Typeddfs

Added

- Reading and writing dataset metadata (attrs)

Changed

- Improved `cli_help`
- Arguments in `write_file` and `read_file` are now keyword-only
- Small parts of `Checksums`

Fixed

- `write_file` checks for existing hashes and write permissions
before attempting to write
- extras "all" and "main"
- docstring rst issues, esp. broken links
- metadata always passed to custom exceptions

Added

- Some uncommon Excel suffixes
- `cli_help`
- `FileFormat.is_recommended`
- `FileFormat.matches`
- `recommended_only()` to builders

Fixed

- `CompressionFormat.strip`
- Deprecation warning in test
- matrix row and column names now always typed as str

Added

- Preview support for .properties, INI, and TOML
- Top-level imports `typed`, `untyped`, etc.
- Q & A in the readme

Changed

- `remap_suffixes()` renamed to `suffix()`
- `.subclass()` now supports multiple inheritance

Removed

Fixed

- `to_parquet()` doesn't change short to int

Added

- `CoreDf.strip_control_chars`
- `Utils.exact_natsort_alg`, `Utils.guess_natsort_alg`, and `Utils.all_natsort_flags`
- Functions from `pandas.api.types` to `Utils`
- `DfSupport.reload()`

Changed

- `DfTyping` is now generic
- `FrozeSet` and `FrozeDict` are now ordered
- Moved checksum utils to `checksums.py`
- Moved `DfSupport` to `_format_support.py`
- `sort_natural` now infers the best algorithm from the data type, by default
- `drop_cols` can now accept \*args
- Split `parse_hash_file` into `parse_hash_file_resolved` and `parse_hash_file_generic`
- `regex` is now a dependency
- Hashing options in `write_file`
- Some `Utils` and `FileFormat` params are keyword-only

Removed

- `Utils.verify_any_hash`
- Positional args from `ffill` and `bfill`

Fixed

- Bugs in `exact_natsort_alg`
- Small bugs in `FrozeDict` and `FrozeSet`

Added

- Dataclass conversion
- Utils for freezing types

Changed

- `from_records` now calls `convert`