Solvation data preparation

[craabreu]

This PR introduces two notebooks designed to prepare solvation data for publication:

1. The data are classified into three categories: transition_states, closed_shell_species, and open_shell_species
2. For each category, data points are split by solvent and stored in files called <solvent-name>.csv, distributed in subdirectories according to the first letter of the solvent's name (ignoring non-letter characters).
3. Reaction SMILES are stored without atom mappings and using canonical SMILES for each reactant/product.

Note: A file solvents.csv is included, which maps solvent names to canonical SMILES without stereochemical information. The solvent SMILES in the original solvation data is not in canonical form, thus requiring an intermediate canonicalization step.