Spooq

------------------
* [MOD] Add support for prepending and appending mappings on input dataframe (Mapper)
* [MOD] Add support for custom spark sql functions in mapper without injecting methods
* [MOD] Add support for "on"/"off" and "enabled"/"disabled" in extended_string_to_boolean custom mapper transformations
* [ADD] New custom mapper transformations:

- extended_string_to_date
- extended_string_unix_timestamp_ms_to_date
- has_value

------------------
* [MOD] `drop_rows_with_empty_array` flag to allow keeping rows with empty array after explosion
* [MOD] Additional test-cases for extended_string mappings (non string inputs)
* [FIX] Remove STDERR logging, don't touch root logging level anymore (needs to be done outside spooq to see some lower log levels)
* [ADD] New custom mapper transformations:

- extended_string_unix_timestamp_ms_to_timestamp

------------------
* [ADD] Python 3 support
* [MOD] `ignore_missing_columns` flag to fail on missing input columns with Mapper transformer (https://github.com/Breaka84/Spooq/pull/6)
* [MOD] Timestamp support for threshold cleaner
* [ADD] New custom mapper transformations:

- meters_to_cm
- unix_timestamp_ms_to_spark_timestamp
- extended_string_to_int
- extended_string_to_long
- extended_string_to_float
- extended_string_to_double
- extended_string_to_boolean
- extended_string_to_timestamp

------------------
* [UPDATE] Upgrade to use Spark 2 (tested for 2.4.3) -> will no longer work for spark 1
* Breaking changes (severe refactoring)

------------------
* [FIX] Logger writes now to std_out and std_err & logger instance is shared across all spooq instances
* [FIX] PyTest version locked to 3.10.1 as 4+ broke the tests
* [MOD] Removes id_function to create names for parameters in test methods (fallback to built-in)
* [ADD] Change SelectNewestByGroup from string eval to pyspark objects
* [FIX] json_string is now able to None values

------------------
* [FIX] PassThrough Extractor (input df now defined at instantiation time)
* [ADD] json_string new custom data type