Feldera

Added

- [SQL] CREATE LOCAL VIEW statement to declare views which are not outputs. ([1851](https://github.com/feldera/feldera/pull/1851))
- Documentation: added the [fraud detection use case](https://feldera.com/docs/use_cases/fraud_detection) to the docs section of the website.
- API: storage class can be configured via pipeline resources ([1829](https://github.com/feldera/feldera/pull/1829))
- [Python] Added support for Kafka connector via Python SDK ([1807](https://github.com/feldera/feldera/pull/1807))
- [Python] Added support for HTTP GET connector via Python SDK
- [Python] Add support for Avro format ([1836](https://github.com/feldera/feldera/pull/1836))
- [Python] allow specifying pipeline resource configuration and compilation profile via SQLContext ([1835](https://github.com/feldera/feldera/pull/1835))
- [Python] Add function to wait for the pipeline to become idle ([1841](https://github.com/feldera/feldera/pull/1841))
- WebConsole: DeltaLake input (1744) and output (1584) conenctors
- WebConsole: Add output buffer configuration UI (1591)

Fixed

- WebConsole: Fix crash when displaying anonymous output columns (1757)

Added

- Added backpressure mode to the `/egress` endpoint, which applies backpressure on the pipeline if the HTTP client cannot keep up
with the output instead of dropping data. ([1780](https://github.com/feldera/feldera/pull/1780))
- `/heap_profile` endpoint that generates the heap profile of a pipeline in the pprof format. ([1767](https://github.com/feldera/feldera/pull/1767))
- Python SDK for Feldera ([1745](https://github.com/feldera/feldera/pull/1745)).

Added

- [SQL] Support for user-defined functions implemented in SQL ([1714](https://github.com/feldera/feldera/pull/1714))
- [Kafka] Allow specifying Kafka headers as part of output Kafka connector
config ([1718](https://github.com/feldera/feldera/pull/1718))
- [SQL] Support for `LEAD`/`LAG` window aggregates ([1706](https://github.com/feldera/feldera/pull/1706))
- [SQL] Support for `WATERMARK` table column annotations ([1694](https://github.com/feldera/feldera/pull/1694))
- Added a new config flag `tcp_metrics_exporter` which allows to inspect
the pipeline metrics for debugging purposes with the `metrics-observer`
CLI. ([1710](https://github.com/feldera/feldera/pull/1710))
- Ability to selectively enable storage per-pipeline. ([1692](https://github.com/feldera/feldera/pull/1692))

Changed

- [Docs] Re-organize documentation sections ([1734](https://github.com/feldera/feldera/pull/1734))

Added

- Initial support for Avro as an output format ([1673](https://github.com/feldera/feldera/pull/1673))

Changed

- Fixed a bug where sometimes the wrong profile for compilation is chosen by the manager. ([1696](https://github.com/feldera/feldera/pull/1696))

Added

- [SQL] `ARG_MIN`, `ARG_MAX` aggregation functions
([1619](https://github.com/feldera/feldera/pull/1619))
- SQL: Support for `ARRAYS_OVERLAP` function ([1570](https://github.com/feldera/feldera/pull/1570))
- Connectors: Initial Delta Table output connector support
([1639](https://github.com/feldera/feldera/pull/1639))

Changed

- [SQL]: `ARRAY_CONTAINS`, `ARRAY_POSITION`, `ARRAY_REMOVE` functions
for arguments of different types ([1636](https://github.com/feldera/feldera/pull/1636))

Added

- Output buffering mechanism to consolidate multiple output
updates over time.
([1558](https://github.com/feldera/feldera/pull/1558))
- pipeline-manager: allow programs to be compiled in unoptimized or
optimized modes, which tradeoff compilation speed for runtime speed
([1524](https://github.com/feldera/feldera/pull/1524))
- WebConsole: Support NULL values on Data Inspection and Insertion page (1392)

Changed

- pipeline-manager: connector transport configuration in the API is now
strongly typed, and the transport names have the `_input` and `_output`
suffix added ([1532](https://github.com/feldera/feldera/pull/1532))

Fixed

- WebConsole: State shows InQueue when program is empty (1443)