Nbstripout

- Fix regression where input file was truncated before reading (190).

- Drop support for Python 3.7 (end of life 2023-06-27), require Python 3.8.
- Add support for Python 3.12.
- Drop backwards compability with IPython.nbformat for IPython <4.
- Rename cell ids to be sequential by default. Disable by passing `--keep-id` (JasonJooste, 184).
- Improve documentation for notebook and cell metadata stripping (187).
- Switch from pytest-flake8 to pytest-ruff.
- Convert all text assets from rST to Markdown format.

* Add `--python` option for `nbstripout --install` to allow overriding the Python interpreter specified in `.git/config` (nobodyinperson, 181, 182)
* Add option `--keep-metadata-keys` to keep specific metadata keys that are stripped by default (davidxia, 177, 78)
* Replace [Cram](https://bitheap.org/cram/) as test runner for integration tests with a custom framework which also supports testing on Windows (arobrien, #176, 178)
* Use `SystemExit` instead of `sys.exit` and do not exit from functions (janosh, 173)

* Removed `setup_requires` and `tests_require` and no longer rely on `pytest-runner` for test execution but invoke `pytest` directly (168)

* Support for stripping init cells (Pugio, 157)
* Added `--drop-tagged-cells="some tags"` option (boun, 161)
* Renamed `--strip-empty-cells` to `--drop-empty-cells`
* Dropped support for Python 3.5, added support for Python 3.10

* Support only stripping outputs larger than a given size (cjblocker, 135)
* Support stripping output from Zeppelin Notebooks (ankitrokdeonsns, 130)
* Switch CI to GitHub actions (151)
* Support attributes file without leading path component (155)