stores.redis.RedisOnlineStore",

"dynamodb": "feast.infra.online_stores.dynamodb.DynamoDBOnlineStore",

"snowflake.online": "feast.infra.online_stores.snowflake.SnowflakeOnlineStore",

"bigtable": "feast.infra.online_stores.bigtable.BigtableOnlineStore",

"postgres": "feast.infra.online_stores.postgres_online_store.postgres.PostgreSQLOnlineStore",

"hbase": "feast.infra.online_stores.hbase_online_store.hbase.HbaseOnlineStore",

"cassandra": "feast.infra.online_stores.cassandra_online_store.cassandra_online_store.CassandraOnlineStore",

"scylladb": "feast.infra.online_stores.scylladb_online_store.scylladb.ScyllaDBOnlineStore",

"mysql": "feast.infra.online_stores.mysql_online_store.mysql.MySQLOnlineStore",

"hazelcast": "feast.infra.online_stores.hazelcast_online_store.hazelcast_online_store.HazelcastOnlineStore",

"elasticsearch": "feast.infra.online_stores.elasticsearch_online_store.elasticsearch.ElasticSearchOnlineStore",

"remote": "feast.infra.online_stores.remote.RemoteOnlineStore",

"singlestore": "feast.infra.online_stores.singlestore_online_store.singlestore.SingleStoreOnlineStore",

"qdrant": "feast.infra.online_stores.qdrant_online_store.qdrant.QdrantOnlineStore",

"couchbase.online": "feast.infra.online_stores.couchbase_online_store.couchbase.CouchbaseOnlineStore",

"milvus": "feast.infra.online_stores.milvus_online_store.milvus.MilvusOnlineStore",

"mongodb": "feast.infra.online_stores.mongodb_online_store.MongoDBOnlineStore",

"hybrid": "feast.infra.online_stores.hybrid_online_store.hybrid_online_store.HybridOnlineStore",

"chronon": "feast.infra.online_stores.chronon_online_store.chronon.ChrononOnlineStore",

**LEGACY_ONLINE_STORE_CLASS_FOR_TYPE,

}

OFFLINE_STORE_CLASS_FOR_TYPE = {

"file": "feast.infra.offline_stores.dask.DaskOfflineStore",

"dask": "feast.infra.offline_stores.dask.DaskOfflineStore",

"bigquery": "feast.infra.offline_stores.bigquery.BigQueryOfflineStore",

"redshift": "feast.infra.offline_stores.redshift.RedshiftOfflineStore",

"snowflake.offline": "feast.infra.offline_stores.snowflake.SnowflakeOfflineStore",

"spark": "feast.infra.offline_stores.contrib.spark_offline_store.spark.SparkOfflineStore",

"trino": "feast.infra.offline_stores.contrib.trino_offline_store.trino.TrinoOfflineStore",

"postgres": "feast.infra.offline_stores.contrib.postgres_offline_store.postgres.PostgreSQLOfflineStore",

"athena": "feast.infra.offline_stores.contrib.athena_offline_store.athena.AthenaOfflineStore",

"mssql": "feast.infra.offline_stores.contrib.mssql_offline_store.mssql.MsSqlServerOfflineStore",

"duckdb": "feast.infra.offline_stores.duckdb.DuckDBOfflineStore",

"remote": "feast.infra.offline_stores.remote.RemoteOfflineStore",

"couchbase.offline": "feast.infra.offline_stores.contrib.couchbase_offline_store.couchbase.CouchbaseColumnarOfflineStore",

"clickhouse": "feast.infra.offline_stores.contrib.clickhouse_offline_store.clickhouse.ClickhouseOfflineStore",

"ray": "feast.infra.offline_stores.contrib.ray_offline_store.ray.RayOfflineStore",

"oracle": "feast.infra.offline_stores.contrib.oracle_offline_store.oracle.OracleOfflineStore",

"chronon": "feast.infra.offline_stores.contrib.chronon_offline_store.chronon.ChrononOfflineStore",

"hybrid": "feast.infra.offline_stores.hybrid_offline_store.HybridOfflineStore",

}

FEATURE_SERVER_CONFIG_CLASS_FOR_TYPE = {

"local": "feast.infra.feature_servers.local_process.config.LocalFeatureServerConfig",

