nx_entity_seeds.nx

nx_entity_seeds.nx -- SEED DISCOVERY: turn one entity page into the list of OTHER pages about that entity. Why this exists: our own index holds exactly ONE page for some entities, so search cannot supply seeds -- a measured dead end, not a guess. Seeds must come from STRUCTURE instead: an entity's page links out to the other pages about it (filmography, official site, profiles, references). This fetches a page over sovereign TLS (dechunk -> gunzip, same wire order as the gatherer), enumerates every <a href>, resolves them absolute, and keeps the ones whose URL carries a token of the entity's name. The output is a plain seed list, one URL per line, so the gatherer can be run once per seed. Cross-seed dedupe is FREE and needs no state: assets are sha256 content-addressed, so the same image reached from two different seeds writes the same filename. usage: nx_entity_seeds <page-url> <name-token> <out-file> [max]