Small implementations and clean-ups

After the discussion with @ania , we agree on incorporating some small changes and additional methods, to facilitate additional downstream tasks:

• When invoking the dataset creation, if it exists:
  • In prompts an error
  • But, we include a flag to override this, delete the previous and create the new one
• We need a function to use the AnalysisCallback on newly generated geometries.
  • This is intended to compute the design error
• Ensure the categorical in the performance attributes works.
• When using the generator, the type returned is correct. Now, returning float for the constellations for example.
• The AnalysisCallback will have an option to indicate the input/output format\
• Integrate the methods inside the generator
  • In the generator, a flag like “df”, “dict”, and default it provides the full thing
• Remove dataset from Generator
• Method inside cae Model to recreate DataModule using the dataset
• Write a Forward_evaluation and Inverse_design, with nice input (as df or numpy), and nice output (as df or numpy)

Some bugs to check

• Generator is providing exact numbers on some attributes not requested
• Error in the Generation plot of generator designs
• Printing of results from Generator