Re-structure Flet Python packages (wheels) for optimized installation on various platforms

[FeodorFitsner]

To-Do:

• "desktop light" for Linux users with excluded Video, Audio (closes libmpv errors on Linux #2637).
• Make an alias for flet.fastapi to flet_web.fastapi.

More info:

• https://packaging.python.org/en/latest/specifications/dependency-specifiers/#extras
• https://peps.python.org/pep-0508/
• https://peps.python.org/pep-0345/

flet is the main and minimal package. Could be installed on any platform.

New packages:

• flet containing:
  • ...
• flet_core containing:
  • ...
• flet-web containing:
  • web directory with a compiled Flet web app.
  • flet_web.fastapi module
  • flet_web.auth module
  • flet_web.utils module
• flet-desktop containing:
  • bin - the full Flet client executable/bundle with all optional packages (Video, Audio, etc.). Platform dependent.
  • flet_desktop module with client download/unpack code.
• flet-desktop-light containing:
  • bin - Flet client executable/bundle without optional packages (Video, Audio, etc.). Provided for Linux only.
• flet-cli containing:
  • flet_cli module

Optional (extra) packages which are dependencies with constraints for flet package:

• flet-web - platform-independent (py3-none-any) package that contains "web" directory with a compiled Flet web app. Condition: platform_system == 'Windows' or platform_system == 'Linux' or platform_system == 'Darwin, 'embedded' not in platform_version'.
• flet-desktop - platform-specific package with Flet client executable/bundle without optional packages (Video, Audio, etc.). Condition: platform_system == 'Windows' or platform_system == 'Linux' or platform_system == 'Darwin, 'embedded' not in platform_version'. Additionally, we could skip installation of flet-desktop in Docker environment.
• flet-cli - platform-independent (py3-none-any) package that contains Flet CLI. Conditions similar to desktop.
• flet-desktop-full - could be optionally installed on Linux if MPV, Audio dependencies are met.

Examples: https://github.com/encode/uvicorn/blob/master/pyproject.toml