"mcp": "feast.infra.mcp_servers.mcp_config.McpFeatureServerConfig",

}

ALLOWED_AUTH_TYPES = ["no_auth", "kubernetes", "oidc"]

AUTH_CONFIGS_CLASS_FOR_TYPE = {

"no_auth": "feast.permissions.auth_model.NoAuthConfig",

"kubernetes": "feast.permissions.auth_model.KubernetesAuthConfig",

"oidc": "feast.permissions.auth_model.OidcAuthConfig",

"oidc_client": "feast.permissions.auth_model.OidcClientAuthConfig",

}

_OIDC_CLIENT_KEYS = frozenset(

{"client_secret", "token", "token_env_var", "username", "password"}

)

def _is_oidc_client_config(auth_dict: dict) -> bool:

"""Decide whether an OIDC auth dict should be routed to OidcClientAuthConfig.

True when the dict carries any client-credential key, or when it is a bare

``{"type": "oidc"}`` dict with no server-side keys (auth_discovery_url /

client_id), which signals token-passthrough via FEAST_OIDC_TOKEN.

"""

if auth_dict.get("type") != AuthType.OIDC.value:

return False

has_client_keys = bool(_OIDC_CLIENT_KEYS & auth_dict.keys())

has_server_keys = "auth_discovery_url" in auth_dict

return has_client_keys or not has_server_keys

class FeastBaseModel(BaseModel):

"""Feast Pydantic Configuration Class"""

model_config = ConfigDict(arbitrary_types_allowed=True, extra="allow")

class FeastConfigBaseModel(BaseModel):

"""Feast Pydantic Configuration Class"""

model_config = ConfigDict(arbitrary_types_allowed=True, extra="forbid")

class McpRegistryConfig(FeastBaseModel):

"""MCP (Model Context Protocol) configuration for the registry REST server."""

enabled: StrictBool = False

""" bool: Enable MCP support on the REST registry server. """

class RegistryConfig(FeastBaseModel):

"""Metadata Store Configuration. Configuration that relates to reading from and writing to the Feast registry."""

registry_type: StrictStr = "file"

""" str: Provider name or a class name that implements Registry."""

registry_store_type: Optional[StrictStr] = None

""" str: Provider name or a class name that implements RegistryStore. """

path: StrictStr = ""

""" str: Path to metadata store.

If registry_type is 'file', then an be a local path, or remote object storage path, e.g. a GCS URI

If registry_type is 'sql', then this is a database URL as expected by SQLAlchemy """

cache_ttl_seconds: StrictInt = 600

"""int: The cache TTL is the amount of time registry state will be cached in memory. If this TTL is exceeded then

the registry will be refreshed when any feature store method asks for access to registry state. The TTL can be

set to infinity by setting TTL to 0 seconds, which means the cache will only be loaded once and will never

expire. Users can manually refresh the cache by calling feature_store.refresh_registry() """

cache_mode: StrictStr = "sync"

"""str: Cache mode type. Possible options are 'sync' (immediate refresh after each write operation) and

'thread' (asynchronous background refresh at cache_ttl_seconds intervals). In 'sync' mode, registry changes

are immediately visible. In 'thread' mode, changes may take up to

cache_ttl_seconds to be visible."""

s3_additional_kwargs: Optional[Dict[str, str]] = None

""" Dict[str, str]: Extra arguments to pass to boto3 when writing the registry file to S3. """

purge_feast_metadata: StrictBool = False

""" bool: Stops using feast_metadata table and delete data from feast_metadata table.

Once this is set to True, it cannot be reverted back to False. Reverting back to False will

only reset the project but not all the projects"""

enable_online_feature_view_versioning: StrictBool = False

""" bool: Enable versioned resolution of version-qualified feature references

(e.g., 'fv@v2:feature'). Despite the 'online' in its name, this flag gates

both online and offline (get_historical_features) versioned reads, which

share the same resolution path; when True, each schema version also gets

its own online store table and can be queried independently. Version

history tracking in the registry is always active regardless of this

setting. """

serve_features_while_materializing: StrictBool = False

""" bool: Allow online serving to continue for a feature view while it is in

the ``MATERIALIZING`` state. When ``feature_store.materialize()`` runs against

a shared registry it transitions the feature view to ``MATERIALIZING`` and

commits that state, which otherwise causes concurrent feature servers to

reject requests until materialization completes. When True, feature views in

the ``MATERIALIZING`` state keep serving their last-materialized values.

