Pad .zip attribute with 0 and ensure proper loading from court list

[nonprofittechy]

Fix #840

Currently, when you use the court_loader class in al_courts.py, there's a chance that a column that has a zip code with leading zeros will be interpreted incorrectly.

This fixes it in 2 ways:

1. transforms the address_zip column inside the dataframe when it is loaded to zerofill it to at least 5 digits
2. when printing a zip code (.zip attribute) as part of an address, zerofill to 5 digits.

It also adds an API to define a transformation for arbitrary columns of the Excel spreadsheet: defining the .converters attribute of the ALCourtLoader will now automatically run the converters when the dataframe is loaded. This honors the standard Pandas syntax, where converters should be a mapping between the column name and a callable with one parameter.

Ex:

---
objects:
  - all_courts: |
      ALCourtLoader.using(file_name='courts_list.xlsx', converters={"address_zip": float})

[samglover]

The logic of this fix makes sense and the ALKiln error doesn't seem to relate to it. (Didn't mean to close; clicked the wrong button to comment. 😬)