Replace the pyMBE Pandas dataframe for a canonical pyMBE database

[pm-blanco]

Fixes #146, #123 and #18 :

Added

• Introduced a canonical pyMBE database backend replacing the previous monolithic Pandas DataFrame storage approach. This lays the foundation for more robust, extensible, and normalized data handling across pyMBE.
• Added support to define reaction templates in the pyMBE database.
• Utility functions to cast information about templates and instances in the pyMBE database into pandas dataframe pmb.get_templates_df, pmb.get_instances_df and pmb.get_reactions_df.
• Utility functions to load and save the new database via the pyMBE API, pmb.save_database and pmb.load_database.
• Added functions to define particle states: pmb.define_particle_states and pmb.define_monoprototic_particle_states.
• Added utility functions in lib/handy_functions to define residue and particle templates for aminoacids en peptides .and residues: define_protein_AA_particles, define_protein_AA_residues and define_peptide_AA_residues.

Changed

• Refactored core modules to use the new database schema based on templates and instances for particles, residues, molecules, hydrogels, proteins and peptides.
• Particle states now are independent templates and are now disentangled from particle templates.
• Pka values are now stored as part of chemical reactions and no longer an attribute of particle templates.
• Amino acid residue templates are no longer defined internally in define_peptide and define_protein. Those definitions are now exposed to the user.
• Molecule templates now need to be defined to be used as templates for hydrogel chains in hydrogels.

Fixed

• Wrong handling of units in get_radius_map when the dimensionless argument was triggered.
• Utility methods get_particle_id_map, calculate_HH, calculate_net_charge, center_object_in_simulation_box now support all template types in pyMBE, including hydrogels. Some of these methods have been renamed to expose directly in the API this change in behavior.

Removed

• Method add_bonds_to_espresso has been removed from the API. pyMBE now adds bonds internally to ESPResSo when molecule instances are created into ESPResSo.
• Tutorial lattice_builder.ipynb has been removed because its content is redundant with sample script build_hydrogel.py.

Documentation

• Cleaned and standardized documentation.

Tests

• Modernized molecule tests and LJ tests.

[1234somesh]

This massive PR brings systematic and safe handling of dataframe. Now mananger.py handles creation of templates and instances classes whenever define_* and create_* of any pyMBE entity, while enforcing validation.