Sync with changes merged after 6502c66025718bf45e0e2ee12398b7b92da41a0c

[zero323]

• SPARK-30377 Make Regressors extend abstract class Regressor - SPARK-30377 Make AFTSurvivalRegression / Model extend JavaPredictor / JavaPredict… #325
• SPARK-30351 BisectingKMeans support instance weighting - SPARK-30351 Add weightCol to BisectingKMeans #323
• SPARK-30498 Fix some ml parity issues between python and scala - SPARK-30498 Adjust for fixes introduced by SPARK-30498 #326
• SPARK-9478 Add sample weights to Random Forest - SPARK-9478 Add weightCol to RandomForest(Regressor|Classifier) #324
• SPARK-29565 add setInputCol/setOutputCol in OHEModel - [SPARK-29565][FOLLOWUP] Add setInputCol / setOutputCol to OneHotEncoderModel #335
• SPARK-30533 Add classes to represent Java Regressors and RegressionModels - [SPARK-30533] Add Regressor and RegressionModel #339
• SPARK-30539 Add DataFrame.tail in PySpark - [SPARK-30539] Add DataFrame.tail #338
• SPARK-28264 Support type hints in pandas UDF and rename/move inconsistent pandas UDF types - [SPARK-28264] Support udf-free operations #349
• SPARK-30607 Add overlay wrappers for SparkR and PySpark - [SPARK-30607] Add pyspark.sql.functions.overlay #347
• SPARK-30543 RandomForest add Param bootstrap to control sampling method - [SPARK-30543] Add bootstrap Param to RandomForest #348
• SPARK-30642 LinearSVC blockify input vectors [SPARK-30642] Add HasBlockSize Params and blockSize Param to LinearSVC #356
• SPARK-30662 ALS/MLP extend HasBlockSize - [SPARK-30662] Add blockSize to ALS #358, [SPARK-30662] Add HasBlockSize to _MultilayerPerceptronParams #359
• SPARK-29093 FOLLOW-UP Remove duplicate setter - [SPARK-29093][FOLLOW-UP] Drop duplicated methods from BucketedRandomProjectionLSH #360
• SPARK-30660 LinearRegression blockify input vectors - [SPARK-30660] Add HasBlockSize to LinearRegressionParams #361

[zero323]

All tasks have been resolved, and Spark 3.0 was cut.