Truly-unavailable states (e.g. ``CREATED``, ``GENERATED``) remain gated. """

mcp: Optional[McpRegistryConfig] = None

""" McpRegistryConfig: MCP (Model Context Protocol) configuration for the registry REST server. """

@staticmethod

def _normalize_postgres_scheme(value: str, field_name: str) -> str:

"""Rewrite a bare ``postgresql://`` URL to the psycopg3 driver, with a warning.

SQLAlchemy resolves a bare ``postgresql://`` to the psycopg2 driver, while

feast standardizes on psycopg3 (``postgresql+psycopg``). Pass an explicit

``postgresql+psycopg2`` to keep psycopg2. Shared by the ``path`` and

``read_path`` validators so both endpoints normalize identically.

"""

if value.startswith("postgresql://"):

_logger.warning(

f"The `{field_name}` of the `RegistryConfig` starts with a plain "

"`postgresql` string. We are updating this to `postgresql+psycopg` "

"to ensure that the `psycopg3` driver is used by `sqlalchemy`. If "

f"you want to use `psycopg2` pass `postgresql+psycopg2` explicitly "

f"to `{field_name}`. To silence this warning, pass `postgresql+psycopg` "

f"explicitly to `{field_name}`."

)

# Rewrite only the leading scheme, not any later occurrence (e.g.

# inside credentials or a query string).

return "postgresql+psycopg://" + value[len("postgresql://") :]

return value

@field_validator("path")

def validate_path(cls, path: str, values: ValidationInfo) -> str:

if values.data.get("registry_type") == "sql":

return cls._normalize_postgres_scheme(path, "path")

return path

@model_validator(mode="after")

def _warn_on_unsafe_serve_while_materializing(self) -> "RegistryConfig":

"""Warn when ``serve_features_while_materializing`` is combined with a

registry configuration where it is a no-op or where its guarantees may

not hold. These are advisory warnings only, never hard failures.

"""

if self.serve_features_while_materializing:

if self.registry_type != "sql":

_logger.warning(

"`serve_features_while_materializing` is enabled but "

f"`registry_type` is '{self.registry_type}'. The flag targets "

"the serving interruption caused by concurrent servers sharing "

"a `sql` registry that materialization flips to `MATERIALIZING`; "

"on other registry types it typically has no effect."

)

if self.cache_mode == "thread":

_logger.warning(

"`serve_features_while_materializing` is enabled together with "

"`cache_mode='thread'`. In thread mode the registry cache may lag "

"by up to `cache_ttl_seconds`, so feature-view state transitions "

"(including the exit from `MATERIALIZING`) may be observed stale."

)

return self

class MaterializationConfig(BaseModel):

"""Configuration options for feature materialization behavior."""

pull_latest_features: StrictBool = False

""" bool: If true, feature retrieval jobs will only pull the latest feature values for each entity.

If false, feature retrieval jobs will pull all feature values within the specified time range. """

online_write_batch_size: Optional[int] = Field(default=None, gt=0)

""" int: Number of rows to write to online store per batch during materialization.

If None (default), all rows are written in a single batch for backward compatibility.

Set to a positive integer (e.g., 10000) to enable batched writes.

Supported compute engines: local, spark, ray. """

class DataQualityMonitoringConfig(FeastConfigBaseModel):

"""Data Quality Monitoring configuration."""

auto_baseline: StrictBool = True

"""Whether baseline distribution is computed automatically on ``feast apply``."""

class OpenLineageConsumerConfig(FeastBaseModel):

"""Configuration for the OpenLineage consumer (event receiver)."""

enabled: StrictBool = False

""" bool: Whether the consumer is enabled. """

store_type: StrictStr = "sql"

""" str: Storage backend type. Currently only 'sql' is supported. """

connection_string: Optional[StrictStr] = None

""" str: Optional separate database connection string. """

api_key: Optional[StrictStr] = None

""" str: API key for authenticating producers sending events. """

namespace_mapping: Optional[Dict[str, str]] = None

""" dict: Read-side RBAC bridge mapping external OL namespaces to Feast project

names. Users who can DESCRIBE a project also see lineage from mapped namespaces.

