Resolve an entity name or alias to its article stem via the alias index.
Normalizes the query the same way aliases are stored (casefold, ws-collapse)
and also retries with a leading article word ("the"/"a"/"an") stripped, so
a "" surface and a "" short form both reach the same stem.
On an exact-alias miss, falls back to a conservative NEAR-NAME match
(nearNameMatch) so a discussed sub-concept that is plainly the same
subject as an existing article — the article title minus a generic suffix
("" for an " Theory" article) — still resolves to it, while a
distinct subject or a purely-generic query does not. No model, no embedding.
Resolve an entity name or alias to its article stem via the alias index. Normalizes the query the same way aliases are stored (casefold, ws-collapse) and also retries with a leading article word ("the"/"a"/"an") stripped, so a "" surface and a "" short form both reach the same stem.
On an exact-alias miss, falls back to a conservative NEAR-NAME match (nearNameMatch) so a discussed sub-concept that is plainly the same subject as an existing article — the article title minus a generic suffix ("" for an " Theory" article) — still resolves to it, while a
distinct subject or a purely-generic query does not. No model, no embedding.