…le output more user-friendly (#8999)

#8978)

* Fix type annotation for buildSqlAndParams

* Rework TEST_CLICKHOUSE_HOST handling

* Port ClickHouseDbRunner to TypeScript

* Move _normaliseResponse to static method

* updates AI API reference

* adds model providers and meta

* fills out env vars

* adds some details

---------

Co-authored-by: Michael Irvine <[email protected]>

…eferencing measures from joined cubes (#8991)

* fix(schema-compiler): Fix incorrect sql generation for view queries referencing measures from joined cubes

* add tests

…ql push down (#9000)

* add single intervals transformation for databricks

* add CURRENTDATE support for Databricks SQL Push down

…hanges (#8985)

* Replaces Accumulator usage in grouped hash aggregation with
  GroupsAccumulator API

* Makes all accumulators in DF grouped hash aggregation use
  GroupsAccumulatorFlatAdapter

* Uses PrimitiveGroupsAccumulator and CountGroupsAccumulator for sum,
  min, max, and count for faster performance

* Improves memory layout and traversal order of group by keys saved
  during grouped aggregation.

…use of disk space (#9003)

…generation (#8981)

* fix(schema-compiler): Fix ungrouped cumulative queries incorrect sql generation

* add tests

* Revert "fix(schema-compiler): Fix ungrouped cumulative queries incorrect sql generation"

This reverts commit ca365fa.

* fix(schema-compiler): Fix ungrouped cumulative queries incorrect sql generation

* fix tests: have to add alias, because of too long identifiers

* automatically start the engine after connection

* add tests

* fix linter

* start engine in the end of the test

…lters in cube's SQL query (#8937)

* fix(schema-compiler): fix FILTER_PARAMS to populate set and notSet filters in cube's SQL query

* Fix {DATE_OPERATORS}Where methods to return '1 = 1' when the required parameters are unavailable

* fix(ksql-driver): Kafka, list of brokers (#9009)

* docs: A few notes

* chore(cubesql): Do not call async Node functions while planning (#8793)

* Add postgress integration tests

---------

Co-authored-by: Konstantin Burkalev <[email protected]>
Co-authored-by: Dmitriy Rusov <[email protected]>
Co-authored-by: Igor Lukanin <[email protected]>
Co-authored-by: Alex Qyoun-ae <[email protected]>

macOS 12 runner is unsupported since 2024-12-03
See actions/runner-images#10721

…8929)

Also disallow pulling up push-to-Cube wrapper over CubeScan with limit

#8775)

Extensions is used meaningfully only for CubeScan, but not for CubeScanWrapper. Both are first-class logical plan nodes for us, so Extension is not helpful in egraph

…preparation in UI side (#8934)

* Implement input data parsing in debugger

* Use simpler state as debug data source

* Enable serde for egg
Used just for Id for now

* Move data preparation to UI side of debugger

* Filter duplicate edges during preparation

* Filter non-canonical eclasses before rendering

…e. In order to query ksql create pre-aggregation first if Kafka download isn't enabled

…ore. In order to query ksql create pre-aggregation first if Kafka download isn't enabled -- remove unloadWithoutTempTable as it's not needed due to readOnly flag on rollups