Turboflow

- Update package to use `poetry` for dependency management
- Format all codebase using `black`

- Complete configuration check:
- config_validation is intended to check what turbomachinery configuration is given and validation configuration accordingly
- turbo_configurations folder gather all turbomchainery configuration files (only config_axial_turbine now)
- see any config validation file for new setup
- New way to define design variables
- Updated all config files according to new config_validation

- Tests for all model options and optimization psuhed
- Added script for checking configurations through pydantic

- The `variables` entries in the configuraion files have gotten default values. If the user now only give a partly set of variables, the remaining ones are given with a default value.

- Update implementation of objective function in configuration file. The objective funciton is now defined by providing the variable, scale, and type (minimize/maximize). Based on the type, the sign of the scal might change.
- The objective function is now fetched from the `results` structure. The variable must be on the form `key.column`, where `key` refer to a key in `results` (e.g. `overall`), and the column, refer to a column in the DataFrame contained in
- `results[key]` (e.g. `efficiency_ts`). The column will be an array, and the first element is used (for `overall`, all columns has one entry).
- Update implementation of constraints. Similarly as for the objective function, the constraint variable are now fethced from `results`, with the name `key.column`. An additional key is added to results: `additional_constraints`.
- This can contain variables that does not fit anywhere else in the results structure.
- A future idea is to restructure `results` to be an object, to have more control over the variables.

- Approximate slope of flow angle curve at critical point with parabolic function
- Use this to calculate the deviation bin the range 0.5 < Ma_exit < Ma_crit