Example: {"spark://ml-team": "ml_team", "airflow://prod-cluster": "ml_team"} """

retention_days: int = 30

""" int: Number of days to retain OpenLineage events and runs. Set to 0 to

disable automatic pruning. Default: 30 days. """

retention_check_interval_hours: int = 6

""" int: How often the background pruning task runs, in hours. Default: 6. """

standalone_server: StrictBool = False

""" bool: When true, the retention background task is delegated to the

standalone lineage server. All consumer API endpoints remain available

on both servers. """

class OpenLineageConfig(FeastBaseModel):

"""Configuration for OpenLineage integration.

This enables automatic data lineage tracking for Feast operations like

materialization, feature retrieval, and registry changes.

Example configuration in feature_store.yaml:

openlineage:

enabled: true

transport_type: http

transport_url: http://localhost:5000

transport_endpoint: api/v1/lineage

namespace: feast

"""

enabled: StrictBool = False

""" bool: Whether OpenLineage integration is enabled. Defaults to False. """

transport_type: Optional[StrictStr] = None

""" str: Type of transport (http, console, file, kafka). Defaults to None (uses OpenLineage SDK defaults). """

transport_url: Optional[StrictStr] = None

""" str: URL for HTTP transport. Required when transport_type is 'http'. """

transport_endpoint: StrictStr = "api/v1/lineage"

""" str: API endpoint for HTTP transport. Defaults to 'api/v1/lineage'. """

api_key: Optional[StrictStr] = None

""" str: Optional API key for authentication with the lineage server. """

namespace: StrictStr = "feast"

""" str: Default namespace for Feast jobs and datasets. """

producer: StrictStr = "feast"

""" str: Producer identifier for OpenLineage events. """

emit_on_apply: StrictBool = True

""" bool: Emit lineage events when 'feast apply' is called. """

emit_on_materialize: StrictBool = True

""" bool: Emit lineage events during materialization. """

consumer: Optional[OpenLineageConsumerConfig] = None

""" OpenLineageConsumerConfig: Consumer (event receiver) configuration. """

def to_openlineage_config(self):

"""Convert to feast.openlineage.OpenLineageConfig."""

from feast.openlineage.config import OpenLineageConfig as OLConfig

from feast.openlineage.config import (

OpenLineageConsumerConfig as OLConsumerConfig,

)

consumer = None

if self.consumer:

consumer = OLConsumerConfig(

enabled=self.consumer.enabled,

store_type=self.consumer.store_type,

connection_string=self.consumer.connection_string,

api_key=self.consumer.api_key,

namespace_mapping=self.consumer.namespace_mapping or {},

retention_days=self.consumer.retention_days,

retention_check_interval_hours=self.consumer.retention_check_interval_hours,

standalone_server=self.consumer.standalone_server,

)

return OLConfig(

enabled=self.enabled,

transport_type=self.transport_type,

transport_url=self.transport_url,

transport_endpoint=self.transport_endpoint,

api_key=self.api_key,

namespace=self.namespace,

producer=self.producer,

emit_on_apply=self.emit_on_apply,

emit_on_materialize=self.emit_on_materialize,

consumer=consumer or OLConsumerConfig(),

)

class EmbeddingModelConfig(FeastConfigBaseModel):

"""Configuration for the query-time embedding model used by the feature server.

Required when using ``openai_search`` or the

``/v1/vector_stores/{vector_store_id}/search`` endpoint.

**Sentence Transformers** (default) — runs locally, no API key required.

Ideal for air-gapped or cost-sensitive deployments. Requires the

``sentence-transformers`` package (``pip install sentence-transformers``).

Example in ``feature_store.yaml``::

embedding_model:

provider: sentence_transformers # default; can be omitted

model: all-MiniLM-L6-v2

Custom providers can be plugged in by implementing the

:class:`~feast.embedder.EmbeddingProvider` protocol and passing an

instance to :class:`~feast.feature_store.FeatureStore`.

"""

provider: str = "sentence_transformers"

"""Embedding backend to use. Supported values:

``'sentence_transformers'`` (default)."""

model: str

"""Model identifier.

