NEAR-NAME match: resolve a surface to an EXISTING article when it is clearly the
SAME subject — the conservative token-subset fuzzy (a query whose content tokens
are a distinctive subset of an article's, e.g. the title minus a generic suffix).
Reuses EntityResolver.match (stem → squeeze → token-subset fuzzy) but
REFUSES a query made only of generic suffix words (just "Theory"/"Framework"),
which would over-merge onto an arbitrary article. Returns the canonical (== the
article stem the resolver was seeded with), or null. Pure.
NEAR-NAME match: resolve a surface to an EXISTING article when it is clearly the SAME subject — the conservative token-subset fuzzy (a query whose content tokens are a distinctive subset of an article's, e.g. the title minus a generic suffix). Reuses EntityResolver.match (stem → squeeze → token-subset fuzzy) but REFUSES a query made only of generic suffix words (just "Theory"/"Framework"), which would over-merge onto an arbitrary article. Returns the canonical (== the article stem the resolver was seeded with), or null. Pure.