Use jinja2 templating to update library to unicode 15.0, jquast#61

- update `tox.ini`, `requirements*` files
  - use `pip-compile` with "requirements.in" files to generate "requirements.txt" files
  - use older testing tools for older python versions
  - ensure for python 3.10 and 3.11
- bin/update-tables.py changes,
  - use jinja2 (`.j2`) templating to generate code and documentation,
    - easier to support code generation of other languages
  - use https and perform retries
  - build cache uses HEAD to compare last-modified for refresh
- `for wcwidth/table*.py` files
  - uses new template system to generate new table data files for unicode version 14 and 15
  - when a description of unicode is single codepoint, describe only as 'A' instead of 'A..A'

Do not depend on pkg_resources module (jquast#42)

Generate code for `wcwidth/unicode_versions.py`, delete `wcwidth/version.json`, and all references to it.

Do not call _get_package_version() on import

Hopefully, helps out an xonsh issue on fedora?

xonsh/xonsh#3607

0.2.3 to include tests in sdist

trailing ws

RST check

Unicode v12 and proj. maintenence (jquast#30)

* remove static analysis
* EastAsianWidth v9 -> v12
* README and tox.ini

add missing PR href

Merge pull request jquast#15 from jquast/license-doc-polish

0.1.6: Include LICENSE file in package distribution

brevity of README.rst and polish