Any HuggingFace model name compatible with ``SentenceTransformer``,

e.g. ``'all-MiniLM-L6-v2'``, ``'BAAI/bge-small-en-v1.5'``.

"""

class RepoConfig(FeastBaseModel):

"""Repo config. Typically loaded from `feature_store.yaml`"""

project: StrictStr

""" str: This acts as a Feast unique project identifier. This can be any alphanumeric string and can have '_' but can not start with '_'.

You can have multiple independent feature repositories deployed to the same cloud

provider account, as long as they have different project identifier.

"""

project_description: Optional[StrictStr] = None

""" str: Optional description of the project to provide context about the project's purpose and usage.

"""

provider: StrictStr = "local"

""" str: local or gcp or aws """

registry_config: Any = Field(alias="registry", default="data/registry.db")

""" Configures the registry.

Can be:

1. str: a path to a file based registry (a local path, or remote object storage path, e.g. a GCS URI)

2. RegistryConfig: A fully specified file based registry or SQL based registry

3. SnowflakeRegistryConfig: Using a Snowflake table to store the registry

"""

online_config: Any = Field(None, alias="online_store")

""" OnlineStoreConfig: Online store configuration (optional depending on provider) """

auth: Any = Field(None, alias="auth")

""" auth: Optional if the services needs the authentication against IDPs (optional depending on provider) """

offline_config: Any = Field(None, alias="offline_store")

""" OfflineStoreConfig: Offline store configuration (optional depending on provider) """

batch_engine_config: Any = Field(None, alias="batch_engine")

""" BatchMaterializationEngine: Batch materialization configuration (optional depending on provider)"""

feature_server: Optional[Any] = None

""" FeatureServerConfig: Feature server configuration (optional depending on provider) """

embedding_model: Optional[EmbeddingModelConfig] = Field(

None, alias="embedding_model"

)

""" EmbeddingModelConfig: Embedding model configuration.

Required when using openai_search or the

OpenAI-compatible vector store search endpoint. """

flags: Any = None

""" Flags (deprecated field): Feature flags for experimental features """

repo_path: Optional[Path] = None

"""When using relative path in FileSource path, this parameter is mandatory"""

entity_key_serialization_version: StrictInt = 3

""" Entity key serialization version: This version is used to control what serialization scheme is

used when writing data to the online store.

A value of 3 uses the latest serialization scheme, supported as of Feast 0.38.

Version Schemas:

v3: add entity_key value length to serialized bytes to enable deserialization, which can be used in retrieval of entity_key in document retrieval.

"""

coerce_tz_aware: Optional[bool] = True

""" If True, coerces entity_df timestamp columns to be timezone aware (to UTC by default). """

materialization_config: MaterializationConfig = Field(

MaterializationConfig(), alias="materialization"

)

""" MaterializationConfig: Configuration options for feature materialization behavior. """

openlineage_config: Optional[OpenLineageConfig] = Field(None, alias="openlineage")

""" Configuration for OpenLineage data lineage integration (optional). """

mlflow_config: Optional[Any] = Field(None, alias="mlflow")

""" MlflowConfig: Configuration for MLflow experiment tracking integration (optional). """

data_quality_monitoring_config: Optional[DataQualityMonitoringConfig] = Field(

None, alias="data_quality_monitoring"

)

""" DataQualityMonitoringConfig: Data Quality Monitoring configuration (optional). """

def __init__(self, **data: Any):

super().__init__(**data)

self._registry = None

if "registry" not in data:

raise FeastRegistryNotSetError()

self.registry_config = data["registry"]

self._offline_store = None

self.offline_config = data.get("offline_store", "dask")

self._online_store = None

self.online_config = data.get("online_store", "sqlite")

self._auth = None

if "auth" not in data:

self.auth = dict()

self.auth["type"] = AuthType.NONE.value

else:

self.auth = data.get("auth")

self._batch_engine = None

if "batch_engine" in data:

self.batch_engine_config = data["batch_engine"]

elif "batch_engine_config" in data:

self.batch_engine_config = data["batch_engine_config"]

else:

# Defaults to using local in-process materialization engine.

self.batch_engine_config = "local"

if isinstance(self.feature_server, Dict):

self.feature_server = get_feature_server_config_from_type(

self.feature_server["type"]

