Refactor `SAPBW` source

[angelika233]

• [ ] Check if _apply_user_mapping() is needed and if possible replace it with df.rename(columns=mapping_dict).

• [ ] api_connection() should be refactored into to_records() or just included as part of to_df()

• [ ] add unit tests

https://github.com/dyvenia/viadot/pull/1053#discussion_r1778744268

[angelika233]

remove mapping_dict from to_df()

[trymzet]

Seems like this _apply_mapping is actually doing 2 things: first it filters a dataframe, and then renames the columns. Eg. if we have a filter + mapping dict like this: {"my_col": "my_new_col_name"} and a df with 2 columns, "my_col" and "my_other_col", after applying this function, we'd get a df with 1 column, "my_new_col_name".

So I guess this function is confusing for 3 reasons:

• it does 2 things
• the naming is misleading, as it only mentions one of the two things it's doing
• the docstring doesn't explain what this func is doing, ony uses a generic term "applies mapping"

So it should probably be split into 2 functions or just native dataframe functions used (filtering a df and renaming columns are done with a single line of code in core pandas, so it's likely wrappers are not needed).