Cmdstanpy

- Fixed a critical bug in how outputs containing complex numbers were read in by CmdStanPy. If you are using models which output complex numbers, please update immediately to ensure you are getting the correct results.

- Fixed a build error in the documentation
- Improved messages when model fails to compile due to C++ errors.

- Fixed a typo in `cmdstanpy.show_versions()`
- Reorganized and updated the documentation
- Reorganized a lot of internal code
- Cleaned up the output of `CmdStanMCMC.draws_pd`
- Cleaned up the output of `CmdStanMCMC.summary`
- Removed the logging which occurred when Python exited with cmdstanpy imported.

- Fix an issue with the `cmdstanpy.install_cmdstan()` function where the installation would report that it had failed even when it had not.

- Fix an issue where Stan fit objects were not `pickle`-able when they previously were.
Note: We still do not recommend pickling cmdstanpy objects, but rather using functions `save_csvfiles` and `from_csv`.

- CmdStanPy can now format (and canonicalize) your Stan files with `CmdStanModel.format()`
- Stan variables can now be accessed from fit objects using the `.` syntax when no naming conflicts occur. For example, previous code `fit.stan_variable("my_cool_variable")` can now be written `fit.my_cool_variable`
- CmdStanPy is more robust to running in threaded environments and tries harder to not overwrite its own output files
- The `install_cmdstan` script can now be run in interactive mode using `--interactive`/`-i`
- CmdStanPy now computes some diagnostics after running HMC and will warn you about post-warmup divergences and treedepth exceptions
- Runtime exceptions in the `generated quantities` block should be recognized better now.
- The default level of precision to `fit.summary()` is now 6, as it is when `stansummary` is used from the command line.\
- Various documentation improvements