)(**self.feature_server)

# Initialize OpenLineage configuration

self._openlineage: Optional[OpenLineageConfig] = None

if "openlineage" in data:

self.openlineage_config = data["openlineage"]

# Initialize MLflow configuration

self._mlflow = None

if "mlflow" in data:

self.mlflow_config = data["mlflow"]

if self.entity_key_serialization_version < 3:

warnings.warn(

"The serialization version below 3 are deprecated. "

"Specifying `entity_key_serialization_version` to 3 is recommended.",

DeprecationWarning,

)

@property

def registry(self):

if not self._registry:

if isinstance(self.registry_config, Dict):

if "registry_type" in self.registry_config:

self._registry = get_registry_config_from_type(

self.registry_config["registry_type"]

)(**self.registry_config)

else:

# This may be a custom registry store, which does not need a 'registry_type'

self._registry = RegistryConfig(**self.registry_config)

elif isinstance(self.registry_config, str):

# Let Registry.__init__ auto-detect the correct store class

# from the URI scheme (e.g. gs:// -> GCSRegistryStore).

# Previously this hardcoded "file" type, which broke gs:// and

# s3:// paths because FileRegistryStore uses pathlib.Path.

self._registry = RegistryConfig(path=self.registry_config)

elif self.registry_config:

self._registry = self.registry_config

return self._registry

@property

def offline_store(self):

if not self._offline_store:

if isinstance(self.offline_config, Dict):

self._offline_store = get_offline_config_from_type(

self.offline_config["type"]

)(**self.offline_config)

elif isinstance(self.offline_config, str):

self._offline_store = get_offline_config_from_type(

self.offline_config

)()

elif self.offline_config:

self._offline_store = self.offline_config

return self._offline_store

@property

def auth_config(self):

if not self._auth:

if isinstance(self.auth, Dict):

config_type = (

"oidc_client"

if _is_oidc_client_config(self.auth)

else self.auth.get("type")

)

self._auth = get_auth_config_from_type(config_type)(**self.auth)

elif isinstance(self.auth, str):

self._auth = get_auth_config_from_type(self.auth)()

elif self.auth:

self._auth = self.auth

return self._auth

@property

def online_store(self):

if not self._online_store:

if isinstance(self.online_config, Dict):

self._online_store = get_online_config_from_type(

self.online_config["type"]

)(**self.online_config)

elif isinstance(self.online_config, str):

self._online_store = get_online_config_from_type(self.online_config)()

elif self.online_config:

self._online_store = self.online_config

return self._online_store

@property

def batch_engine(self):

if not self._batch_engine:

if isinstance(self.batch_engine_config, Dict):

self._batch_engine = get_batch_engine_config_from_type(

self.batch_engine_config["type"]

)(**self.batch_engine_config)

elif isinstance(self.batch_engine_config, str):

self._batch_engine = get_batch_engine_config_from_type(

self.batch_engine_config

)()

elif self.batch_engine_config:

self._batch_engine = self._batch_engine

return self._batch_engine

@property

def openlineage(self) -> Optional[OpenLineageConfig]:

"""Get the OpenLineage configuration."""

if not self._openlineage:

if isinstance(self.openlineage_config, Dict):

self._openlineage = OpenLineageConfig(**self.openlineage_config)

elif self.openlineage_config:

self._openlineage = self.openlineage_config

return self._openlineage

@property

def mlflow(self):

"""Get the MLflow configuration."""

if not self._mlflow:

if isinstance(self.mlflow_config, Dict):

from feast.mlflow_integration.config import MlflowConfig

self._mlflow = MlflowConfig(**self.mlflow_config)

elif self.mlflow_config:

self._mlflow = self.mlflow_config

return self._mlflow

@model_validator(mode="before")

def _validate_auth_config(cls, values: Any) -> Any:

from feast.permissions.auth_model import AuthConfig

if "auth" in values:

if isinstance(values["auth"], Dict):

if values["auth"].get("type") is None:

raise ValueError(

f"auth configuration is missing authentication type. Possible values={ALLOWED_AUTH_TYPES}"

)

elif values["auth"]["type"] not in ALLOWED_AUTH_TYPES:

raise ValueError(

f"auth configuration has invalid authentication type={values['auth']['type']}. Possible "

f"values={ALLOWED_AUTH_TYPES}"

