Applies five regex pattern families, filters single-word sentence-start
stopwords, normalizes whitespace, dedupes by exact match, and caps at
MAX_ENTITIES_PER_TEXT.
Parameters
text: string
Free-form text (chunk content, query, etc.).
Returns string[]
Deduplicated array of entity labels in document order.
Extract entity labels from free-form text.
Applies five regex pattern families, filters single-word sentence-start stopwords, normalizes whitespace, dedupes by exact match, and caps at MAX_ENTITIES_PER_TEXT.