Pdtable

Fixed

- Bug causing `write_excel` `engine_kwargs` arguments to not be passed on to the engine.

Added

- Support for writing Excel files with `xlsxwriter`.

Fixed

- Can now read empty tables (tables with only name and origin, no column or unit information).
- Fixes deprecation warning related to numpy type
- Fixes warning when calling `groupby` method

Fixed

- Issue in `make_table_dataframe` when using `unit_map`.

Added

- [106] Support for detailed origin information in table metadata.
- [106] Support for loading complete input sets.

Changed

- Table origin information has been moved from `TableMetadata` to an aggregate object.
- The error type raised by the `fixer` has been changed to integrate with `load` framework.

Fixed

- In response to a Numpy deprecation warning, changed one `dtype` argument from deprecated `numpy.str` to its designated successor, built-in `str`.

Changed

- `write_excel()` parameter `style` renamed to `styles`.
- Optionally specify custom styles in `write_excel()` by passing to parameter `styles` a JSON-like structure of dicts.
- When `write_excel()` parameter `styles` is truthy, transposed tables' column units and values are horizontally centered by default. This default is overridden by any horizontal alignment specified explicitly in `styles`.
- `write_excel()` parameter `num_blank_rows_between_tables` renamed to `sep_lines` for conciseness.

Fixed

- Replaced the one remaining reference to `numpy.bool` (now deprecated from numpy) with its designated successor, plain ol' built-in `bool`.