Sea-serpent

This is a major release with many new goodies, improvements and fixes.

New:
- added `Column.freeze()`, `Column.unfreeze()`, `Column.resize()` and `Column.value_counts()` methods
- added `Table.server_info` property
- added `Table.fetch_row_logs` and `Table.delete` methods
- added `BundleEdits` context manager which can be used to bundle edits before upload (faster)
- added admin functions: `list_users`, `find_user`, `share_base`, `list_common_datasets`, `import_common_datasets`, `list_script_tasks`

Improved:
- error messages when trying to write wrong data type to table
- comparisons with categorical data
- indexing table now works with pandas arrays and boolean dtypes
- filter queries using checkbox now work without the need to use `table.checkbox_column == True`
- `Column.contains` can now be used on single- and multi-select columns
- improved `Table.delete_rows()` and `Table.from_frame()` methods

Changed:
- `Table.row_ids` now always returns strings

Fixed:
- `<=` (`__le__` method) filter queries were broken up until this release
- previously environment variables `SEATABLE_SERVER` and `SEATABLE_TOKEN` were ignored under certain circumstances, this is fixed now

This is a bug fix release.

- new: `Table.get_view` (works with most filter settings)
- improved `Table.get_logs`
- `Table.loc` now accepts row IDs (strings)
- functions that return `DataFrame` should now always use the row IDs as index

Lots of additions since the last release. Here are some highlights:

- you can now ask `value in table.column` (translates via SQL query - i.e. no data loading required)
- the `.loc` and `[]` is now (even) more similar to that of pandas
- automatic checks for 64bit integer issues (when creating and writing to tables)
- you can now access shared tables
- added some convenience functions/shorthands to columns: `.astype()`, `.map()`, `.contains()`, `startswith()` and `endswith()` (the latter three are string-related)

New stuff:
- `Table.delete_rows()` method
- parse dates for reading and writing
- `Column.add_options` method for single- and multi-select columns
- for "text" and "long text" columns: parse `"None"` and `""` strings to `None`