)

elif isinstance(values["auth"], AuthConfig):

if values["auth"].type not in ALLOWED_AUTH_TYPES:

raise ValueError(

f"auth configuration has invalid authentication type={values['auth'].type}. Possible "

f"values={ALLOWED_AUTH_TYPES}"

)

return values

@model_validator(mode="before")

def _validate_online_store_config(cls, values: Any) -> Any:

# This method will validate whether the online store configurations are set correctly. This explicit validation

# is necessary because Pydantic Unions throw very verbose and cryptic exceptions. We also use this method to

# impute the default online store type based on the selected provider. For the time being this method should be

# considered tech debt until we can implement https://github.com/samuelcolvin/pydantic/issues/619 or a more

# granular configuration system

# Set empty online_store config if it isn't set explicitly

if "online_store" not in values:

values["online_store"] = dict()

# Skip if we aren't creating the configuration from a dict or online store is null or it is a string like "None" or "null"

if not isinstance(values["online_store"], Dict):

if isinstance(values["online_store"], str) and values[

"online_store"

].lower() in {"none", "null"}:

values["online_store"] = None

return values

# Set the default type

# This is only direct reference to a provider or online store that we should have

# for backwards compatibility.

if "type" not in values["online_store"]:

values["online_store"]["type"] = "sqlite"

online_store_type = values["online_store"]["type"]

# Validate the dict to ensure one of the union types match

try:

online_config_class = get_online_config_from_type(online_store_type)

online_config_class(**values["online_store"])

except ValidationError as e:

raise e

return values

@model_validator(mode="before")

@classmethod

def _validate_offline_store_config(cls, values: Any) -> Any:

# Set empty offline_store config if it isn't set explicitly

if "offline_store" not in values:

values["offline_store"] = dict()

# Skip if we aren't creating the configuration from a dict

if not isinstance(values["offline_store"], Dict):

return values

# Set the default type

if "type" not in values["offline_store"]:

values["offline_store"]["type"] = "dask"

offline_store_type = values["offline_store"]["type"]

# Validate the dict to ensure one of the union types match

try:

offline_config_class = get_offline_config_from_type(offline_store_type)

offline_config_class(**values["offline_store"])

except ValidationError as e:

raise e

return values

@model_validator(mode="before")

@classmethod

def _validate_feature_server_config(cls, values: Any) -> Any:

# Having no feature server is the default.

if "feature_server" not in values:

return values

# Skip if we aren't creating the configuration from a dict

if not isinstance(values["feature_server"], Dict):

return values

defined_type = values["feature_server"].get("type", "local")

values["feature_server"]["type"] = defined_type

# Validate the dict to ensure one of the union types match

try:

feature_server_config_class = get_feature_server_config_from_type(

defined_type

)

feature_server_config_class(**values["feature_server"])

except ValidationError as e:

raise e

return values

@field_validator("project")

@classmethod

def _validate_project_name(cls, v: str, info: ValidationInfo) -> str:

# Deferred import to avoid circular dependency during package initialization.

from feast.repo_operations import is_valid_name

sqlite_compatible = False

online_store = info.data.get("online_store") if info else None

if online_store is None or online_store == {}:

sqlite_compatible = True

elif isinstance(online_store, dict):

sqlite_compatible = online_store.get("type", "sqlite") == "sqlite"

if not is_valid_name(v):

raise ValueError(

f"Project name, {v}, should only have "

f"alphanumerical values, underscores, and hyphens but not start with an underscore or hyphen."

)

if sqlite_compatible and "-" in v:

raise ValueError(

"Project names for SQLite online stores cannot contain hyphens because they are used in table names."

)

return v

@field_validator("flags")

@classmethod

def _validate_flags(cls, v: Optional[dict]) -> Optional[dict]:

if not isinstance(v, dict):

return v

_logger.warning(

"Flags are no longer necessary in Feast. Experimental features will log warnings instead."

