Format mismatch in print_examples() prevents example collection for lexical and syntactic grammars #14
Description
The print_examples() function fails to find examples for lexical and syntactic grammars due to a format mismatch between construction learning and example collection phases.
During learning (process_grammar()):
Lexical grammar: converts enriched data to [(lex, -1, -1), ...]
Syntactic grammar: converts to [(-1, syn, -1), ...]
Full grammar: keeps as [(lex, syn, sem), ...]
During example collection (print_examples()):
All grammar types: uses unmodified enriched format [(lex, syn, sem), ...]
Without examples, get_token_similarity() fails (requires 75% character overlap between examples), causing all constructions to collapse into token cluster "1" regardless of their actual similarity.
Suggested fix: apply the same format masking in print_examples().