Tablib

Bugfixes

- Support solo CR in text input imports (518).

Changes

- Dropped Python 3.6 support (513).

Bugfixes

- Corrected order of arguments to a regex call in `safe_xlsx_sheet_title` (510).

Improvements

- Add support for Python 3.10 (504).
- The csv, xls, and xlsx formats gained support for the `skip_lines` keyword
argument for their `import_set()` method to be able to skip the nth first
lines of a read file (497).

Bugfixes

- Avoided mutable parameter defaults (494).
- Specify build backend for editable installs (501).
- Doubled sample size passed to `csv.Sniffer()` in `_csv.detect()` (503).

Breaking changes

- Dropped Python 3.5 support.
- JSON-exported data is no longer forced to ASCII characters.
- YAML-exported data is no longer forced to ASCII characters.

Improvements

- Added Python 3.9 support.
- Added read_only option to xlsx file reader (482).

Bugfixes

- Prevented crash in rst export with only-space strings (469).

Breaking changes

- The `Row.lpush/rpush` logic was reversed. `lpush` was appending while `rpush`
and `append` were prepending. This was fixed (reversed behavior). If you
counted on the broken behavior, please update your code (453).

Bugfixes

- Fixed minimal openpyxl dependency version to 2.6.0 (457).
- Dates from xls files are now read as Python datetime objects (373).
- Allow import of "ragged" xlsx files (465).

Improvements

- When importing an xlsx file, Tablib will now read cell values instead of formulas (462).

Deprecations

- Upcoming breaking change in Tablib 2.0.0: the `Row.lpush/rpush` logic is reversed.
`lpush` is appending while `rpush` and `append` are prepending. The broken behavior
will remain in Tablib 1.x and will be fixed (reversed) in Tablib 2.0.0 (453). If you
count on the broken behavior, please update your code when you upgrade to Tablib 2.x.

Improvements

- Tablib is now able to import CSV content where not all rows have the same
length. Missing columns on any line receive the empty string (226).