[NOTICKET] fix(knowledge_extraction): show the model the section heading

[NOTICKET] feat(knowledge_extraction): consume the parsing half's ParsedDocument

[NOTICKET] feat(knowledge_extraction): paid extraction stage + validate, diff, queue

[NOTICKET] feat(knowledge_extraction): span filter — E1 recall 0.8537, matches the frozen baseline

[NOTICKET] fix(knowledge_extraction): move runner into the package; fix two ranking defects

[NOTICKET] feat(knowledge_extraction): free stages — models, seam adapter, filters, cluster, ranking