Spooq

-------------------
* [MOD] (BREAKING CHANGE!) rename package back to Spooq (dropping "2").
This means you need to update all imports from spooq2.xxx.yyy to spooq.xxx.yyy in your code!
* [MOD] Prepare for PyPi release
* [MOD] Drop official support for Spark 2.x (it most probably still works without issues,
but some tests fail on Spark2 due to different columns ordering and the effort is too high to
maintain both versions with respect to tests)

-------------------
* [MOD] Add functionality to log cleansed values into separate struct column (column_to_log_cleansed_values)
* [MOD] Add ignore_ambiguous_columns to Mapper
* [MOD] Log spooq version when importing
* [REM] Drop separate spark package (bin-folder) as pip package can now handle all tests as well
* [ADD] Github action to test on label (test-it) or merge into master

-------------------
* [ADD] EnumCleaner Transformer
* [MOD] Add support for dynamic default values with the ThresholdCleaner

-------------------
* [MOD] extended_string_to_timestamp: now keeps milli seconds (no more cast to LongType) for conversion to Timestamp

-------------------
* [ADD] Spark 3 support (different handling in tests via `only_sparkX` decorators)
* [FIX] Fix null types in schema for custom transformations on missing columns
* [MOD] (BREAKING CHANGE!) set default for `ignore_missing_columns` of Mapper to False (fails on missing input columns)

------------------
* [MOD] extended_string_to_timestamp: it can now handle unix timestamps in seconds and in milliseconds
* [MOD] extended_string_to_date: it can now handle unix timestamps in seconds and in milliseconds