Xmu

-----

- Added `clean_xml()` function to removed restricted characters from XML
- Added `sum` keyword for handling repeated keys in
`from_xml_parallel()`
- Changed handling of reference field views and field groups, including
group definitions with overlapping fields
- Removed date attribute setters from EMuDate

-----

- Transferred repository to NMNH-IDSC
- Added `from_xml_parallel()` method to EMuReader to allow parallel
reading of XML files (experimental)
- Added support for pickle and json to EMuRecord objects
- Added check for missing grid fields to the EMuReader and EMuGrid
objects. These methods are intended to allow users to catch when a
required grid field is missing from an export.
- Changed custom data types to be essentially immutable. Previously
attributes could be directly modified.
- Changed loading of schema and config files such that these files are
explicitly loaded when the module is imported. Previously they were
lazy loaded when an associated attribute was accessed.

-----

- Added support for Python 3.12
- Added functions to write CSV from XML files
- Changed EMuRecord to coerce NAs to None and to coerce empty nested
tables to empty lists
- Changed write_import() to write_xml(). The write_import() function
remains as an alias.
- Changed EMuCoord to catch out-of-bounds minutes and seconds
- Changed EMuDate to handle up to three arguments (date or (year, month,
day))
- Fixed bug identifying the module in a reference

-----

- Added support for Python 3.11 and removed support for Python 3.7
- Added check for sequence data in atomic fields
- Changed parsing of config files, including autoloading schemas
- Changed EMuCoord to better represent decimal coordinates
- Changed EMuFloat to accept floats with no formatting string
- Fixed bugs with EMuType comparisons, reading/writing XML, and
populating lookup list parents

-----

- Changed EMuDate to handle dates outside the MINYEAR and MAXYEAR
allowed in python’s datetime module. Dates outside that range can be
expressed but do not support most date operations.
- Changed EMuRecord.to_xml() and EMuColumn.to_xml() to (1) include blank
values when records are created, (2) to add blank values to parent
fields in a lookup hierarchy based a key in the config file, and (3)
to check for an emu_str() method when writing values. The emu_str()
method is used by some EMuType classes to allow greater control in how
values are depicted in imports.
- Fixed bugs selecting grid rows, parsing IRNs, reporting progress, and
marking normally inaccessible fields as visible.

-----

- Added EMuReader.report_progress() to notify users about progress
reading through an XML file
- Changed EMuColumn.to_xml() method so that table rows are numbered
sequentially using the row attribute if no modifiers are given. This
prevents the EMu client from skipping empty nested table rows in valid
XML during import.
- Changed behavior of EMuRow.\_\_getitem\_\_() when a dictionary is
passed to:
1. fix a bug where empty row values matched any query string
2. return only the row. Previously the method returned the index and
row as a tuple, but this was redundant as the index is accessible
directly from the row object itself.
- Changed modifier pattern to catch additional modifiers supported by
EMu. Acceptable formats include “=”, “+”, “-”, “1=”, “1+”, and “1-”.
- Renamed EMuRow mod attribute to replace_mod to clarify that it returns
the replacement modifier