Featurebyte

+ `service` fillna operation on a Target object now preserves the original target name automatically
+ `credential` Add support for key-pair authentication for Snowflake feature store
+ `api` Support creating batch feature tables from source tables and views without creating batch request tables.
+ `middleware` Merging ExecutionContext and ExceptionMiddleware functionality
+ `api` Introduce AddTimestampSchema column cleaning operation
+ `target` Introduce a target_type attribute to the target object, allowing explicit specification and updates of target prediction types.
+ `session` Sort database object listings in lexical order

⚠️ Deprecations

+ `python` Deprecating python version 3.9

🐛 Bug Fixes

+ `service` Fix an error when joining TimeSeriesView with SCDView in Snowflake due to timezone handling
+ `service` Fix feature version number generation bug when a previous version is deleted in the same day
+ `service` Fix grouping of calendar aggregation features when materializing in batch
+ `service` Fix handling of effective timestamp schema when serving parent features and deployment
+ `session` Fix large dates (e.g. 9999-01-01) causing source table preview to fail
+ `service` Fix concurrent materialization of feature lists with overlapping features
+ `service` Fix syntax error due to malformed DATEDIFF expressions
+ `session` Fix fallback calls for table and schema listing when user has not access to catalog INFORMATION_SCHEMA
+ `service` Fix SCD joins to take end timestamp column into account when available

🛑 Breaking Changes

+ `api` Make `fill_value` mandatory argument in `forward_aggregate`, `forward_aggregate_asat` and `as_target`.

💡 Enhancements

+ `service` Updates observation table construction logic to store invalidate rows from the observation table.
+ `service` Support timestamp schema for event timestamp column
+ `service` Support cron based default feature job setting for EventTable
+ `service` Introduce feature type to feature namespace.

💡 Enhancements

+ `warehouse_session` SessionManager caching logic is refactored to be in SessionManagerService
Deprecation of SessionManager in its entirety in favor of SessionManagerService
+ `warehouse_session` Simplification of FeatureStoreController DWH creation logic
+ `route` Add route to preview a feature from a historical feature table.
+ `session` Add support to configure max query concurrency for a feature store
+ `service` Implementing UserGroups for credential management
+ `task_manager` Add support for re-submitting failed jobs in the task manager
+ `service` Add validation for event ID column dtypes between item and event table during item table registration.
+ `warehouse_session` Deprecating of SessionManager instance variables
+ `api` Add overwrite parameter to file download methods in the SDK.
+ `session` Enforce application-wide concurrency limit on long running SQL execution
+ `api` Add support for sampling by time range during registration of observation tables from source table or view.
+ `service` Add data structures, routes and api for creation and manipulation of new Time Series table type
Feature creation functionality for Time Series table type to be added later
+ `service` Improve tile cache working table query efficiency in BigQuery
+ `warehouse_session` Merging functionality of SessionManagerService and SessionValidatorService
+ `service` Support non-timestamp special columns in SCDTable
+ `session` Cancel the query when interrupted by timeout or asyncio task cancellation.
+ `service` Prevent users to tag multiple columns with the same entity in the same table.
+ `service` Entity tags will be automatically removed when the table is deprecated.
+ `service` Implemented a consistency check to ensure entity data types are consistent between associated tables.

⚠️ Deprecations

+ `service` Merging MongodbBackedCredentialProvider with CredentialService

🐛 Bug Fixes

+ `service` Fixed an error in historical feature materialization when performing min/max aggregations on timestamp columns.
+ `service` Fixed overflow error for specific UDFs in Databricks SQL Warehouse.
+ `service` Fix computation of complex count distinct aggregation features involving multiple windows
+ `model` Move the `store_info` attribute from feature list model to deployment model.
+ `service` Fix scheduled tile task error due to missing `tile_compute_query`
+ `service` Fixed online serving code using serving entities from feature list instead of deployment.
+ `service` Fix entity untagging bug when removing ancestor IDs that can be reached by multiple paths
+ `service` Fix observation table creation failure when using a datetime object with timezone as sample_from_timestamp or sample_to_timestamp.
The created observation table is also filtered to exclude rows where POINT_IN_TIME is too recent for | feature materialization, or where mapped columns have NULL values.

💡 Enhancements

+ `session` Add support for BigQuery as a feature store backend
+ `service` Update ItemTable.get_view() method to auto-resolve column name conflicts by removing the conflicting column from the event table
+ `dependencies` Bump vulnerable dependencies.
+ jupyterlab to `^4.2.5`
+ aiohttp to `^3.10.2`
+ `service` Tighten asset name length validation in the service from 255 to 230 characters
+ `service` Allow some special columns (event_id_column, item_id_column, natural_key_column) to be optional during table registration.
+ `service` Support VARCHAR columns with a maximum length (previously detected as UNKNOWN)

🐛 Bug Fixes

+ `websocket` Fixes issue with websocket connection not disconnecting properly

🛑 Breaking Changes

+ `service` Remove default values for mandatory arguments in aggregation methods such as aggregate_over.
The value_column parameter must now be provided.

💡 Enhancements

+ `linting` Using ruff as the linter for the project
+ `package` Upgrade Pydantic to V2

⚠️ Deprecations

+ `dependencies` Deprecation of `sasl` library to support python 3.11

🐛 Bug Fixes

+ `service` Fix feature metadata extraction throwing KeyError during feature info retrieval
+ `session` Handle schema and table listing for catalogs without information schema in DataBricks Unity.

💡 Enhancements

+ `service` Validate aggregation method is supported in more aggregation methods
+ `service` Added support for count_distinct aggregation method

🐛 Bug Fixes

+ `numpy` Explicitly Set lower bound for numpy version to <2.0.0