)

return v

def write_to_path(self, repo_path: Path):

config_path = repo_path / "feature_store.yaml"

with open(config_path, mode="w") as f:

yaml.dump(

yaml.safe_load(

self.json(

exclude={"repo_path"},

exclude_unset=True,

)

),

f,

sort_keys=False,

)

model_config = ConfigDict(populate_by_name=True)

class FeastConfigError(Exception):

def __init__(self, error_message, config_path):

self._error_message = error_message

self._config_path = config_path

super().__init__(self._error_message)

def __str__(self) -> str:

return f"{self._error_message}\nat {self._config_path}"

def __repr__(self) -> str:

return (

f"FeastConfigError({repr(self._error_message)}, {repr(self._config_path)})"

)

def get_data_source_class_from_type(data_source_type: str):

module_name, config_class_name = data_source_type.rsplit(".", 1)

return import_class(module_name, config_class_name, "DataSource")

def get_registry_config_from_type(registry_type: str):

# We do not support custom registry's right now

if registry_type not in REGISTRY_CLASS_FOR_TYPE:

raise FeastRegistryTypeInvalidError(registry_type)

registry_type = REGISTRY_CLASS_FOR_TYPE[registry_type]

module_name, registry_class_type = registry_type.rsplit(".", 1)

config_class_name = f"{registry_class_type}Config"

return import_class(module_name, config_class_name, config_class_name)

def get_batch_engine_config_from_type(batch_engine_type: str):

if batch_engine_type in BATCH_ENGINE_CLASS_FOR_TYPE:

batch_engine_type = BATCH_ENGINE_CLASS_FOR_TYPE[batch_engine_type]

else:

assert batch_engine_type.endswith("Engine")

module_name, batch_engine_class_type = batch_engine_type.rsplit(".", 1)

config_class_name = f"{batch_engine_class_type}Config"

return import_class(module_name, config_class_name, config_class_name)

def get_online_config_from_type(online_store_type: str):

if online_store_type in ONLINE_STORE_CLASS_FOR_TYPE:

online_store_type = ONLINE_STORE_CLASS_FOR_TYPE[online_store_type]

elif not online_store_type.endswith("OnlineStore"):

raise FeastOnlineStoreInvalidName(online_store_type)

module_name, online_store_class_type = online_store_type.rsplit(".", 1)

config_class_name = f"{online_store_class_type}Config"

return import_class(module_name, config_class_name, config_class_name)

def get_auth_config_from_type(auth_config_type: str):

if auth_config_type in AUTH_CONFIGS_CLASS_FOR_TYPE:

auth_config_type = AUTH_CONFIGS_CLASS_FOR_TYPE[auth_config_type]

elif not auth_config_type.endswith("AuthConfig"):

raise FeastInvalidAuthConfigClass(auth_config_type)

module_name, online_store_class_type = auth_config_type.rsplit(".", 1)

config_class_name = f"{online_store_class_type}"

return import_class(module_name, config_class_name, config_class_name)

def get_offline_store_type(offline_store_type: str):

if offline_store_type in OFFLINE_STORE_CLASS_FOR_TYPE:

return OFFLINE_STORE_CLASS_FOR_TYPE[offline_store_type]

elif not offline_store_type.endswith("OfflineStore"):

raise FeastOfflineStoreInvalidName(offline_store_type)

return offline_store_type

def get_offline_config_from_type(offline_store_type: str):

offline_store_type = get_offline_store_type(offline_store_type)

module_name, offline_store_class_type = offline_store_type.rsplit(".", 1)

config_class_name = f"{offline_store_class_type}Config"

return import_class(module_name, config_class_name, config_class_name)

def get_feature_server_config_from_type(feature_server_type: str):

# We do not support custom feature servers right now.

if feature_server_type not in FEATURE_SERVER_CONFIG_CLASS_FOR_TYPE:

raise FeastFeatureServerTypeInvalidError(feature_server_type)

feature_server_type = FEATURE_SERVER_CONFIG_CLASS_FOR_TYPE[feature_server_type]

module_name, config_class_name = feature_server_type.rsplit(".", 1)

return import_class(module_name, config_class_name, config_class_name)

def load_repo_config(repo_path: Path, fs_yaml_file: Path) -> RepoConfig:

config_path = fs_yaml_file

with open(config_path) as f:

raw_config = yaml.safe_load(os.path.expandvars(f.read()))

try:

c = RepoConfig(**raw_config)

c.repo_path = repo_path

return c

except ValidationError as e:

raise FeastConfigError(e, config_path)